Google Git
Sign in
android / platform / external / webkit / 1368e05e8875f00e8d2529fe6050d08b55ea4d87 / . / Tools / ChangeLog-2011-02-16
blob: c736f76970dbedf08c234d1406d65743d6dbb0f9 [file] [log] [blame]
2011-02-16 Anders Carlsson <andersca@apple.com>
Reviewed by Adam Roben.
Assertion failure in DrawingAreaImpl::display when calling WKPageForceRepaint on a page that uses accelerated compositing
https://bugs.webkit.org/show_bug.cgi?id=54575
<rdar://problem/8979594>
Add test.
* TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
* TestWebKitAPI/Tests/WebKit2/ForceRepaint.cpp: Added.
(TestWebKitAPI::didForceRepaint):
(TestWebKitAPI::didFinishLoadForFrame):
(TestWebKitAPI::TEST):
* TestWebKitAPI/Tests/WebKit2/simple-accelerated-compositing.html: Added.
* TestWebKitAPI/win/TestWebKitAPI.vcproj:
* TestWebKitAPI/win/copy-resources.cmd:
2011-02-16 Sergio Villar Senin <svillar@igalia.com>
Reviewed by Martin Robinson.
[Gtk] Add support for layoutTestController.setWillSendRequestClearHeader
https://bugs.webkit.org/show_bug.cgi?id=54537
Added handler for resource-request-starting emitted in
FrameLoader's willSendRequest(). The current implementation
removes the headers set in setWillSendRequestClearHeader.
* DumpRenderTree/gtk/DumpRenderTree.cpp:
(willSendRequestCallback):
(createWebView):
2011-02-16 Csaba Osztrogonác <ossy@webkit.org>
Reviewed by Andreas Kling.
[Qt] Qt build system can't handle removing headers from tree
https://bugs.webkit.org/show_bug.cgi?id=54466
* Scripts/build-webkit: Remove Qt makefiles from productDir before build.
2011-02-16 Ilya Tikhonovsky <loislo@chromium.org>
Reviewed by Yury Semikhatsky.
Web Inspector: chromium: DRT is crashing on the tests in inspector-enabled folder.
https://bugs.webkit.org/show_bug.cgi?id=54548
* DumpRenderTree/chromium/TestShell.cpp:
(TestShell::runFileTest):
2011-02-16 Sheriff Bot <webkit.review.bot@gmail.com>
Unreviewed, rolling out r77898.
http://trac.webkit.org/changeset/77898
https://bugs.webkit.org/show_bug.cgi?id=54541
Leopard Debug hasn't compiled for a long time (Requested by
abarth on #webkit).
* Scripts/webkitpy/common/net/buildbot/buildbot.py:
* Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
2011-02-15 Xianzhu Wang <wangxianzhu@google.com>
Reviewed by Kent Tamura.
Fixed a small bug about --no-timeout option.
https://bugs.webkit.org/show_bug.cgi?id=52873
* DumpRenderTree/chromium/DumpRenderTree.cpp:
(main):
2011-02-15 Kenneth Russell <kbr@google.com>
Reviewed by Darin Fisher.
Allow controlling minimum DOMTimer interval on a per-page basis
https://bugs.webkit.org/show_bug.cgi?id=54312
Implemented new method setMinimumTimerInterval in
LayoutTestController for all ports.
* DumpRenderTree/LayoutTestController.cpp:
(setMinimumTimerIntervalCallback):
(LayoutTestController::staticFunctions):
* DumpRenderTree/LayoutTestController.h:
* DumpRenderTree/chromium/LayoutTestController.cpp:
(LayoutTestController::LayoutTestController):
(LayoutTestController::setMinimumTimerInterval):
* DumpRenderTree/chromium/LayoutTestController.h:
* DumpRenderTree/chromium/WebViewHost.cpp:
(WebViewHost::reset):
* DumpRenderTree/gtk/DumpRenderTree.cpp:
(resetDefaultsToConsistentValues):
* DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
(LayoutTestController::setMinimumTimerInterval):
* DumpRenderTree/mac/DumpRenderTree.mm:
(resetWebViewToConsistentStateBeforeTesting):
* DumpRenderTree/mac/LayoutTestControllerMac.mm:
(LayoutTestController::setMinimumTimerInterval):
* DumpRenderTree/qt/DumpRenderTreeQt.cpp:
(WebCore::WebPage::resetSettings):
* DumpRenderTree/qt/LayoutTestControllerQt.cpp:
(LayoutTestController::setMinimumTimerInterval):
* DumpRenderTree/qt/LayoutTestControllerQt.h:
* DumpRenderTree/win/DumpRenderTree.cpp:
(resetWebViewToConsistentStateBeforeTesting):
* DumpRenderTree/win/LayoutTestControllerWin.cpp:
(LayoutTestController::setMinimumTimerInterval):
2011-02-15 Martin Robinson <mrobinson@igalia.com>
Reviewed by Xan Lopez.
[Freetype] Better map CSS font weight to Fontconfig font weight
https://bugs.webkit.org/show_bug.cgi?id=54323
Add the WeightWatcher fonts to the list of sourced fonts when intializing
DRT. This will allow us to pass tests that use WeightWatcher.
* DumpRenderTree/gtk/DumpRenderTree.cpp:
(initializeFonts): Source the WebkitWeightWatcher fonts.
2011-02-15 Dirk Pranke <dpranke@chromium.org>
Reviewed by Tony Chang.
Make new-run-webkit-tests --processes work under Windows' weird
multiprocessing model. This requires us to be running a Python
file that can be imported as a module, so I've modified the
new-run-webkit-tests script to spawn a subprocess runing
webkitpy.layout_tests.run_webkit_tests.py. I've also modified
the manager_worker_broker tests because you can't pickle a
nested class. This requires us to use globals to pass queues
back and forth for testing, but the alternative is to pass extra
arguments to start_worker() and that feels like a bad tradeoff
just for testing.
https://bugs.webkit.org/show_bug.cgi?id=54431
* Scripts/new-run-webkit-tests:
* Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker_unittest.py:
2011-02-15 Brady Eidson <beidson@apple.com>
Reviewed by Darin Adler.
<rdar://problem/9002047> and https://bugs.webkit.org/show_bug.cgi?id=54426
WebFrameLoaderClient::shouldGoToHistoryItem needs implementation
Add dummy implementations to keep our tools building.
* MiniBrowser/mac/BrowserWindowController.m:
(-[BrowserWindowController awakeFromNib]):
* WebKitTestRunner/TestController.cpp:
(WTR::TestController::initialize):
2011-02-15 David Levin <levin@chromium.org>
Reviewed by Shinichiro Hamaji.
check-webkit-style should do some detection for valid uses of WEBKIT_API (in chromium public api).
https://bugs.webkit.org/show_bug.cgi?id=52271
* Scripts/webkitpy/style/checkers/cpp.py: Added the check for WEBKIT_API usage.
* Scripts/webkitpy/style/checkers/cpp_unittest.py: Added tests.
2011-02-14 Dirk Pranke <dpranke@chromium.org>
Unreviewed, build fix.
Fix minor regression introduced in r78522. For some reason, Python
provides a platform.mac_ver() routine on Linux, so we need to
check to make sure it's returning something in addition to
checking for its existence.
* Scripts/webkitpy/layout_tests/port/mac.py:
2011-02-14 Dirk Pranke <dpranke@chromium.org>
Reviewed by Mihai Parparita.
Clean up the way we handle platform names in
new-run-webkit-tests. Previously, if you specified
'--platform mac' or '--platform chromium-mac' on the command
line, it would use that platform name to determine baseline search
paths, etc. This is almost never the right thing to do, because
it ignores the existence of version-specific baselines.
So, I changed this so that ports that support versions must
always use a version, and the version should default to the
version of the operating system that the port is being run on
(e.g., if you're running on Leopard, then passing '--platform
mac' will default to the 'mac-leopard' port). This means that
port.name() should always include a version-specific suffix if
the port supports versions.
However, we still support things like '--platform mac' as input
so that the user doesn't have to explicitly specify a version,
both for convenience and so we don't have to make buildbot
configs more complicated than they already are.
Also, I cleaned up how the baselines were being searched
on the chromium-mac port (this is one of the two bugs that
prompted this refactoring), and made the logic easier to follow
and more consistent across the mac and chromium ports
Lastly, I added a port.get_all_names() method that could be used
to get the names of all of the ports to search, rather than having
to guess (incorrectly) from the list of directories under
LayoutTests/platform (this last thing was necessary because the
deduplicate_tests unittest broke as the fallback dirs changed).
https://bugs.webkit.org/show_bug.cgi?id=54248
* Scripts/webkitpy/layout_tests/deduplicate_tests.py:
* Scripts/webkitpy/layout_tests/port/factory.py:
* Scripts/webkitpy/layout_tests/port/factory_unittest.py:
* Scripts/webkitpy/layout_tests/port/chromium_gpu_unittest.py:
* Scripts/webkitpy/layout_tests/port/chromium_mac.py:
* Scripts/webkitpy/layout_tests/port/chromium_mac_unittest.py:
* Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
* Scripts/webkitpy/layout_tests/port/chromium_win.py:
* Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
* Scripts/webkitpy/layout_tests/port/mac.py:
* Scripts/webkitpy/layout_tests/port/mac_unittest.py:
* Scripts/webkitpy/layout_tests/port/port_testcase.py:
2011-02-14 Dirk Pranke <dpranke@chromium.org>
Reviewed by Adam Barth.
update-webkit hangs in a git-only checkout
Previously it would hang forever trying to do a 'git svn
rebase'. Now, it will default to a 'git pull' if the
svn-remote.svn.fetch variable isn't set, and bail out if it
doesn't know where to pull from.
https://bugs.webkit.org/show_bug.cgi?id=54265
* Scripts/update-webkit:
* Scripts/VCSUtils.pm:
2011-02-14 Dirk Pranke <dpranke@chromium.org>
Reviewed by Mihai Parparita.
This patch modifies NRWT's Worker clas to use the new
WorkerMixin class to actually run tests. With this patch, the
multiprocess version of NRWT is fully functional and just needs
the values for --worker-model and --child-processes to be corrected
in the port/* classes to start using the new code by default.
https://bugs.webkit.org/show_bug.cgi?id=54082
* Scripts/webkitpy/layout_tests/layout_package/worker.py:
2011-02-14 Dirk Pranke <dpranke@chromium.org>
Reviewed by Mihai Parparita.
This patch shuffles code out of dump_render_tree_thread and
single_test_runner and into a new worker_mixin class that will
be shared by both TestShellThread and Worker. This change
restores single_test_runner to its state as of r77606 - it is
stateless again. It also prunes TestShellThread to pretty much
just the logic used for the original shared-memory access.
https://bugs.webkit.org/show_bug.cgi?id=54079
* Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
* Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py:
* Scripts/webkitpy/layout_tests/layout_package/worker_mixin.py: Added.
2011-02-14 Dirk Pranke <dpranke@chromium.org>
Reviewed by Ojan Vafai.
This patch updates NRWT to add per-worker and per-shard test stats
into TestRunner2, bringing it to parity with TestRunner.
https://bugs.webkit.org/show_bug.cgi?id=54074
* Scripts/webkitpy/layout_tests/layout_package/test_runner2.py:
2011-02-14 Dirk Pranke <dpranke@chromium.org>
Reviewed by Tony Chang.
nrwt multiprocessing: add code to handle interrupts and wedged
threads.
https://bugs.webkit.org/show_bug.cgi?id=54072
* Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker.py:
Adds the cancel(), is_alive(), join(), and log_wedged_worker()
methods to the WorkerConnection class
* Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker_unittest.py:
* Scripts/webkitpy/layout_tests/layout_package/test_runner2.py:
* Scripts/webkitpy/layout_tests/layout_package/worker.py:
Adds the cancel() method to the Worker class
2011-02-14 Dirk Pranke <dpranke@chromium.org>
Reviewed by Tony Chang.
update the NRWT multiprocessing code to spawn multiple workers
instead of just using one.
https://bugs.webkit.org/show_bug.cgi?id=54071
* Scripts/webkitpy/layout_tests/layout_package/test_runner2.py:
2011-02-12 Chang Shu <cshu@webkit.org>
Unreviewed.
Update my own email addresses and IRC nickname.
* Scripts/webkitpy/common/config/committers.py:
2011-02-11 Dirk Pranke <dpranke@chromium.org>
Reviewed by Tony Chang.
This patch adds to NRWT most of the support needed to run the new
message-based workers in separate threads or processes. The code
isn't fully complete yet because we don't support cancel() or
is_alive().
https://bugs.webkit.org/show_bug.cgi?id=54070
* Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker.py:
* Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker_unittest.py:
* Scripts/webkitpy/layout_tests/layout_package/test_runner2.py:
* Scripts/webkitpy/layout_tests/port/base.py:
* Scripts/webkitpy/layout_tests/port/mock_drt.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2011-02-11 Sailesh Agrawal <sail@chromium.org>
Reviewed by Kenneth Russell.
plugins/invalidate_rect.html fails on chromium-mac
https://bugs.webkit.org/show_bug.cgi?id=54051
This change fixes the invalidate_rect.html test failure on Windows.
There were two problems. First, the test specified that the plugin was window less by doing <embed ... windowedPlugin="false"></embed>. The windowedPlugin parameter was never being read by the plugin. Fix was to simply set the NPPVpluginWindowBool variable based on the parameter.
The second problem was that the plugin never handled paint events on Windows. Fix was to simply copy the Mac code to handle paint events.
This change also updates the build path in chromium_win.py to use the new Source directory.
* DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
(invalidateRect):
* DumpRenderTree/TestNetscapePlugIn/main.cpp:
(NPP_New):
(handleEventWin):
(NPP_HandleEvent):
* Scripts/webkitpy/layout_tests/port/chromium_win.py:
2011-02-11 Csaba Osztrogonác <ossy@webkit.org>
Reviewed by Andreas Kling.
[Qt] Increase the timeout of Qt API tests to 120 seconds.
* BuildSlaveSupport/build.webkit.org-config/master.cfg:
2011-02-10 Zhenyao Mo <zmo@google.com>
Unreviewed, build fix.
* DumpRenderTree/DumpRenderTree.gypi:
2011-02-10 Dirk Pranke <dpranke@chromium.org>
Reviewed by Tony Chang.
This patch adds the bulk of the remaining functionality for
the NRWT message-passing implementation. The patch adds a new
Worker class that will eventually replace the TestShellThread
class in dump_render_tree_thread.py, and implements enough of
TestRunner2 and the inline version of the manager_worker_broker
to actually be able to send a full set of messages back and
forth. The Worker stubs do not actually run tests, and there's
lots of error handling and stats needed, but this is the core
logic.
https://bugs.webkit.org/show_bug.cgi?id=54068
* Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker.py:
* Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker_unittest.py:
* Scripts/webkitpy/layout_tests/layout_package/test_runner2.py:
* Scripts/webkitpy/layout_tests/layout_package/worker.py: Added.
* Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2011-02-10 Zhenyao Mo <zmo@google.com>
Unreviewed, build fix.
* DumpRenderTree/DumpRenderTree.gypi:
2011-02-10 Zhenyao Mo <zmo@google.com>
Unreviewed, build fix.
* DumpRenderTree/DumpRenderTree.gypi:
2011-02-10 Zhenyao Mo <zmo@google.com>
Unreviewed, attempt to fix crashing plugin tests.
* DumpRenderTree/DumpRenderTree.gypi:
2011-02-10 Zhenyao Mo <zmo@google.com>
Unreviewed, build fix.
* DumpRenderTree/DumpRenderTree.gypi:
2011-02-10 Sam Weinig <sam@webkit.org>
Try and fix some crashing tests on the chromium build bot.
* DumpRenderTree/DumpRenderTree.gypi:
2011-02-10 Anders Carlsson <andersca@apple.com>
Reviewed by Sam Weinig.
Log an error if a plug-in test can't be found
https://bugs.webkit.org/show_bug.cgi?id=54252
* DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp:
(PluginTest::create):
Return null if the test wasn't found.
* DumpRenderTree/TestNetscapePlugIn/main.cpp:
(NPP_New):
Report an error if PluginTest::create returns null.
2011-02-10 Robert Hogan <robert@webkit.org>
Reviewed by Laszlo Gombos.
[Qt] Return menu items from eventSender.contextMenu()
https://bugs.webkit.org/show_bug.cgi?id=53039
* DumpRenderTree/qt/EventSenderQt.cpp:
(EventSender::contextClick):
* DumpRenderTree/qt/EventSenderQt.h:
2011-02-10 Adam Roben <aroben@apple.com>
Test showing and hiding the find indicator on Windows
Test for <http://webkit.org/b/54213> <rdar://problem/8983261> REGRESSION (r78198): Crash in
FindIndicator::contentImage when scrolling page
Reviewed by Steve Falkenburg.
* TestWebKitAPI/Tests/WebKit2/win/HideFindIndicator.cpp: Added.
(TestWebKitAPI::didFinishLoadForFrame): Record that the load finished.
(TestWebKitAPI::findIndicatorCallback): Record that the callback was called, and save the
bitmap.
(TestWebKitAPI::initialize): Hook up our callbacks.
(TestWebKitAPI::TEST): Test showing then hiding the find indicator to see if we crash.
* TestWebKitAPI/PlatformWebView.h:
* TestWebKitAPI/mac/PlatformWebViewMac.mm:
(TestWebKitAPI::PlatformWebView::page):
* TestWebKitAPI/win/PlatformWebViewWin.cpp:
(TestWebKitAPI::PlatformWebView::page):
Made page a const member function.
* TestWebKitAPI/win/TestWebKitAPI.vcproj: Added the new test.
2011-02-10 Mario Sanchez Prada <msanchez@igalia.com>
Reviewed by Martin Robinson.
[GTK] Log signals from AtkDocument interface also in AccessibilityController
https://bugs.webkit.org/show_bug.cgi?id=54198
Trace AtkDocument's signals emission through AccessibilityController.
Also, taken the code related to adding and removing global
listeners for ATK signals out to a separate file, for the sake of
clarity and to ease future additions.
* DumpRenderTree/gtk/AccessibilityCallbacks.h: Added.
* DumpRenderTree/gtk/AccessibilityCallbacks.cpp: Added
(printAccessibilityEvent): Print information about an event.
(axObjectEventListener): Global listener for AtkObject's signals.
(axDocumentEventListener): Global listener for AtkDocument's signals.
(connectAccessibilityCallbacks): Connect all global listeners.
(disconnectAccessibilityCallbacks): Disconnect all global listeners.
* DumpRenderTree/gtk/AccessibilityControllerGtk.cpp:
(AccessibilityController::setLogAccessibilityEvents): Call to
connectAccessibilityCallbacks and disconnectAccessibilityCallbacks.
* GNUmakefile.am: Added new files.
2011-02-09 Martin Robinson <mrobinson@igalia.com>
Reviewed by Gustavo Noronha Silva.
[GTK] Default error page is interfering with tests which require failed loads
https://bugs.webkit.org/show_bug.cgi?id=54157
Disable the default error page during DRT runs. This prevents unexpected loads
from interfering with tests that have planned failed loads.
* DumpRenderTree/gtk/DumpRenderTree.cpp:
(webViewLoadError): Added, disable default handler.
(createWebView): Connect the new load-error handler.
2011-02-10 Zoltan Horvath <zoltan@webkit.org>
Reviewed by Andreas Kling.
[Qt] Add UrlLoader and command line argument handling to MiniBrowser
https://bugs.webkit.org/show_bug.cgi?id=54192
Copy and modify files from QtTestBrowser's implementation.
* MiniBrowser/qt/BrowserWindow.cpp:
(BrowserWindow::BrowserWindow):
(BrowserWindow::newWindow):
* MiniBrowser/qt/BrowserWindow.h:
* MiniBrowser/qt/MiniBrowser.pro:
* MiniBrowser/qt/MiniBrowserApplication.cpp: Copied from QtTestBrowser/launcherwindow.cpp
(MiniBrowserApplication::MiniBrowserApplication):
(MiniBrowserApplication::handleUserOptions):
* MiniBrowser/qt/MiniBrowserApplication.h: Copied from QtTestBrowser/launcherwindow.h
(WindowOptions::WindowOptions):
(MiniBrowserApplication::urls):
(MiniBrowserApplication::isRobotized):
(MiniBrowserApplication::robotTimeout):
(MiniBrowserApplication::robotExtraTime):
* MiniBrowser/qt/UrlLoader.cpp: Copied from QtTestBrowser/urlloader.cpp
(UrlLoader::UrlLoader):
(UrlLoader::loadNext):
(UrlLoader::checkIfFinished):
(UrlLoader::frameLoadStarted):
(UrlLoader::frameLoadFinished):
(UrlLoader::loadUrlList):
(UrlLoader::getUrl):
* MiniBrowser/qt/UrlLoader.h: Copied from QtTestBrowser/urlloader.h
* MiniBrowser/qt/main.cpp:
(main):
* MiniBrowser/qt/utils.cpp: Copied from QtTestBrowser/utils.cpp
(takeOptionValue):
(formatKeys):
(enumToKeys):
(appQuit):
(urlFromUserInput):
* MiniBrowser/qt/utils.h: Copied from QtTestBrowser/utils.h
2011-02-10 Zoltan Horvath <zoltan@webkit.org>
Reviewed by Andreas Kling.
check-webkit-style: Add -build/include exemption for Tools/MiniBrowser/qt
https://bugs.webkit.org/show_bug.cgi?id=54200
* Scripts/webkitpy/style/checker.py:
* Scripts/webkitpy/style/checker_unittest.py:
2011-02-10 Peter Varga <pvarga@webkit.org>
Reviewed by Csaba Osztrogonác.
Remove PCRE source from trunk
https://bugs.webkit.org/show_bug.cgi?id=54188
* wx/build/settings.py:
2011-02-10 David Levin <levin@chromium.org>
Reviewed by Shinichiro Hamaji.
check-webkit-style should be able to figure out function modifiers and return type.
https://bugs.webkit.org/show_bug.cgi?id=54124
* Scripts/webkitpy/style/checkers/cpp.py:
(_rfind_in_lines): A way to search backwards in lines.
(_FunctionState.modifiers_and_return_type):
* Scripts/webkitpy/style/checkers/cpp_unittest.py: Added tests for the new functionality.
2011-02-09 Adam Roben <aroben@apple.com>
Test that WM_PRINT doesn't trigger an assertion
Test for <http://webkit.org/b/54145> <rdar://problem/8979559> WM_PRINT doesn't work with new
drawing area (assertion failure in DrawingAreaProxyImpl in Debug builds)
Reviewed by Darin Adler.
* TestWebKitAPI/Tests/WebKit2/win/WMPrint.cpp: Added.
(TestWebKitAPI::TEST): Send a WM_PRINT message to a WKView.
2011-02-09 Adam Roben <aroben@apple.com>
Test that having a WKView paint after being resized while hidden doesn't cause a crash
Test for <http://webkit.org/b/54142> <rdar://problem/8979365> Crash in
DrawingAreaProxyImpl::paint when WKView paints after being resized while hidden
Reviewed by Darin Adler.
* TestWebKitAPI/Tests/WebKit2/win/ResizeViewWhileHidden.cpp:
(TestWebKitAPI::TEST): Force the WKView to paint after resizing it while hidden.
2011-02-09 Adam Roben <aroben@apple.com>
Add a test that resizes a WKView while it's hidden
Test for <http://webkit.org/b/54141> <rdar://problem/8979195> WKView draws at wrong size,
performs badly if is resized while hidden (assertion fails in
BackingStore::incorporateUpdate in Debug builds)
Reviewed by Darin Adler.
* TestWebKitAPI/Tests/WebKit2/win/ResizeViewWhileHidden.cpp: Added.
(TestWebKitAPI::didFinishLoadForFrame): Record that the load finished.
(TestWebKitAPI::setPageLoaderClient): Hook up our didFinishLoadForFrame callback.
(TestWebKitAPI::flushMessages): Load a URL and wait for the load to complete. This ensures
that all pending messages have been handled by the UI and web process.
(TestWebKitAPI::timerCallback): Kill the timer and record that it fired.
(TestWebKitAPI::runForDuration): Set a timer and run the run loop until it fires.
(TestWebKitAPI::waitForBackingStoreUpdate): Wait for half a second to give the web process a
chance to display, then flush all pending messages.
(TestWebKitAPI::TEST): Resize the WKView while it's hidden, then show it again and wait for
the backing store to update. This triggers the assertion from bug 54141.
* TestWebKitAPI/win/PlatformWebViewWin.cpp:
(TestWebKitAPI::PlatformWebView::resizeTo): Implemented.
* TestWebKitAPI/win/TestWebKitAPI.vcproj: Added new test.
2011-02-10 Philippe Normand <pnormand@igalia.com>
Unreviewed, GTK build fix after roll out of r78157...
* GNUmakefile.am:
2011-02-10 Eric Seidel <eric@webkit.org>
Unreviewed, rolling out r78157.
http://trac.webkit.org/changeset/78157
https://bugs.webkit.org/show_bug.cgi?id=54150
Fails on a bunch of bots
* DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
* DumpRenderTree/TestNetscapePlugIn/Tests/EvaluateJSDestroyingPluginFromDestroyStream.cpp: Removed.
* DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
* DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro:
2011-02-10 Mario Sanchez Prada <msanchez@igalia.com>
Reviewed by Xan Lopez.
[GTK] Reset GTK' DRT's AccessibilityController to consistent value before every test
https://bugs.webkit.org/show_bug.cgi?id=54185
* DumpRenderTree/gtk/DumpRenderTree.cpp:
(resetDefaultsToConsistentValues):
2011-02-10 Philippe Normand <pnormand@igalia.com>
Unreviewed GTK build fix.
Adding
Tools/DumpRenderTree/TestNetscapePlugIn/Tests/EvaluateJSDestroyingPluginFromDestroyStream.cpp
in the build, fixing the
plugins/npruntime/evaluate-js-destroying-plugin-from-destroy-stream.html
test.
* GNUmakefile.am:
2011-02-10 Mario Sanchez Prada <msanchez@igalia.com>
Reviewed by Chris Fleizach.
[GTK] Add support in DRT to log "accessibility events"
https://bugs.webkit.org/show_bug.cgi?id=54116
Added a way to log accessibility related events, in a
platform-dependant way.
Add new function to AccessibilityController.
* DumpRenderTree/AccessibilityController.h:
* DumpRenderTree/AccessibilityController.cpp:
(logAccessibilityEventsCallback): New callback.
(AccessibilityController::getJSClass): Added new function.
(AccessibilityController::resetToConsistentState): Initialized
calling setLogAccessibilityEvents(false).
Provided implementation for the GTK port.
* DumpRenderTree/gtk/AccessibilityControllerGtk.cpp:
(accessibility_event_listener): Common listener for logging
information about all the signals emitted by any AtkObject.
(AccessibilityController::setLogAccessibilityEvents): Add or
remove listeners for signals, as specified by the parameter.
Provide dummy implementations for mac and win ports.
* DumpRenderTree/mac/AccessibilityControllerMac.mm:
(AccessibilityController::setLogAccessibilityEvents):
* DumpRenderTree/win/AccessibilityControllerWin.cpp:
(AccessibilityController::setLogAccessibilityEvents):
2011-02-09 Hayato Ito <hayato@chromium.org>
Reviewed by Tony Chang.
[NRWT] Remove encoding parameters where we can assume data can be
written in binary mode.
https://bugs.webkit.org/show_bug.cgi?id=54066
* Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py:
* Scripts/webkitpy/layout_tests/port/base.py:
2011-02-09 Dirk Pranke <dpranke@chromium.org>
Reviewed by Tony Chang.
nrwt multiprocessing: minor cleanup prior to implementing the new worker
This patch renames AbstractManager to ManagerConnection and
changes some argument names to be more consistent between
manager_worker_broker and message_broker. It also fixes a couple
of typos in message_broker. These changes will be tested by code
introduced in the next patch.
https://bugs.webkit.org/show_bug.cgi?id=54067
* Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker.py:
* Scripts/webkitpy/layout_tests/layout_package/message_broker2.py:
2011-02-09 Anders Carlsson <andersca@apple.com>
Reviewed by Sam Weinig.
Repro crash with Sony Google TV ad at Gizmodo
https://bugs.webkit.org/show_bug.cgi?id=54150
<rdar://problem/8782346>
Add a new plug-in test that runs JavaScript that destroys the plug-in from within its NPN_DestroyStream callback.
* DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
* DumpRenderTree/TestNetscapePlugIn/Tests/EvaluateJSDestroyingPluginFromDestroyStream.cpp: Added.
(EvaluateJSDestroyingPluginFromDestroyStream::EvaluateJSDestroyingPluginFromDestroyStream):
(EvaluateJSDestroyingPluginFromDestroyStream::NPP_Destroy):
(EvaluateJSDestroyingPluginFromDestroyStream::NPP_DestroyStream):
* DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
* DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro:
2011-02-09 David Levin <levin@chromium.org>
Reviewed by Eric Seidel.
check-webkit-style: The error message about NULL should be more clear for comments.
https://bugs.webkit.org/show_bug.cgi?id=53786
* Scripts/webkitpy/style/checkers/cpp.py: Clarified the error message.
* Scripts/webkitpy/style/checkers/cpp_unittest.py: Fixed the corresponding test.
2011-02-09 Nebojsa Ciric <cira@chromium.org>
Reviewed by Darin Fisher.
Implements Locale object of JavaScript internationalization API proposal, as an
v8 extension. Extension code is hosted in v8 project (src/extensions/experimental/i18n-extension.{cc,h})
and in this patch we just provide flags, tests and build rules for chromium port.
https://bugs.webkit.org/show_bug.cgi?id=49414
* DumpRenderTree/chromium/TestShell.cpp:
(TestShell::TestShell):
2011-02-09 Laszlo Gombos <laszlo.1.gombos@nokia.com>
Reviewed by Kenneth Rohde Christiansen.
[Qt] Remove UiTools dependency from MiniBrowser
https://bugs.webkit.org/show_bug.cgi?id=54096
* MiniBrowser/qt/MiniBrowser.pro:
2011-02-08 Hayato Ito <hayato@chromium.org>
Reviewed by Tony Chang.
[NRWT] Pull up rebaseline code from compare_output() function defined
in text_diff.py and image_diff.py into a SingleTestRunner.
This patch is a first step for eliminating test_type/* classes.
https://bugs.webkit.org/show_bug.cgi?id=53071
* Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py:
* Scripts/webkitpy/layout_tests/test_types/image_diff.py:
* Scripts/webkitpy/layout_tests/test_types/test_type_base.py:
* Scripts/webkitpy/layout_tests/test_types/text_diff.py:
2011-02-07 Ojan Vafai <ojan@chromium.org>
Reviewed by Mihai Parparita.
stop generating results.json files
https://bugs.webkit.org/show_bug.cgi?id=53977
We've only used incremental_results.json for a while now
and there are plans to start generating a results.json file that matches
the format of unexpected_results.json.
* Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
* Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
* Scripts/webkitpy/layout_tests/layout_package/json_results_generator_unittest.py:
* Scripts/webkitpy/layout_tests/layout_package/test_runner.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2011-02-08 Sailesh Agrawal <sail@chromium.org>
Reviewed by Kenneth Russell.
Invalidate rect doesn't work for windowless plugins on Chromium
https://bugs.webkit.org/show_bug.cgi?id=53117
Added two new utility methods.
- layoutTestController.displayInvalidatedRegion() does a paint of any area that has been invalidated. This is different from layoutTestController.display() which explicitly invalidates the entire page then paints.
- plugin.invalidateRect(left, top, right, bottom) - invalidates the given rect
This is used to test that invalidating a rect correctly causes a repaint of the plugin.
* DumpRenderTree/LayoutTestController.cpp:
(displayInvalidatedRegionCallback):
(LayoutTestController::staticFunctions):
* DumpRenderTree/LayoutTestController.h:
* DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
(invalidateRect):
(pluginInvoke):
* DumpRenderTree/chromium/LayoutTestController.cpp:
(LayoutTestController::LayoutTestController):
(LayoutTestController::displayInvalidatedRegion):
* DumpRenderTree/chromium/LayoutTestController.h:
* Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
2011-02-08 Dirk Pranke <dpranke@chromium.org>
Reviewed by Tony Chang.
new-run-webkit-tests: move the logic that starts and stops the
servers from dump_render_tree_thread into single_test_runner
so that we can reuse it in the new multiprocessing worker class
as well.
https://bugs.webkit.org/show_bug.cgi?id=53840
* Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
* Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py:
2011-02-08 Dirk Pranke <dpranke@chromium.org>
Reviewed by Mihai Parparita.
new-run-webkit-tests: split out thread stack logging code into a sharable module
This patch splits out the code used to find and log thread
stacks from NRWT-specific packages to something generic and
shareable by other python modules. It will be shared in the near
future by the manager_worker_broker module, for example.
https://bugs.webkit.org/show_bug.cgi?id=53656
* Scripts/webkitpy/common/system/stack_utils.py: Added.
* Scripts/webkitpy/common/system/stack_utils_unittest.py: Added.
* Scripts/webkitpy/layout_tests/layout_package/message_broker.py:
* Scripts/webkitpy/layout_tests/layout_package/message_broker_unittest.py:
2011-02-08 Dirk Pranke <dpranke@chromium.org>
Reviewed by Tony Chang.
new-run-webkit-tests: move a bunch of testing logic out of
dump_render_tree_thread and into single_test_runner so that we
will be able to reuse it in the new multiprocessing worker class as well.
https://bugs.webkit.org/show_bug.cgi?id=53838
* Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
* Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py:
2011-02-08 Dirk Pranke <dpranke@chromium.org>
Reviewed by Ojan Vafai.
new-run-webkit-tests: remove no longer needed WatchableThread
class.
https://bugs.webkit.org/show_bug.cgi?id=53839
* Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
2011-02-08 Dirk Pranke <dpranke@chromium.org>
Reviewed by Tony Chang.
new-run-webkit-tests: add stubs for the Manager objects that
will abstract the concurrency implementation (inline / threads /
processes). These classes do nothing yet and are not wired up to
anything.
https://bugs.webkit.org/show_bug.cgi?id=53477
* Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker.py: Added.
* Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker_unittest.py: Added.
2011-02-08 Dirk Pranke <dpranke@chromium.org>
Reviewed by Tony Chang.
new-run-webkit-tests: add simplified message broker for new-style
messaging. This change adds a very simple message broker that
will be used to shuttle messages between the TestRunner2 manager
thread and the Worker threads. For now the class is not used by
anything, but the eventual usage can be seen in the patches
attached to bug 49566.
https://bugs.webkit.org/show_bug.cgi?id=53158
* Scripts/webkitpy/layout_tests/layout_package/message_broker2.py: Added.
* Scripts/webkitpy/layout_tests/layout_package/message_broker2_unittest.py: Added.
2011-02-08 Adam Barth <abarth@webkit.org>
Reviewed by Dimitri Glazkov.
chromium-win builder shouldn't run python or perl tests
https://bugs.webkit.org/show_bug.cgi?id=54032
These tests don't pass on this builder because the builder isn't
running in cygwin. There isn't really any point in running them and
making the bot red forever.
* BuildSlaveSupport/build.webkit.org-config/master.cfg:
2011-02-08 Brady Eidson <beidson@apple.com>
Reviewed by Darin Adler.
<rdar://problem/8972913> and https://bugs.webkit.org/show_bug.cgi?id=54036
didChangeBackForwardList should include some context about what changed
* MiniBrowser/mac/BrowserWindowController.m:
(didChangeBackForwardList):
2011-02-08 Sam Weinig <sam@webkit.org>
Reviewed by Anders Carlsson.
WK2: Add ability to pass context to policy delegate methods
https://bugs.webkit.org/show_bug.cgi?id=54031
* MiniBrowser/mac/BrowserWindowController.m:
(decidePolicyForNavigationAction):
(decidePolicyForNewWindowAction):
(decidePolicyForMIMEType):
* TestWebKitAPI/Tests/WebKit2/PageLoadBasic.cpp:
(TestWebKitAPI::decidePolicyForNavigationAction):
(TestWebKitAPI::decidePolicyForNewWindowAction):
(TestWebKitAPI::decidePolicyForMIMEType):
Update policy client for new API.
2011-02-08 Martin Robinson <mrobinson@igalia.com>
Reviewed by Xan Lopez.
[GTK] DRT needs an implementation of LayoutTestController.setIconDatabaseEnabled
https://bugs.webkit.org/show_bug.cgi?id=54033
Add an implementation of LayoutTestController.setIconDatabaseEnabled that just
call DumpRenderTreeSupportGtk. Turn off the icon database between tests.
* DumpRenderTree/gtk/DumpRenderTree.cpp:
(resetDefaultsToConsistentValues): Turn off the icon database.
* DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
(LayoutTestController::setIconDatabaseEnabled): Call the appropriate DumpRenderTreeSupportGtk
method.
2011-02-08 Kundu Suchismita <suchi.kundu@nokia.com>
Reviewed by Laszlo Gombos.
Local Storage settings can be enable/disable from "Develop" menu
https://bugs.webkit.org/show_bug.cgi?id=52296
* QtTestBrowser/launcherwindow.cpp:
(LauncherWindow::createChrome):
(LauncherWindow::toggleLocalStorage):
(LauncherWindow::toggleOfflineStorageDatabase):
(LauncherWindow::toggleOfflineWebApplicationCache):
(LauncherWindow::setOfflineStorageDefaultQuota):
* QtTestBrowser/launcherwindow.h:
(WindowOptions::WindowOptions):
* QtTestBrowser/main.cpp:
(LauncherApplication::handleUserOptions):
2011-02-08 Martin Robinson <mrobinson@igalia.com>
Reviewed by Eric Seidel.
[GTK] EventSender.keyDown does not support non-array modifier arguments
https://bugs.webkit.org/show_bug.cgi?id=53962
Support handling either an array or a string for the modifier argument
to EventSender.keyDown.
* DumpRenderTree/gtk/EventSender.cpp:
(gdkModifierFromJSValue): Added this helper which factors out the
string comparison bits.
(gdkModifersFromJSValue): Test up front whether the value is a string,
to handle it specially.
2011-02-08 Benjamin Poulain <ikipou@gmail.com>
Reviewed by Csaba Osztrogonác.
openDatabaseSync() stop responding after too many call
https://bugs.webkit.org/show_bug.cgi?id=53945
Reset the database quota to a known state between each test.
Define a quota for the known origins at each test run and delete
all the databases.
This way, the database related test do not depend on previous
allocations.
* DumpRenderTree/qt/DumpRenderTreeQt.cpp:
(WebCore::DumpRenderTree::resetToConsistentStateBeforeTesting):
(WebCore::DumpRenderTree::dumpDatabaseQuota):
* DumpRenderTree/qt/LayoutTestControllerQt.cpp:
(LayoutTestController::reset):
2011-02-08 Carlos Garcia Campos <cgarcia@igalia.com>
Reviewed by Martin Robinson.
[GTK] DRT's TextInputController is unimplemented on GTK
https://bugs.webkit.org/show_bug.cgi?id=52997
Initial implementation of TextInputController for GTK port.
* DumpRenderTree/gtk/DumpRenderTree.cpp:
(webViewWindowObjectCleared):
* DumpRenderTree/gtk/TextInputController.cpp: Added.
(setMarkedTextCallback):
(insertTextCallback):
(unmarkTextCallback):
(firstRectForCharacterRangeCallback):
(selectedRangeCallback):
(getClass):
(makeTextInputController):
* DumpRenderTree/gtk/TextInputController.h: Added.
* GNUmakefile.am:
2011-02-07 Tony Chang <tony@chromium.org>
Reviewed by Adam Barth.
clean up python imports
https://bugs.webkit.org/show_bug.cgi?id=53966
Convert "import A.B.C.D as D" to "from A.B.C import D" and
make some imports absolute as required by PEP-8
* Scripts/webkitpy/common/checkout/scm.py:
* Scripts/webkitpy/common/net/irc/ircbot.py:
* Scripts/webkitpy/common/system/logutils.py:
* Scripts/webkitpy/common/system/logutils_unittest.py:
* Scripts/webkitpy/layout_tests/deduplicate_tests.py:
* Scripts/webkitpy/layout_tests/port/chromium.py:
* Scripts/webkitpy/layout_tests/port/mac.py: Remove some unused imports
* Scripts/webkitpy/layout_tests/port/webkit.py:
* Scripts/webkitpy/layout_tests/update_webgl_conformance_tests.py:
* Scripts/webkitpy/tool/bot/irc_command.py:
* Scripts/webkitpy/tool/bot/sheriffircbot.py:
* Scripts/webkitpy/tool/commands/download.py:
* Scripts/webkitpy/tool/commands/prettydiff.py:
* Scripts/webkitpy/tool/commands/queries.py:
* Scripts/webkitpy/tool/commands/roll.py:
* Scripts/webkitpy/tool/commands/stepsequence.py:
* Scripts/webkitpy/tool/commands/upload.py:
* Scripts/webkitpy/tool/main.py:
2011-02-07 James Robinson <jamesr@chromium.org>
Reviewed by Adam Barth.
Teach do-webcore-rename about the WebKit move to Source/
https://bugs.webkit.org/show_bug.cgi?id=53967
* Scripts/do-webcore-rename:
2011-02-07 Maciej Stachowiak <mjs@apple.com>
Reviewed by Dan Bernstein.
Add resource load client for injected bundle and move willSendRequest there
https://bugs.webkit.org/show_bug.cgi?id=53972
* MiniBrowser/mac/WebBundle/WebBundleMain.m:
(didCreatePage):
* WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
(WTR::InjectedBundlePage::InjectedBundlePage):
(WTR::InjectedBundlePage::didInitiateLoadForResource):
(WTR::InjectedBundlePage::didReceiveResponseForResource):
(WTR::InjectedBundlePage::didReceiveContentLengthForResource):
(WTR::InjectedBundlePage::didFinishLoadForResource):
(WTR::InjectedBundlePage::didFailLoadForResource):
(WTR::InjectedBundlePage::willSendRequestForFrame):
* WebKitTestRunner/InjectedBundle/InjectedBundlePage.h:
2011-02-07 Adam Barth <abarth@webkit.org>
Add Leopard Debug back to the core builders.
* Scripts/webkitpy/common/net/buildbot/buildbot.py:
* Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
2011-02-07 Adam Klein <adamk@chromium.org>
Reviewed by Jian Li.
[chromium] Remove deprecated method WebNotification::dir
https://bugs.webkit.org/show_bug.cgi?id=53735
* DumpRenderTree/chromium/NotificationPresenter.cpp:
(NotificationPresenter::show):
2011-02-07 Adam Barth <abarth@webkit.org>
Reviewed by Eric Seidel.
Remove Leopard Debug as a core builder
https://bugs.webkit.org/show_bug.cgi?id=53971
The Leopard Debug builder has been failing to compile for several days.
My understanding is that folks are working on fixing the build, but for
the time being we should probably remove it from the list of core
builders so that the core waterfall can be green. We can certainly add
it back once the build is fixed.
Also, I've removed the Tiger builders because those no longer appear to
be attached to the buildbot master.
* Scripts/webkitpy/common/net/buildbot/buildbot.py:
* Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
2011-02-01 Ojan Vafai <ojan@chromium.org>
Reviewed by Adam Roben.
include svn revisions in git diffs for the code review tool to use
https://bugs.webkit.org/show_bug.cgi?id=53569
* Scripts/webkitpy/common/checkout/scm.py:
* Scripts/webkitpy/common/checkout/scm_unittest.py:
2011-02-03 MORITA Hajime <morrita@google.com>
Reviewed by Darin Fisher.
[Chromium] Should implement EditorClientImpl::requestCheckingOfString()
https://bugs.webkit.org/show_bug.cgi?id=51013
Gave DRT implementation for requestTextCheck().
* DumpRenderTree/chromium/LayoutTestController.cpp:
(LayoutTestController::setAsynchronousSpellCheckingEnabled): Implemented.
* DumpRenderTree/chromium/WebViewHost.cpp:
(invokeFinishLastTextCheck): Added.
(WebViewHost::requestTextCheck): Added.
(WebViewHost::finishLastTextCheck): Added.
* DumpRenderTree/chromium/WebViewHost.h:
2011-02-07 Joone Hur <joone.hur@collabora.co.uk>
Reviewed by Martin Robinson.
[Gtk] Implement layoutTestController.findString
https://bugs.webkit.org/show_bug.cgi?id=50237
* DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
(LayoutTestController::findString):
2011-02-07 Chang Shu <chang.shu@nokia.com>
Reviewed by Darin Adler.
We should disable spatial navigation explicitly during the initialization
of DumpRenderTree.
https://bugs.webkit.org/show_bug.cgi?id=53928
* DumpRenderTree/mac/DumpRenderTree.mm:
(resetDefaultsToConsistentValues):
2011-02-07 Maciej Stachowiak <mjs@apple.com>
Not reviewed.
Fix WebKitTestRunner build on the SL bot.
* WebKitTestRunner/Configurations/Base.xcconfig: Look for the JSC copy
of ICU headers.
2011-02-07 Csaba Osztrogonác <ossy@webkit.org>
Unreviewed.
[Qt][WK2] Buildfix after r77794.
WebKitTestRunner does not block remote resources or complain about them
https://bugs.webkit.org/show_bug.cgi?id=42139
<rdar://problem/8183942>
* WebKitTestRunner/InjectedBundle/qt/InjectedBundle.pro: Add missing includepaths.
2011-02-07 Maciej Stachowiak <mjs@apple.com>
Not reviewed.
More bot appeasement.
* WebKitTestRunner/WebKitTestRunner.xcodeproj/project.pbxproj:
2011-02-07 Maciej Stachowiak <mjs@apple.com>
Not reviewed.
Remove accidental references to directories on my laptop.
* WebKitTestRunner/WebKitTestRunner.xcodeproj/project.pbxproj:
2011-02-07 Maciej Stachowiak <mjs@apple.com>
Reviewed by Antti Koivisto.
WebKitTestRunner does not block remote resources or complain about them
https://bugs.webkit.org/show_bug.cgi?id=42139
<rdar://problem/8183942>
* WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
(WTR::InjectedBundlePage::willSendRequestForFrame): Implement the required
checks (using KURL, to avoid need to invent a whole URL API).
* WebKitTestRunner/WebKitTestRunner.xcodeproj/project.pbxproj: Link WebCore
directly to get at KURL symbols.
2011-02-06 Maciej Stachowiak <mjs@apple.com>
Reviewed by Daniel Bates.
Add WebKit2 bot to list of expected builders in Python regression test results.
https://bugs.webkit.org/show_bug.cgi?id=53905
* Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
2011-02-06 Maciej Stachowiak <mjs@apple.com>
Reviewed by Ryosuke Niwa.
Update test expectations for new core builder
https://bugs.webkit.org/show_bug.cgi?id=53904
* Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
2011-02-06 Maciej Stachowiak <mjs@apple.com>
Reviewed by David Levin.
Add WebKit2 test bot to core builders
https://bugs.webkit.org/show_bug.cgi?id=53901
* Scripts/webkitpy/common/net/buildbot/buildbot.py:
2011-02-06 Ryosuke Niwa <rniwa@webkit.org>
Another unreviewed Chromium build fix.
* DumpRenderTree/chromium/LayoutTestController.cpp:
(LayoutTestController::pathToLocalResource):
2011-02-06 Ryosuke Niwa <rniwa@webkit.org>
Unreviewed Chromium build fix.
* DumpRenderTree/chromium/LayoutTestController.cpp:
(LayoutTestController::pathToLocalResource):
2011-02-06 Ryosuke Niwa <rniwa@webkit.org>
Reviewed by Darin Adler.
OwnArraryPtr.h uses deleteOwnedPtr but doesn’t include OwnPtrCommon.h
https://bugs.webkit.org/show_bug.cgi?id=52867
* DumpRenderTree/chromium/ImageDiff.cpp:
(Image::craeteFromStdin): Call adoptArrayPtr.
2011-02-06 James Kozianski <koz@chromium.org>
Reviewed by Ojan Vafai.
Add classes for representing test outputs.
https://bugs.webkit.org/show_bug.cgi?id=52136
These classes will make it easier to write scripts that deal with test
outputs such as rebaselining and deduping scripts. The intent is that
eventually we will be have buildbot return TestOutputs for a
particular build which can be compared with TestOutputs derived from a
local LayoutTests directory.
* Scripts/webkitpy/common/net/testoutput.py: Added.
* Scripts/webkitpy/common/net/testoutput_unittest.py: Added.
* Scripts/webkitpy/common/net/testoutputset.py: Added.
* Scripts/webkitpy/common/net/testoutputset_unittest.py: Added.
* Scripts/webkitpy/common/system/zip_mock.py: Added.
2011-02-06 Robert Hogan <robert@webkit.org>
Reviewed by Andreas Kling.
[Qt] Clear page's groupName even when not in DRT
https://bugs.webkit.org/show_bug.cgi?id=53874
* DumpRenderTree/qt/DumpRenderTreeQt.cpp:
(WebCore::DumpRenderTree::windowCloseRequested):
2011-02-06 Patrick Gansterer <paroga@webkit.org>
Reviewed by Andreas Kling.
[CMake] Add dependencies for Visual Studio projects
https://bugs.webkit.org/show_bug.cgi?id=53773
Add a WebKit dependecy to WinCELauncher, so CMake can
generate the correct build order for the solution.
* CMakeListsWinCE.txt:
2011-02-06 Andreas Kling <kling@webkit.org>
Reviewed by Antonio Gomes.
[Qt] MiniBrowser: Clean up handling of titleChanged() signal
https://bugs.webkit.org/show_bug.cgi?id=53869
* MiniBrowser/qt/BrowserView.cpp:
(BrowserView::BrowserView): Remove unnecessary titleChanged() handling.
* MiniBrowser/qt/BrowserWindow.cpp:
(BrowserWindow::BrowserWindow): Connect titleChanged() directly to setWindowTitle().
* MiniBrowser/qt/BrowserWindow.h:
2011-02-06 Maciej Stachowiak <mjs@apple.com>
Reviewed by Dan Bernstein.
WebKitTestRunner needs layoutTestController.setWillSendRequestReturnsNull
https://bugs.webkit.org/show_bug.cgi?id=42690
<rdar://problem/8213851>
* WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl:
* WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
(WTR::InjectedBundlePage::willSendRequestForFrame):
* WebKitTestRunner/InjectedBundle/LayoutTestController.cpp:
(WTR::LayoutTestController::LayoutTestController):
* WebKitTestRunner/InjectedBundle/LayoutTestController.h:
(WTR::LayoutTestController::willSendRequestReturnsNull):
(WTR::LayoutTestController::setWillSendRequestReturnsNull):
2011-02-05 Sheriff Bot <webkit.review.bot@gmail.com>
Unreviewed, rolling out r77720.
http://trac.webkit.org/changeset/77720
https://bugs.webkit.org/show_bug.cgi?id=53854
"Broke nrwt on Chromium win." (Requested by dglazkov|away on
#webkit).
* Scripts/webkitpy/common/system/executive.py:
2011-02-05 Sheriff Bot <webkit.review.bot@gmail.com>
Unreviewed, rolling out r77725.
http://trac.webkit.org/changeset/77725
https://bugs.webkit.org/show_bug.cgi?id=53844
It broke Qt minimal build (Requested by Ossy_ on #webkit).
* QtTestBrowser/launcherwindow.cpp:
(LauncherWindow::createChrome):
* QtTestBrowser/launcherwindow.h:
(WindowOptions::WindowOptions):
* QtTestBrowser/main.cpp:
(LauncherApplication::handleUserOptions):
2011-02-04 Martin Robinson <mrobinson@igalia.com>
Reviewed by Adam Barth.
[GTK] plugins/plugin-document-back-forward.html fails
https://bugs.webkit.org/show_bug.cgi?id=53833
Add the new "alert on load" functionality for the Unix version of the
test plugin. This functionality was originally introduced in r77706.
* DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
(webkit_test_plugin_new_instance): Add new plugin fuctionality.
2011-02-04 Kundu Suchismita <suchi.kundu@nokia.com>
Reviewed by Andreas Kling.
Local Storage settings can be enable/disable from "Develop" menu
https://bugs.webkit.org/show_bug.cgi?id=52296
* QtTestBrowser/launcherwindow.cpp:
(LauncherWindow::createChrome):
(LauncherWindow::toggleLocalStorage):
(LauncherWindow::toggleOfflineStorageDatabase):
(LauncherWindow::toggleOfflineWebApplicationCache):
(LauncherWindow::setOfflineStorageDefaultQuota):
* QtTestBrowser/launcherwindow.h:
(WindowOptions::WindowOptions):
* QtTestBrowser/main.cpp:
(LauncherApplication::handleUserOptions):
2011-02-04 Tony Chang <tony@chromium.org>
Reviewed by Eric Seidel.
get test-webkitpy running on win32 python
https://bugs.webkit.org/show_bug.cgi?id=53822
The test harness crashes with a WindowsError because it can't find
'svn' when using subprocess.Popen. This gets us past the error
so we can see the failing tests on the Chromium Win Release Tests
bot.
* Scripts/webkitpy/common/system/executive.py:
2011-02-04 Martin Robinson <mrobinson@igalia.com>
Reviewed by Gustavo Noronha Silva.
[GTK] fast/events/pagehide-timeout.html fails
https://bugs.webkit.org/show_bug.cgi?id=53771
Add knowledge of the document browser cache model the DRT. When resetting
WebKit settings to consistent values between tests, default to the document
browser cache model.
* DumpRenderTree/gtk/DumpRenderTree.cpp:
(resetDefaultsToConsistentValues): Reset the cache model to the document browser
cache model between tests.
* DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
(LayoutTestController::setCacheModel): An int value of 2 here corresponds to the
document browser cache model.
2011-02-04 Adam Barth <abarth@webkit.org>
Reviewed by Eric Seidel.
PluginDocuments don't create widgets for plugins on back/forward
https://bugs.webkit.org/show_bug.cgi?id=53474
Teach the test plugin how to call alert on load.
* DumpRenderTree/TestNetscapePlugIn/main.cpp:
(NPP_New):
2011-02-04 Adam Roben <aroben@apple.com>
Add a test case for encode/decode of FormData/FormDataElement
If the fix made in r77401 is broken, this new test will crash.
Fixes <http://webkit.org/b/53616> <rdar://problem/8949883>.
Reviewed by Sam Weinig.
* TestWebKitAPI/Test.h: Moved the bulk of TEST_ASSERT into a new _TEST_ASSERT_HELPER macro.
Added a new TEST_ASSERT_RETURN macro that can be used in functions with a return value.
* TestWebKitAPI/Tests/WebKit2/RestoreSessionStateContainingFormData.cpp: Added.
(TestWebKitAPI::didFinishLoadForFrame): Record that the load is finished.
(TestWebKitAPI::setPageLoaderClient): Hook up the loader client.
(TestWebKitAPI::createSessionStateContainingFormData): Load simple-form.html, submit the
form, an return the session state data.
(TestWebKitAPI::TEST): Create some session state that contains form data from one WKPage,
and restore it into another WKPage.
* TestWebKitAPI/Tests/WebKit2/simple-form.html: Added.
* TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
* TestWebKitAPI/win/TestWebKitAPI.vcproj:
* TestWebKitAPI/win/copy-resources.cmd:
Added new files.
2011-02-04 Adam Roben <aroben@apple.com>
Move code to run JavaScript tests into its own files
This will allow other tests to use this mechanism.
Fixes <http://webkit.org/b/53806> SpacebarScrolling should share its JavaScript-fu with the
world
Reviewed by Sam Weinig.
* TestWebKitAPI/JavaScriptTest.cpp: Added.
* TestWebKitAPI/JavaScriptTest.h: Added.
Moved code here...
* TestWebKitAPI/Tests/WebKit2/SpacebarScrolling.cpp: ...from here.
* TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
* TestWebKitAPI/win/TestWebKitAPI.vcproj:
Added the new files.
2011-02-04 Dirk Pranke <dpranke@chromium.org>
Reviewed by Mihai Parparita.
merge test expectations for chromium, chromium-gpu. The
chromium-gpu port will no longer maintain its own
test expectations file.
Also add in a graphics_type() call on the port object to
determine what graphics type to use in the test configuration,
and log the configuration as a config setting.
https://bugs.webkit.org/show_bug.cgi?id=53562
* Scripts/webkitpy/layout_tests/port/base.py:
* Scripts/webkitpy/layout_tests/layout_package/test_runner.py:
* Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
* Scripts/webkitpy/layout_tests/port/chromium_gpu_unittest.py:
2011-02-04 Mikhail Naganov <mnaganov@chromium.org>
Reviewed by Pavel Feldman.
Web Inspector: fix processing Unicode literals from .js sources in check-inspector-strings.
https://bugs.webkit.org/show_bug.cgi?id=53781
* Scripts/check-inspector-strings:
2011-02-04 Adam Roben <aroben@apple.com>
Include the crashing function in the link to a crash log on Mac
Fixes <http://webkit.org/b/53780> Crash log links in results.html should include the
function that crashed on Mac
Reviewed by David Kilzer.
* Scripts/old-run-webkit-tests:
(crashLocation): Moved all the Windows code inside an if instead of using an early return,
but didn't otherwise change it. Added an if for Mac that parses the crashing function out of
the crash log.
2011-02-04 Adam Roben <aroben@apple.com>
Link to Mac crash logs from results.html
Fixes <http://webkit.org/b/14861> run-webkit-tests should link to Mac crash logs in
results.html
Reviewed by David Kilzer.
* Scripts/old-run-webkit-tests:
(captureSavedCrashLog): Added a case for Apple's Mac port. We get the crash log out of
~/Library/Logs/CrashReporter, and wait for ReportCrash to exit before trying to get the log.
2011-02-03 Maciej Stachowiak <mjs@apple.com>
Reviewed by Dan Bernstein.
WebKit2: Need WebKit2 equivalent of WebResourceLoadDelegate::willSendRequest in the Bundle
https://bugs.webkit.org/show_bug.cgi?id=52897
<rdar://problem/8898294>
* WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
(WTR::InjectedBundlePage::InjectedBundlePage):
(WTR::InjectedBundlePage::willSendRequestForFrame):
* WebKitTestRunner/InjectedBundle/InjectedBundlePage.h:
2011-02-03 Ivan Krstić <ike@apple.com>
Unreviewed.
Adding myself to committers.py.
* Scripts/webkitpy/common/config/committers.py:
2011-02-03 James Kozianski <koz@chromium.org>
Reviewed by Dimitri Glazkov.
Add navigator.registerProtocolHandler behind a flag.
https://bugs.webkit.org/show_bug.cgi?id=52609
* Scripts/build-webkit:
2011-02-03 Hayato Ito <hayato@chromium.org>
Reviewed by Eric Seidel.
[NRWT] Remove TestArgs class, which is wrongly used.
https://bugs.webkit.org/show_bug.cgi?id=53063
* Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
* Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py:
* Scripts/webkitpy/layout_tests/test_types/image_diff.py:
* Scripts/webkitpy/layout_tests/test_types/test_type_base.py:
* Scripts/webkitpy/layout_tests/test_types/test_type_base_unittest.py:
* Scripts/webkitpy/layout_tests/test_types/text_diff.py:
2011-02-03 Andrew Wason <rectalogic@rectalogic.com>
Reviewed by Kenneth Russell.
Change ENABLE_3D_CANVAS to ENABLE_WEBGL
https://bugs.webkit.org/show_bug.cgi?id=53714
* Scripts/build-webkit: change --3d-canvas build option
to set ENABLE_WEBGL
2011-02-03 Hayato Ito <hayato@chromium.org>
Reviewed by Tony Chang.
[NRWT] Introduces Input/Output class used by base.Driver into port/base.py and
move _run_single_test() and _process_output() functions from
dump_render_tree_thread.py to a single_test_runner.py as an individual module.
This is clean up and acts as a preparation for elimination of test_types/*
classes. These classes will move to the single_test_runner module introduced
in this patch.
https://bugs.webkit.org/show_bug.cgi?id=53004
* Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
* Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py: Added.
* Scripts/webkitpy/layout_tests/layout_package/test_input.py:
* Scripts/webkitpy/layout_tests/layout_package/test_output.py: Removed.
* Scripts/webkitpy/layout_tests/port/base.py:
* Scripts/webkitpy/layout_tests/port/chromium.py:
* Scripts/webkitpy/layout_tests/port/dryrun.py:
* Scripts/webkitpy/layout_tests/port/test.py:
* Scripts/webkitpy/layout_tests/port/webkit.py:
* Scripts/webkitpy/layout_tests/test_types/image_diff.py:
* Scripts/webkitpy/layout_tests/test_types/test_type_base.py:
* Scripts/webkitpy/layout_tests/test_types/text_diff.py:
2011-02-03 Dirk Pranke <dpranke@chromium.org>
Unreviewed, build fix.
Delete a line that was accidentally missed in r77586; without it
there is still a race between the two threads.
https://bugs.webkit.org/show_bug.cgi?id=51572
* Scripts/webkitpy/layout_tests/layout_package/message_broker_unittest.py:
2011-02-03 Dirk Pranke <dpranke@chromium.org>
Reviewed by Tony Chang.
nrwt: handle "broken pipe" notifications from DRT more cleanly.
https://bugs.webkit.org/show_bug.cgi?id=52927
* Scripts/webkitpy/layout_tests/port/server_process.py:
* Scripts/webkitpy/layout_tests/port/server_process_unittest.py:
2011-02-03 Dirk Pranke <dpranke@chromium.org>
Reviewed by Kenneth Russell.
Update, resubmit change to bug 51572. I think the race that was
fixed in r76703 caused the hang that caused us to roll this
patch out before, so I'd like to try again.
https://bugs.webkit.org/show_bug.cgi?id=51572
* Scripts/webkitpy/layout_tests/layout_package/message_broker.py:
* Scripts/webkitpy/layout_tests/layout_package/message_broker_unittest.py:
2011-02-03 Daniel Cheng <dcheng@chromium.org>
Unreviewed.
Adding myself to committers.py.
* Scripts/webkitpy/common/config/committers.py:
2011-02-03 Adam Roben <aroben@apple.com>
Include the crashing function in the link to a crash log
Fixes <http://webkit.org/b/53739> Crash log links in results.html should include the
function that crashed
Reviewed by David Kilzer.
* Scripts/old-run-webkit-tests:
(crashLocation): Added. Returns the location of the crash.
(linksForErrorTest): Include the crash location in the link text for the crash log, if one
could be determined.
2011-02-03 Dirk Pranke <dpranke@chromium.org>
Unreviewed, build fix.
Actually make the changes suggested by Mihai in his review
of bug 53720 :(
https://bugs.webkit.org/show_bug.cgi?id=53720
* Scripts/webkitpy/layout_tests/port/test_files.py:
* Scripts/webkitpy/layout_tests/port/test_files_unittest.py:
2011-02-03 Dirk Pranke <dpranke@chromium.org>
Reviewed by Mihai Parparita.
new-run-webkit-tests: Fix bug introduced in r77434 that was
causing us to run the canvas GPU tests on the Mac GPU port. This
bug revealed that we were not setting the port.name() field
properly in many cases, so I've cleaned up all of that code, and
removed a few comments about "version-specific" GPU ports that
don't exist and just confused things.
Testing also revealed that port.abspath_for_test() wasn't
normalizing paths on Windows properly, so I fixed that as well.
https://bugs.webkit.org/show_bug.cgi?id=53719
* Scripts/webkitpy/layout_tests/port/base.py:
* Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
* Scripts/webkitpy/layout_tests/port/chromium_gpu_unittest.py:
2011-02-03 Dirk Pranke <dpranke@chromium.org>
Reviewed by Mihai Parparita.
new-run-webkit-tests: fix normalization of paths on windows when gathering files
r77434 introduced a bug that was causing the test expectations
for the GPU ports to not be treated properly. It turns out that
when we gathered the list of test files on Windows, we would
return paths of the form "c:\LayoutTests/fast/canvas", and the
mixture of backslashes and forward slashes was confusing things.
This patch normalizes all of the filenames returned from
test_files.find(), and adds better tests for this (fixing a
couple of other bugs found in the meantime).
https://bugs.webkit.org/show_bug.cgi?id=53720
* Scripts/webkitpy/layout_tests/port/test.py:
* Scripts/webkitpy/layout_tests/port/test_files.py:
* Scripts/webkitpy/layout_tests/port/test_files_unittest.py:
2011-02-03 Adam Roben <aroben@apple.com>
Add links to Windows crash logs in results.html
Fixes <http://webkit.org/b/53718> run-webkit-tests should link to Windows crash logs in
results.html
Reviewed by David Kilzer.
* Scripts/old-run-webkit-tests:
(top level): Added $crashLogTag and $windowsCrashLogFilePrefix (which came from
createDebuggerCommandFile).
(testCrashedOrTimedOut): If the test crashed, capture any saved crash log after the dump
tool has exited.
(captureSavedCrashLog): Added. Finds the crash log for the test that just crashed and moves
it into the test results directory tree. Only implemented for Cygwin currently.
(findNewestFileMatchingGlob): Added. Does what it says.
(htmlForResultsSection): Only link to files that exist.
(linksForErrorTest): Add a link to the crash log.
(deleteExpectedAndActualResults): Delete any old crash log for this test.
(createDebuggerCommandFile): Use the new $windowsCrashLogFilePrefix constant instead of
hardcoding it here.
2011-02-03 Adam Roben <aroben@apple.com>
Tell the debugger the path to the WebKit source tree when saving a crash log
This allows the debugger to include the crashing line of code in the log.
Fixes <http://webkit.org/b/53678> Crash logs from buildslaves don't show the crashing line
of code
Reviewed by Sam Weinig.
* Scripts/old-run-webkit-tests:
(createDebuggerCommandFile): Added. Saves commands that we'd like the debugger to run to a
file and returns the path to that file. The commands we pass came from
setUpWindowsCrashLogSaving, but I've added a .srcpath command to tell the debugger where the
WebKit source code lives.
(setUpWindowsCrashLogSaving): Instead of specifying the commands directly on the command
line using -c, save them to a file and specify the path to that file using -cf. This works
around what is presumably a bug in Windows's command line parsing, where having multiple
quoted paths in the debugger commands causes the post-mortem debugger not to be invoked at
all. Also pulled the options we pass to the debugger out into a list that is then join()ed
together to make them easier to modify in the future.
2011-02-02 Sam Weinig <sam@webkit.org>
Reviewed by Anders Carlsson.
Add notification of the end of a rubber band.
<rdar://problem/8940648>
* MiniBrowser/mac/BrowserWindowController.m:
(-[BrowserWindowController awakeFromNib]):
* WebKitTestRunner/TestController.cpp:
(WTR::TestController::createOtherPage):
(WTR::TestController::initialize):
Stub out the new WKUIClient function.
2011-02-03 Gabor Rapcsanyi <rgabor@webkit.org>
Unreviewed.
Adding myself to committers.py.
* Scripts/webkitpy/common/config/committers.py:
2011-02-03 Balazs Kelemen <kbalazs@webkit.org>
Reviewed by Csaba Osztrogonác.
[Qt][WK2] MiniBrowser's window size should be 800x600
https://bugs.webkit.org/show_bug.cgi?id=53670
Make layout test failure debugging easier.
* MiniBrowser/qt/BrowserWindow.cpp:
(BrowserWindow::BrowserWindow):
2011-02-02 Andy Estes <aestes@apple.com>
Reviewed by Mark Rowe.
'update-webkit' should handle OpenSource and Internal using different
VCSs.
* Scripts/update-webkit: Re-check the VCS type when updating Apple's
Internal directory.
2011-02-02 Dirk Pranke <dpranke@chromium.org>
Reviewed by Mihai Parparita.
new-run-webkit-tests: hardcode the directories to scan for tests
for GPU bots instead of SKIPping them. See bug 53562 for some
context, but we will now hard-code the list of directories to
use by default instead of skipping over directories in the
expectations file. We do this so that we will be able to merge
the expectations files without getting conflicting SKIP
directives. Note that this change will reduce the # of tests
being run on the Mac GPU bots, because we're accidentally
including some today.
https://bugs.webkit.org/show_bug.cgi?id=53631
* Scripts/webkitpy/layout_tests/port/base.py:
* Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
* Scripts/webkitpy/layout_tests/port/chromium_gpu_unittest.py:
2011-02-02 Sam Weinig <sam@webkit.org>
Reviewed by Beth Dakin.
Add ChromeClient function to paint custom overhang areas.
https://bugs.webkit.org/show_bug.cgi?id=53639
* WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
(WTR::InjectedBundlePage::InjectedBundlePage):
Stub out new callback.
2011-02-02 Dirk Pranke <dpranke@chromium.org>
Reviewed by Tony Chang.
new-run-webkit-tests: support chromium code paths in mock_drt
https://bugs.webkit.org/show_bug.cgi?id=53471
* Scripts/webkitpy/layout_tests/port/mock_drt.py:
* Scripts/webkitpy/layout_tests/port/mock_drt_unittest.py:
2011-02-02 Adam Klein <adamk@chromium.org>
Reviewed by Tony Chang.
Update new-run-webkit-tests --chromium to put output under Source/...
https://bugs.webkit.org/show_bug.cgi?id=53612
* BuildSlaveSupport/test-result-archive:
* Scripts/webkitpy/layout_tests/port/chromium.py:
2011-01-28 Tony Chang <tony@chromium.org>
Reviewed by Eric Seidel.
[chromium] remove --use-test-shell from NRWT
https://bugs.webkit.org/show_bug.cgi?id=53346
I've already started removing some of the bindings like eventSender so
this already doesn't work isn't useful. Also updated some comments to
refer to DRT instead of test_shell.
* Scripts/webkitpy/layout_tests/layout_package/test_expectations.py:
* Scripts/webkitpy/layout_tests/port/chromium.py:
* Scripts/webkitpy/layout_tests/port/chromium_linux.py:
* Scripts/webkitpy/layout_tests/port/chromium_mac.py:
* Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
* Scripts/webkitpy/layout_tests/port/chromium_win.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
* Scripts/webkitpy/layout_tests/test_types/text_diff.py:
2011-02-02 Steve Lacey <sjl@chromium.org>
Reviewed by Eric Carlson.
Implement basic media statistics on media elements.
https://bugs.webkit.org/show_bug.cgi?id=53322
* Scripts/build-webkit:
2011-02-02 Kevin Ollivier <kevino@theolliviers.com>
[wx] Build fixes for wxWebKit.
* DumpRenderTree/wx/LayoutTestControllerWx.cpp:
(LayoutTestController::setSerializeHTTPLoads):
* wx/browser/wscript:
* wx/build/build_utils.py:
* wx/build/settings.py:
2011-02-01 Dirk Pranke <dpranke@chromium.org>
Reviewed by Mihai Parparita.
new-run-webkit-tests: stop skipping so many tests
r77163 introduced a regression where we weren't resetting test
expectations properly and were skipping too many tests as a
result. This patch fixes that and adds a test for it.
https://bugs.webkit.org/show_bug.cgi?id=53551
* Scripts/webkitpy/layout_tests/layout_package/test_expectations.py:
* Scripts/webkitpy/layout_tests/layout_package/test_expectations_unittest.py:
2011-02-01 Dirk Pranke <dpranke@chromium.org>
Reviewed by Tony Chang.
new-run-webkit-tests: add first stub of test_runner2. This
will add support for the 'inline', 'threads', and 'processes'
flags to --worker-model, but for now the implementatios just
fall back on the old ones.
https://bugs.webkit.org/show_bug.cgi?id=53157
* Scripts/webkitpy/layout_tests/layout_package/test_runner2.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2011-02-01 Mihai Parparita <mihaip@chromium.org>
Reviewed by James Robinson.
[Chromium] run-chromium-webkit-unit-tests looks for binaries in the wrong place
https://bugs.webkit.org/show_bug.cgi?id=53522
* Scripts/run-chromium-webkit-unit-tests:
2011-02-01 Adam Barth <abarth@webkit.org>
Reviewed by Mihai Parparita.
Using Control-C to cancel webkit-patch or other python tools causes
unpleasant traceback console spew
https://bugs.webkit.org/show_bug.cgi?id=53462
Catch the exception and don't print the stack trace.
* Scripts/webkit-patch:
2011-02-01 Scott Cameron <sccameron@rim.com>
Reviewed by Daniel Bates.
Use Windows format for MinGW HTTPD path.
https://bugs.webkit.org/show_bug.cgi?id=53503
* Scripts/webkitperl/httpd.pm:
- Use single-quotes around MySys value for $httpdPath in getHTTPDPath()
so that we don't have to escape the space characters in the path. Also,
changed path to Windows-style path and removed FIXME comment.
2011-02-01 Balazs Kelemen <kbalazs@webkit.org>
Reviewed by Andreas Kling.
[Qt][WK2] Add a way to use shared process model in MiniBrowser
https://bugs.webkit.org/show_bug.cgi?id=53090
* MiniBrowser/qt/BrowserView.cpp:
(BrowserView::BrowserView): Removed the m_context member.
From now the context is guaranteed to be non-null and we
don't need to store that in the object.
* MiniBrowser/qt/BrowserView.h:
* MiniBrowser/qt/BrowserWindow.cpp:
Added static bool to determine that new windows need to be
created with their own context or not. Use the same context
and web process by default to be inilne with the other ports.
(BrowserWindow::BrowserWindow):
(BrowserWindow::newWindow):
* MiniBrowser/qt/BrowserWindow.h:
* MiniBrowser/qt/main.cpp:
(main): Added command line switch to be able to use the
non-shared process model. Simplify the handling of the command line
switches a little bit.
2011-02-01 Zoltan Horvath <zoltan@webkit.org>
Reviewed by Andreas Kling.
[Qt] Add Window menu and Toggle FullScreen action to MiniBrowser.
https://bugs.webkit.org/show_bug.cgi?id=53491
* MiniBrowser/qt/BrowserWindow.cpp:
(BrowserWindow::BrowserWindow):
(BrowserWindow::toggleFullScreenMode):
* MiniBrowser/qt/BrowserWindow.h:
2011-02-01 Zoltan Horvath <zoltan@webkit.org>
Reviewed by Andreas Kling.
[Qt] Add Toggle Frame Flattening, Auto Load Images and Disable JavaScript actions to MiniBrowser.
https://bugs.webkit.org/show_bug.cgi?id=53489
Add Toggle Frame Flattening action to Develop menu. Add Auto Load Images and Disable JavaScript actions
to Settings menu.
* MiniBrowser/qt/BrowserWindow.cpp:
(BrowserWindow::BrowserWindow):
(BrowserWindow::toggleFrameFlattening):
(BrowserWindow::toggleDisableJavaScript):
(BrowserWindow::toggleAutoLoadImages):
* MiniBrowser/qt/BrowserWindow.h:
2011-01-31 Mihai Parparita <mihaip@chromium.org>
Reviewed by Tony Chang.
[Chromium] Switch chromium-mac to use ChromiumDriver
https://bugs.webkit.org/show_bug.cgi?id=53461
For the sake of consistency with the other Chromium platforms (and so
that chromium-mac picks up special flags like --enable-hardware-gpu),
switch chromium-mac from the WebKitDriver to ChromiumDriver.
* Scripts/webkitpy/layout_tests/port/chromium.py:
2011-01-31 Dirk Pranke <dpranke@chromium.org>
Reviewed by Mihai Parparita.
The current modifier parsing code in test_expectations is
fragile and hard-coded, so it's not easy to understand the logic
or easily add new types of modifiers (like GPU vs. CPU testing
for graphics tests, or 32-bit vs. 64-bit differences).
This is the first of two patches that will add in more generic
support and then eliminate the GPU-specific test expectations
files for Chromium.
This patch adds two standalone objects for handling modifiers. The
rules for interpreting modifiers, precedence, and conflicts are
given in the docstring to the ModifierMatcher class, which
returns ModifierMatchResult objects.
This patch also adds routines to the Port interface and a
default set of values in the base object, in order to obtain the
values needed on a given test run. These values are then passed
to the expectation parser. This also allows us to clean up the
logic used to lint all of the different configurations in a
single test_expectations.txt file.
The next patch will merge in the separate GPU expectations file.
https://bugs.webkit.org/show_bug.cgi?id=51222
* Scripts/webkitpy/layout_tests/layout_package/printing_unittest.py:
* Scripts/webkitpy/layout_tests/layout_package/test_expectations.py:
* Scripts/webkitpy/layout_tests/layout_package/test_expectations_unittest.py:
* Scripts/webkitpy/layout_tests/layout_package/test_runner.py:
* Scripts/webkitpy/layout_tests/port/base.py:
* Scripts/webkitpy/layout_tests/port/base_unittest.py:
* Scripts/webkitpy/layout_tests/port/chromium.py:
* Scripts/webkitpy/layout_tests/port/port_testcase.py:
* Scripts/webkitpy/layout_tests/port/test.py:
* Scripts/webkitpy/layout_tests/port/webkit.py:
* Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
* Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests_unittest.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
* Tools/Scripts/webkitpy/style/checkers/test_expectations.py:
* Tools/Scripts/webkitpy/style/checkers/test_expectations_unittest.py:
2011-01-31 Scott Cameron <sccameron@rim.com>
Reviewed by Daniel Bates.
update run-webkit-httpd to be able to launch Apache2.2 in a MSYS environment
https://bugs.webkit.org/show_bug.cgi?id=50036
This will add an Apache2.2 configuration file and modifies scripts to allow
running an httpd server in an MSYS environment.
The default Apache2.2 installation path has been preserved and PHP5 modules disabled
in order to allow for simple installation/execution. Simply install the latest
Apache2.2 version with OpenSSL from http://httpd.apache.org/download.cgi#apache22 to
be able to execute run-webkit-httpd.
* Scripts/run-webkit-httpd:
* Scripts/webkitperl/httpd.pm:
2011-01-31 Tony Chang <tony@chromium.org>
Reviewed by Eric Seidel.
mac DRT should report RTL scroll offset relative to top right corner
https://bugs.webkit.org/show_bug.cgi?id=53324
* DumpRenderTree/mac/DumpRenderTree.mm:
(dumpFrameScrollPosition):
2011-01-31 Zoltan Horvath <zoltan@webkit.org>
Reviewed by Andreas Kling.
[Qt] Add Take Screen Shot action to MiniBrowser
https://bugs.webkit.org/show_bug.cgi?id=53422
Add Take Screen Shot action to MiniBrowser's view menu.
* MiniBrowser/qt/BrowserWindow.cpp:
(BrowserWindow::BrowserWindow):
(BrowserWindow::screenshot):
* MiniBrowser/qt/BrowserWindow.h:
2011-01-31 Gustavo Noronha Silva <gustavo.noronha@collabora.co.uk>
Reviewed by Xan Lopez.
[GTK] REGRESSION: http/tests/media/video-{cookie,referer}.html failing
https://bugs.webkit.org/show_bug.cgi?id=53379
Remove left over #ifdef's. I thought all of the conditional code
had been freed from the condition, but I forgot to check DRT.
* DumpRenderTree/gtk/DumpRenderTree.cpp:
(resetDefaultsToConsistentValues):
* DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
(LayoutTestController::setAlwaysAcceptCookies):
2011-01-30 Dirk Pranke <dpranke@chromium.org>
Unreviewed, build fix.
Fix regression introduced in r77093 - path.rsplit() doesn't
take keyword arguments.
https://bugs.webkit.org/show_bug.cgi?id=53326
* Scripts/webkitpy/common/system/filesystem_mock.py:
2011-01-30 Dirk Pranke <dpranke@chromium.org>
Reviewed by Mihai Parparita.
Add more unit tests for rebaseline-chromium-webkit-tests. This
change involves restructuring a bunch of r-c-w-t code to make it
more testable as well. We also add wrapper classes for handling
testing zip files and fetching URLs.
https://bugs.webkit.org/show_bug.cgi?id=53040
* Scripts/webkitpy/common/system/urlfetcher.py:
* Scripts/webkitpy/common/system/urlfetcher_mock.py:
* Scripts/webkitpy/common/system/zipfileset_mock.py:
* Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
* Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests_unittest.py:
* Scripts/webkitpy/tool/mocktool.py:
2011-01-30 Dirk Pranke <dpranke@chromium.org>
Reviewed by Eric Seidel.
Clean up of the filesystem-related modules used in webkitpy.
I've added relpath() to the filesystem interface, modified
ospath.relpath() so that it could work with the filesystem
interface, and modified the fileset* routines to use the
filesystem interface consistently.
This patch also adds a close() routine to the fileset routines
to indicate that the caller is done with the fileset. This
allows zipfileset to clean up after itself when it creates
tempfiles to store downloads.
https://bugs.webkit.org/show_bug.cgi?id=53326
* Scripts/webkitpy/common/system/directoryfileset.py:
* Scripts/webkitpy/common/system/fileset.py:
* Scripts/webkitpy/common/system/filesystem.py:
* Scripts/webkitpy/common/system/filesystem_mock.py:
* Scripts/webkitpy/common/system/filesystem_unittest.py:
* Scripts/webkitpy/common/system/ospath.py:
* Scripts/webkitpy/common/system/zipfileset.py:
* Scripts/webkitpy/common/system/zipfileset_unittest.py:
2011-01-30 Balazs Kelemen <kbalazs@webkit.org>
Reviewed by Csaba Osztrogonác.
[Qt][WK2]REGRESSION (r76991): Fix build errors
https://bugs.webkit.org/show_bug.cgi?id=53400
Revert the temporary build fix (http://trac.webkit.org/changeset/77088)
and remove WebKit2Prefix.h from the build.
* MiniBrowser/qt/MiniBrowser.pro:
2011-01-30 Csaba Osztrogonác <ossy@webkit.org>
Unreviewed.
[Qt][WK2] Buildfix.
* MiniBrowser/qt/MiniBrowser.pro:
2011-01-28 Simon Fraser <simon.fraser@apple.com>
Reviewed by Adam Roben.
define NOMINMAX when building ImageDiff, as we do elsewhere.
Add various clampToInt() methods to MathExtras.h
https://bugs.webkit.org/show_bug.cgi?id=52910
* DumpRenderTree/win/ImageDiffCommon.vsprops:
2011-01-28 Maciej Stachowiak <mjs@apple.com>
Reviewed by Darin Adler.
WebKitTestRunner needs layoutTestController.setPOSIXLocale
https://bugs.webkit.org/show_bug.cgi?id=42682
* WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl:
* WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
(WTR::InjectedBundle::resetLocalSettings):
(WTR::InjectedBundle::didReceiveMessage):
* WebKitTestRunner/InjectedBundle/InjectedBundle.h:
* WebKitTestRunner/InjectedBundle/LayoutTestController.cpp:
(WTR::LayoutTestController::setPOSIXLocale):
* WebKitTestRunner/InjectedBundle/LayoutTestController.h:
2011-01-28 Dirk Pranke <dpranke@chromium.org>
Reviewed by Mihai Parparita.
test-webkitpy: fix webkitpy.layout_tests.port.mac_unittest.MacTest.test_skipped_file_paths
This patch re-enables this test and changes it to
handle all of the mac platform versions, not just the one
it is running on.
https://bugs.webkit.org/show_bug.cgi?id=53356
* Scripts/webkitpy/layout_tests/port/mac_unittest.py:
2011-01-28 Dirk Pranke <dpranke@chromium.org>
Unreviewed, build fix.
Take two. The fix in 77023 didn't work, because we were
still calling path.abspath_to_uri, which calls _cygpath under
the covers, and it appears the cygpath on the bots does
something different than it does on my machine. This patch
removes the calls to path.abspath_to_uri, so it should be safe.
If it doesn't work, I'll roll it out along with r76982 and 77023.
https://bugs.webkit.org/show_bug.cgi?id=53126
* Scripts/webkitpy/layout_tests/port/test.py:
2011-01-28 David Kilzer <ddkilzer@apple.com>
<rdar://problem/8930699> build-webkit gives a bogus warning with newer versions of Xcode
<http://webkit.org/b/53353>
Reviewed by Mark Rowe.
* Scripts/webkitdirs.pm:
(checkRequiredSystemConfig): Check the Xcode marketing version
in addition to the DevCoreTools build version before complaining
about an old version of Xcode. Also make the Mac OS X version
check use Perl's built-in version string comparitor.
2011-01-28 Dirk Pranke <dpranke@chromium.org>
Unreviewed, build fix.
Work around breakage on Win 7 Release bot caused by r76982
and the fact that windows ports use "file:////" instead of
"file:///". Ideally the test code should be isolated from
this, but it isn't yet. Will fix properly in a bit.
* Scripts/webkitpy/layout_tests/port/mock_drt_unittest.py:
2011-01-28 Dirk Pranke <dpranke@chromium.org>
Reviewed by Tony Chang.
new-run-webkit-tests: change worker model values to
"old-inline", "old-threads" in preparation for test_runner2 /
multiprocessing changes.
https://bugs.webkit.org/show_bug.cgi?id=53156
* Scripts/webkitpy/layout_tests/layout_package/test_runner.py:
* Scripts/webkitpy/layout_tests/port/base.py:
* Scripts/webkitpy/layout_tests/port/chromium_mac.py:
* Scripts/webkitpy/layout_tests/port/mac.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2011-01-28 Dirk Pranke <dpranke@chromium.org>
Reviewed by Tony Chang.
committers.py - add an IRC nickname for dpranke
https://bugs.webkit.org/show_bug.cgi?id=53335
* Scripts/webkitpy/common/config/committers.py:
2011-01-28 Dirk Pranke <dpranke@chromium.org>
Reviewed by Tony Chang.
new-run-webkit-tests: add a "mock DRT" port implementation
and a separate class that emulates what we expect the
DumpRenderTree behavior to be.
This will eventually replace port/dryrun.py and allow us to get
better test coverage of the new-run-webkit-tests code as well as
a reference for what new-run-webkit-tests expects from DRT.
This is the first attempt at this, and it is pretty bare-boned. It
really only has been tested on the 'mac' port (and a little on
the 'chromium-mac' port.
https://bugs.webkit.org/show_bug.cgi?id=53126
* Scripts/webkitpy/common/system/filesystem_mock.py:
* Scripts/webkitpy/layout_tests/port/dryrun.py:
* Scripts/webkitpy/layout_tests/port/factory.py:
* Scripts/webkitpy/layout_tests/port/mock_drt.py: Added.
* Scripts/webkitpy/layout_tests/port/mock_drt_unittest.py: Added.
* Scripts/webkitpy/layout_tests/port/test.py:
2011-01-28 Pratik Solanki <psolanki@apple.com>
Unreviewed. Removing .swp file checked in by mistake.
* Scripts/.webkitdirs.pm.swp: Removed.
2011-01-28 Pratik Solanki <psolanki@apple.com>
Reviewed by David Kilzer.
Space not necessary for undefined feature
https://bugs.webkit.org/show_bug.cgi?id=53317
* Scripts/build-webkit:
2011-01-28 Chang Shu <chang.shu@nokia.com>
Reviewed by Andreas Kling.
Reset NetworkAccessManager to clean up credentials from previous tests.
This change causes no performance overhead either.
https://bugs.webkit.org/show_bug.cgi?id=36688
* DumpRenderTree/qt/DumpRenderTreeQt.cpp:
(WebCore::DumpRenderTree::resetToConsistentStateBeforeTesting):
(WebCore::DumpRenderTree::open):
* DumpRenderTree/qt/DumpRenderTreeQt.h:
2011-01-28 Zoltan Horvath <zoltan@webkit.org>
Reviewed by Andreas Kling.
[Qt] Add Zoom Text Only action to MiniBrowser
https://bugs.webkit.org/show_bug.cgi?id=53297
Add Zoom Text Only action to MiniBrowser's view menu, change zoom levels to avoid unnecessary divides.
* MiniBrowser/qt/BrowserWindow.cpp:
(BrowserWindow::BrowserWindow):
(BrowserWindow::zoomIn):
(BrowserWindow::zoomOut):
(BrowserWindow::resetZoom):
(BrowserWindow::toggleZoomTextOnly):
(BrowserWindow::applyZoom):
* MiniBrowser/qt/BrowserWindow.h:
2011-01-28 Adam Barth <abarth@webkit.org>
Reviewed by Eric Seidel.
Add webkit-patch roll-chromium-deps
https://bugs.webkit.org/show_bug.cgi?id=53288
This command updates the Source/WebKit/chromium/DEPS file with the
last-known good revision of Chromium (or a revision specified on the
command line). I'd eventually like to turn this into a SheriffBot
command, but this is the first step.
This patch somewhat sprawled because I needed to move a bunch of code
out of ChangeLog that should never have been there in the first place.
Also, I had to fix a bug in MockUser in order to test the new command.
* Scripts/webkitpy/common/checkout/api.py:
* Scripts/webkitpy/common/checkout/changelog.py:
* Scripts/webkitpy/common/checkout/changelog_unittest.py:
* Scripts/webkitpy/common/checkout/deps.py: Added.
* Scripts/webkitpy/common/config/urls.py:
* Scripts/webkitpy/tool/commands/__init__.py:
* Scripts/webkitpy/tool/commands/download_unittest.py:
* Scripts/webkitpy/tool/commands/roll.py: Added.
* Scripts/webkitpy/tool/commands/roll_unittest.py: Added.
* Scripts/webkitpy/tool/commands/upload_unittest.py:
* Scripts/webkitpy/tool/mocktool.py:
* Scripts/webkitpy/tool/steps/__init__.py:
* Scripts/webkitpy/tool/steps/preparechangelogfordepsroll.py: Added.
* Scripts/webkitpy/tool/steps/preparechangelogforrevert.py:
* Scripts/webkitpy/tool/steps/preparechangelogforrevert_unittest.py: Added.
* Scripts/webkitpy/tool/steps/suggestreviewers_unittest.py:
* Scripts/webkitpy/tool/steps/updatechromiumdeps.py: Added.
* Scripts/webkitpy/tool/steps/validatechangelogs_unittest.py:
2011-01-27 Greg Coletta <greg.coletta@nokia.com>
Reviewed by Laszlo Gombos.
Get rid of prefix header dependency for WebKit2 build system
https://bugs.webkit.org/show_bug.cgi?id=50174
Change the style checker so that it enforces config.h include for
WebKit2.
* Scripts/webkitpy/style/checker.py:
2011-01-27 Maciej Stachowiak <mjs@apple.com>
Reviewed by Darin Adler.
execCommand("Paste") doesn't work in WebKitTestRunner
https://bugs.webkit.org/show_bug.cgi?id=52785
Enable both of the settings needed to allow paste.
* WebKitTestRunner/TestController.cpp:
(WTR::TestController::resetStateToConsistentValues):
2011-01-27 Jacob Dinu <dinu.jacob@nokia.com>
Reviewed by Andreas Kling.
[Qt] QtTestBrowser: User input lost when toggling use of QGraphicsView
https://bugs.webkit.org/show_bug.cgi?id=48440
Preserve user input when toggling use of QGraphicsView
* QtTestBrowser/launcherwindow.cpp:
(LauncherWindow::initializeView):
(LauncherWindow::loadFinished):
* QtTestBrowser/launcherwindow.h:
* QtTestBrowser/mainwindow.cpp:
(MainWindow::addressUrl):
* QtTestBrowser/mainwindow.h:
2011-01-27 Dirk Pranke <dpranke@chromium.org>
Reviewed by Mihai Parparita.
new-run-webkit-tests: turn off pixel tests correctly by default
for webkit-based ports. r70013 (bug 47510) used
port.set_option_default() to attempt to set default values, but
that didn't work correctly. I have removed set_option_default
for now since it was only being used in two places and in three
useless unit tests. There is a separate bug open to fix the
option parsing (48095), so this workaround is fine for now.
https://bugs.webkit.org/show_bug.cgi?id=53217
* Scripts/webkitpy/layout_tests/port/webkit.py:
* Scripts/webkitpy/layout_tests/port/base.py:
* Scripts/webkitpy/layout_tests/port/base_unittest.py:
2011-01-27 Balazs Kelemen <kbalazs@webkit.org>
Rubber-stamped by Ariya Hidayat.
Trivial crash fix in WTR.
* WebKitTestRunner/TestController.cpp:
(WTR::TestController::initialize): Increase the size
of the vector by one to be able to store the null character.
2011-01-27 Balazs Kelemen <kbalazs@webkit.org>
Reviewed by Csaba Osztrogonác.
[Qt][WK2] WTR should be initialized in the same way as DRT
https://bugs.webkit.org/show_bug.cgi?id=53240
* WebKitTestRunner/InjectedBundle/qt/ActivateFontsQt.cpp:
(WTR::activateFonts): Added missing initialization steps.
2011-01-27 Joone Hur <joone.hur@collabora.co.uk>
Unreviewed.
Adding myself to committers.py.
* Scripts/webkitpy/common/config/committers.py:
2011-01-27 Laszlo Gombos <laszlo.1.gombos@nokia.com>
Reviewed by Csaba Osztrogonác.
[Qt] [Symbian] Move project files into Source
https://bugs.webkit.org/show_bug.cgi?id=52891
Fix the Symbian build after the project files
are moved to Source directory. On Symbian qmake
needs to run in the same directory where the main
pro files (WebKit.pro, Tools.pro) are located.
* Scripts/webkitdirs.pm: Change to the directory where the pro file
is located before running qmake for Symbian. Qmake on Symbian
does not properly honor the "-o" option, work it around by setting
the name of the Makefile to bld.inf.
2011-01-27 Zoltan Horvath <zoltan@webkit.org>
Reviewed by Andreas Kling.
[Qt] Add simple zooming features to MiniBrowser
https://bugs.webkit.org/show_bug.cgi?id=53231
Add View menu and Zoom In, Zoom Out, Zoom Reset actions to MiniBrowser.
* MiniBrowser/qt/BrowserWindow.cpp:
(BrowserWindow::BrowserWindow):
(BrowserWindow::zoomIn):
(BrowserWindow::zoomOut):
(BrowserWindow::resetZoom):
(BrowserWindow::updateUserAgentList):
(BrowserWindow::applyZoom):
* MiniBrowser/qt/BrowserWindow.h:
2011-01-26 Maciej Stachowiak <mjs@apple.com>
Reviewed by Dan Bernstein.
WebKitTestRunner needs to support layoutTestController.evaluateInWebInspector
https://bugs.webkit.org/show_bug.cgi?id=42319
Add evaluateInWebInspector and other APIs needed for inspector tests to run.
* WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl:
* WebKitTestRunner/InjectedBundle/LayoutTestController.cpp:
(WTR::LayoutTestController::showWebInspector):
(WTR::LayoutTestController::closeWebInspector):
(WTR::LayoutTestController::evaluateInWebInspector):
(WTR::LayoutTestController::setTimelineProfilingEnabled):
* WebKitTestRunner/InjectedBundle/LayoutTestController.h:
* WebKitTestRunner/TestInvocation.cpp:
(WTR::shouldOpenWebInspector):
(WTR::TestInvocation::invoke):
* WebKitTestRunner/WebKitTestRunnerPrefix.h:
2011-01-26 Martin Robinson <mrobinson@igalia.com>
Reviewed by Xan Lopez.
[GTK] Pixel dumps do not include scrollbars in output images
https://bugs.webkit.org/show_bug.cgi?id=53216
Show scrollbars in pixel dumps. This will fix pixel dumps for tests
that have scrollbars.
* DumpRenderTree/gtk/PixelDumpSupportGtk.cpp:
(createBitmapContextFromWebView): Take the snapshot of the containing GtkScrolledWindow
instead of the web view itself.
2011-01-26 Dirk Pranke <dpranke@chromium.org>
Reviewed by Mihai Parparita.
Fix regression introduced in r76322 ... new-run-webkit-tests
was attempting to save the image diff output as a UTF-8-encoded
file.
https://bugs.webkit.org/show_bug.cgi?id=53210
* Scripts/webkitpy/layout_tests/port/webkit.py:
2011-01-26 Sheriff Bot <webkit.review.bot@gmail.com>
Unreviewed, rolling out r76709.
http://trac.webkit.org/changeset/76709
https://bugs.webkit.org/show_bug.cgi?id=53194
"broke python tests on non-snowleopard platforms" (Requested
by dpranke on #webkit).
* Scripts/webkitpy/layout_tests/port/factory.py:
* Scripts/webkitpy/layout_tests/port/mock_drt.py: Removed.
* Scripts/webkitpy/layout_tests/port/mock_drt_unittest.py: Removed.
2011-01-26 Dirk Pranke <dpranke@chromium.org>
Reviewed by Tony Chang.
new-run-webkit-tests: add a "mock DRT" port implementation
and a separate class that emulates what we expect the
DumpRenderTree behavior to be.
This will eventually replace port/dryrun.py and allow us to get
better test coverage of the new-run-webkit-tests code as well as
a reference for what new-run-webkit-tests expects from DRT.
This is the first attempt at this, and it is pretty bare-boned. It
really only has been tested on the 'mac' port (and a little on
the 'chromium-mac' port.
https://bugs.webkit.org/show_bug.cgi?id=53126
* Scripts/webkitpy/layout_tests/port/mock_drt.py: Added.
* Scripts/webkitpy/layout_tests/port/mock_drt_unittest.py: Added.
* Scripts/webkitpy/layout_tests/port/factory.py:
2011-01-26 Xianzhu Wang <wangxianzhu@google.com>
Reviewed by Tony Chang.
Add '--no-timeout' option to Chromium DRT to ease debugging.
https://bugs.webkit.org/show_bug.cgi?id=52873
* DumpRenderTree/chromium/DumpRenderTree.cpp:
(main):
2011-01-26 Csaba Osztrogonác <ossy@webkit.org>
Reviewed by Andreas Kling.
[Qt] Fix qt_minimal build
https://bugs.webkit.org/show_bug.cgi?id=53172
* QtTestBrowser/locationedit.h:
2011-01-26 Mansi Mithal <mansi.mithal@nokia.com>
Reviewed by Antonio Gomes.
QtTestBrowser should have a UI Setting to disable plugins
https://bugs.webkit.org/show_bug.cgi?id=52408
Added a new action item named "Disable Plugins"
under the "Settings" menu.
* QtTestBrowser/launcherwindow.cpp:
(LauncherWindow::createChrome):
(LauncherWindow::togglePlugins):
* QtTestBrowser/launcherwindow.h:
2011-01-25 Patrick Gansterer <paroga@webkit.org>
Reviewed by Adam Barth.
Move main CMakeLists.txt into Source directory
https://bugs.webkit.org/show_bug.cgi?id=52888
* Scripts/webkitdirs.pm:
2011-01-25 Patrick Gansterer <paroga@webkit.org>
Reviewed by Eric Seidel.
Improve label text in submit-to-ews
https://bugs.webkit.org/show_bug.cgi?id=53130
* QueueStatusServer/templates/submittoews.html:
2011-01-25 Mansi Mithal <mansi.mithal@nokia.com>
Reviewed by Antonio Gomes.
QtTestBrowser should have a UI Settings to prevent loading images
https://bugs.webkit.org/show_bug.cgi?id=52409
Added a new action item named "DisableAutoLoadImages"
under the "Settings" menu
* QtTestBrowser/launcherwindow.cpp:
(LauncherWindow::createChrome):
(LauncherWindow::toggleAutoLoadImages):
* QtTestBrowser/launcherwindow.h:
2011-01-25 Dirk Pranke <dpranke@chromium.org>
Reviewed by Tony Chang.
Minor bug fixes and cleanup for filesystem wrappers, port/* test
classes, test_expectations.py. This change adds "test-win" and
"test-mac" variants to the test port so that we can better test
rebaselining, and adds a MockUser() object for reuse in testing.
https://bugs.webkit.org/show_bug.cgi?id=53036
* Scripts/webkitpy/common/system/filesystem.py:
* Scripts/webkitpy/common/system/filesystem_mock.py:
* Scripts/webkitpy/layout_tests/layout_package/test_expectations.py:
* Scripts/webkitpy/layout_tests/port/factory.py:
* Scripts/webkitpy/layout_tests/port/base.py:
* Scripts/webkitpy/layout_tests/port/test.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
* Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests_unittest.py:
* Scripts/webkitpy/tool/mocktool.py:
2011-01-25 Tony Chang <tony@chromium.org>
Reviewed by Dimitri Glazkov.
[chromium] remove unused layoutTestController callback
https://bugs.webkit.org/show_bug.cgi?id=53103
This method is used by a test_shell_test, but not needed by DRT.
* DumpRenderTree/chromium/LayoutTestController.cpp:
(LayoutTestController::LayoutTestController):
* DumpRenderTree/chromium/LayoutTestController.h:
2011-01-25 Gustavo Noronha Silva <gustavo.noronha@collabora.co.uk>
Reviewed by Andreas Kling.
[GTK] Remove 64 bits release bot from the buildbot master
https://bugs.webkit.org/show_bug.cgi?id=52899
Removes the 64 bits release slave from the buildbot configuration,
so that we can use it as an EWS.
* BuildSlaveSupport/build.webkit.org-config/config.json:
2011-01-24 Maciej Stachowiak <mjs@apple.com>
Reviewed by Dan Bernstein.
Implement database quota callback to fix storage tests on WebKit2 bot
https://bugs.webkit.org/show_bug.cgi?id=53064
* WebKitTestRunner/TestController.cpp:
(WTR::exceededDatabaseQuota):
(WTR::TestController::createOtherPage):
(WTR::TestController::initialize):
2011-01-24 Maciej Stachowiak <mjs@apple.com>
Reviewed by Anders Carlsson.
Use designated temp directory for the database for WebKit2
https://bugs.webkit.org/show_bug.cgi?id=53052
Adopt the new WK2 API for this.
* WebKitTestRunner/TestController.cpp:
(WTR::TestController::libraryPathForTesting):
(WTR::TestController::initialize):
* WebKitTestRunner/TestController.h:
* WebKitTestRunner/mac/TestControllerMac.mm:
(WTR::TestController::platformLibraryPathForTesting):
* WebKitTestRunner/qt/TestControllerQt.cpp:
(WTR::TestController::platformLibraryPathForTesting):
* WebKitTestRunner/win/TestControllerWin.cpp:
(WTR::TestController::platformLibraryPathForTesting):
2011-01-24 Lucas Forschler <lforschler@apple.com>
Reviewed by Stephanie Lewis.
Add a new Leopard test bot.
* BuildSlaveSupport/build.webkit.org-config/config.json:
2011-01-24 Gyuyoung Kim <gyuyoung.kim@samsung.com>
Reviewed by Adam Barth.
Add EFL Builder to core waterfall
https://bugs.webkit.org/show_bug.cgi?id=52704
Add EFL Builder to the core waterfall again.
* Scripts/webkitpy/common/net/buildbot/buildbot.py:
* Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
2011-01-24 Xianzhu Wang <wangxianzhu@google.com>
Reviewed by Tony Chang.
new-run-webkit-tests fails to start HTTP server if there are proxy settings
https://bugs.webkit.org/show_bug.cgi?id=52872
Force disabling proxy to resolve the problem.
* Scripts/webkitpy/layout_tests/port/http_server_base.py:
* Scripts/webkitpy/layout_tests/port/websocket_server.py:
2011-01-24 Andreas Kling <kling@webkit.org>
Reviewed by Darin Adler.
check-webkit-style: Update exemptions following Source/ move.
https://bugs.webkit.org/show_bug.cgi?id=53017
* Scripts/webkitpy/style/checker.py:
* Scripts/webkitpy/style/checker_unittest.py:
2011-01-24 Andras Becsi <abecsi@webkit.org>
Rubber-stamped by Csaba Osztrogonác.
[Qt] Move project files into Source
https://bugs.webkit.org/show_bug.cgi?id=52891
* DumpRenderTree/qt/ImageDiff.pro: Add missing "Source" to WebKit.pri path
2011-01-24 Mikhail Naganov <mnaganov@chromium.org>
Reviewed by Yury Semikhatsky.
Web Inspector: Find duplicate strings in localizedStrings.js
https://bugs.webkit.org/show_bug.cgi?id=53005
* Scripts/check-inspector-strings:
2011-01-24 Zoltan Horvath <zoltan@webkit.org>
Reviewed by Andreas Kling.
[Qt] Regroup MiniBrowser's menubar and add open file action
https://bugs.webkit.org/show_bug.cgi?id=53000
Add File and Develop menu to MiniBrowser's menubar and add open file action to File menu.
* MiniBrowser/qt/BrowserWindow.cpp:
(BrowserWindow::BrowserWindow):
(BrowserWindow::openFile):
(BrowserWindow::~BrowserWindow):
* MiniBrowser/qt/BrowserWindow.h:
2011-01-24 Csaba Osztrogonác <ossy@webkit.org>
Unreviewed fix after r76496
[Qt] Move project files into Source
https://bugs.webkit.org/show_bug.cgi?id=52891
* BuildSlaveSupport/build.webkit.org-config/master.cfg: Remove unnecessary "Source" from the path.
* BuildSlaveSupport/built-product-archive: Remove unnecessary "Source" from the path.
* Scripts/webkitdirs.pm: Add the accidentally removed slash to the path.
2011-01-24 Andras Becsi <abecsi@webkit.org>
Reviewed by Csaba Osztrogonác.
[Qt] Move project files into Source
https://bugs.webkit.org/show_bug.cgi?id=52891
* DerivedSources.pro: Renamed from DerivedSources.pro.
* DumpRenderTree/qt/DumpRenderTree.pro:
* DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro:
* MiniBrowser/DerivedSources.pro:
* MiniBrowser/qt/MiniBrowser.pro:
* QtTestBrowser/QtTestBrowser.pro:
* Scripts/webkitdirs.pm:
* Tools.pro: Added.
* WebKitTestRunner/DerivedSources.pro:
* WebKitTestRunner/InjectedBundle/qt/InjectedBundle.pro:
* WebKitTestRunner/qt/WebKitTestRunner.pro:
2011-01-22 Geoffrey Garen <ggaren@apple.com>
Reviewed by Dan Bernstein.
Beefed up --threaded mode to catch even more kinds of errors.
https://bugs.webkit.org/show_bug.cgi?id=52971
* DumpRenderTree/pthreads/JavaScriptThreadingPthreads.cpp: Use a shared
context group to force JSC to mark multiple threads. (This used to be
the default, but it changed in SnowLeopard.)
(runJavaScriptThread): Do more locking and unlocking, and more allocation,
to give threading mistakes more chances to show themselves.
(startJavaScriptThreads):
(stopJavaScriptThreads):
2011-01-22 Robert Hogan <robert@webkit.org>
Reviewed by Andreas Kling.
[Qt] Support layoutTestController.addURLToRedirect()
https://bugs.webkit.org/show_bug.cgi?id=52956
* DumpRenderTree/qt/LayoutTestControllerQt.cpp:
(LayoutTestController::addURLToRedirect):
* DumpRenderTree/qt/LayoutTestControllerQt.h:
2011-01-22 Robert Hogan <robert@webkit.org>
Reviewed by Kenneth Rohde Christiansen.
[Qt] REGRESSION (r72360): http/tests/security/aboutBlank/xss-DENIED-* tests fail
https://bugs.webkit.org/show_bug.cgi?id=49802
* DumpRenderTree/qt/DumpRenderTreeQt.cpp:
(WebCore::DumpRenderTree::resetToConsistentStateBeforeTesting):
(WebCore::DumpRenderTree::open):
2011-01-21 Adam Barth <abarth@webkit.org>
Reviewed by Maciej Stachowiak.
Disable mac_unittest.py that fails on some of the Mac bots
https://bugs.webkit.org/show_bug.cgi?id=52947
As noted in the comment, this test does not appear to be correct
because the function this test is testing returns different results
depending on which flavor of Mac the test is run on. This patch
disable the test. Hopefully we can re-enable the test once we've
probably insulated the test from its environment.
* Scripts/webkitpy/layout_tests/port/mac_unittest.py:
2011-01-21 Dirk Pranke <dpranke@chromium.org>
Reviewed by Mihai Parparita.
Fix bug introduced in r76322 that caused NRWT to not actually
read the Skipped files properly.
https://bugs.webkit.org/show_bug.cgi?id=52771
* Scripts/webkitpy/layout_tests/port/mac_unittest.py:
* Scripts/webkitpy/layout_tests/port/webkit.py:
* Scripts/webkitpy/layout_tests/port/webkit_unittest.py:
2011-01-21 Sam Weinig <sam@webkit.org>
Reviewed by Geoffrey Sean Garen and Mark Rowe.
WebKitTestRunner needs the HOME environment variable to be set.
<rdar://problem/8896573>
* Scripts/old-run-webkit-tests: Set the HOME environment variable if
it exists.
2011-01-21 Dirk Pranke <dpranke@chromium.org>
Reviewed by Eric Seidel.
Suppress a few remaining integration tests so that none of the
layout_test unit tests ever read from the filesystem or launch
subprocesses that aren't part of the unit tests.
Also fix a minor bug in the printing unit tests that was
incorrectly relying on sys.argv.
https://bugs.webkit.org/show_bug.cgi?id=52863
* Scripts/webkitpy/layout_tests/layout_package/printing_unittest.py:
* Scripts/webkitpy/layout_tests/port/port_testcase.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2011-01-21 Darin Adler <darin@apple.com>
Fix Mac build.
* MiniBrowser/mac/BrowserWindowController.m:
(-[BrowserWindowController awakeFromNib]): Pass 0 for runModal function.
2011-01-20 Darin Adler <darin@apple.com>
Reviewed by Dan Bernstein.
WebKit2: Implement showModalDialog
https://bugs.webkit.org/show_bug.cgi?id=52855
This fixes WebKitTestRunner to compile, but more work is probably
needed to get it to pass the tests.
* WebKitTestRunner/TestController.cpp:
(WTR::TestController::runModal): Added. Calls through to the
platform-specific version of runModal.
(WTR::TestController::createOtherPage): Changed to be a private
static member function so it can refer to runModal, which is
a private static member function.
(WTR::TestController::initialize): Pass 0 for the runModal
function since we don't need to run the main window modal.
I suspect this is wrong and will need to change.
* WebKitTestRunner/TestController.h: Added declarations for
the functions added above.
* WebKitTestRunner/mac/TestControllerMac.mm:
(WTR::TestController::runModal): Added. Untested implementation.
* WebKitTestRunner/qt/TestControllerQt.cpp:
(WTR::TestController::runModal): Added.
* WebKitTestRunner/win/TestControllerWin.cpp:
(WTR::TestController::runModal): Added.
2011-01-21 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
Split Host object off from WebKitPatch for easier re-use
https://bugs.webkit.org/show_bug.cgi?id=52284
new-run-webkit-tests and other tools are eventually going to want one of these.
Basically you need to inherit from this class if you want to pretend to be a
self.tool object for Steps/Commands.
* Scripts/webkitpy/common/host.py: Added.
* Scripts/webkitpy/tool/main.py:
2011-01-20 Dirk Pranke <dpranke@chromium.org>
Reviewed by Adam Barth.
Disable the tests that start and stop the http server and web
server, so that we don't have to worry about port collisions on
the bots.
https://bugs.webkit.org/show_bug.cgi?id=52861
* Scripts/webkitpy/layout_tests/port/port_testcase.py:
2011-01-20 Dirk Pranke <dpranke@chromium.org>
Reviewed by Eric Seidel.
nrwt: rewrite google_chrome_unittest to use mock filesystem.
https://bugs.webkit.org/show_bug.cgi?id=52772
* Scripts/webkitpy/layout_tests/port/google_chrome_unittest.py:
2011-01-20 Dirk Pranke <dpranke@chromium.org>
Reviewed by Eric Siedel.
nrwt: clean up almost all remaining port references, remove
unnecessary import clauses. The only remaining references are
in places where a mock filesystem makes no sense or can't be
used, and in one routine in port/google_chrome_unittest that
I'll rewrite in a separate patch.
https://bugs.webkit.org/show_bug.cgi?id=52771
* Scripts/webkitpy/layout_tests/layout_package/test_runner.py:
* Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
* Scripts/webkitpy/layout_tests/port/chromium_gpu_unittest.py:
* Scripts/webkitpy/layout_tests/port/chromium_linux.py:
* Scripts/webkitpy/layout_tests/port/chromium_mac.py:
* Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
* Scripts/webkitpy/layout_tests/port/chromium_win.py:
* Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
* Scripts/webkitpy/layout_tests/port/google_chrome.py:
* Scripts/webkitpy/layout_tests/port/google_chrome_unittest.py:
* Scripts/webkitpy/layout_tests/port/gtk.py:
* Scripts/webkitpy/layout_tests/port/mac.py:
* Scripts/webkitpy/layout_tests/port/port_testcase.py:
* Scripts/webkitpy/layout_tests/port/qt.py:
* Scripts/webkitpy/layout_tests/port/webkit.py:
* Scripts/webkitpy/layout_tests/port/win.py:
* Scripts/webkitpy/layout_tests/test_types/image_diff.py:
2011-01-20 Dirk Pranke <dpranke@chromium.org>
Reviewed by Mihai Parparita.
remove fs refs from run_webkit_tests, rebaseline_chromium_webkit_tests
https://bugs.webkit.org/show_bug.cgi?id=52762
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2011-01-20 Tony Chang <tony@chromium.org>
Reviewed by Ojan Vafai.
[chromium] fix the paths used for the rebaseline tool
https://bugs.webkit.org/show_bug.cgi?id=52857
Some of the bots were renamed so the path to grab the results zip
file has changed.
* Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
2011-01-20 Dirk Pranke <dpranke@chromium.org>
Reviewed by Tony Chang.
Fix bug introduced in r72688 / bug 52768 that broke the
generation of the output filename links in the results.html
output. Apparently none of the unit tests actually checked to
make sure the generated links were correct :(
https://bugs.webkit.org/show_bug.cgi?id=52854
* Scripts/webkitpy/layout_tests/layout_package/test_failures.py:
* Scripts/webkitpy/layout_tests/layout_package/test_failures_unittest.py:
2011-01-20 Alexey Proskuryakov <ap@apple.com>
Reviewed by Darin Adler.
https://bugs.webkit.org/show_bug.cgi?id=52849
Make window.print work with WebKit2
* MiniBrowser/mac/BrowserWindowController.m:
(-[BrowserWindowController awakeFromNib]):
* WebKitTestRunner/TestController.cpp:
(WTR::createOtherPage):
(WTR::TestController::initialize):
Added zeroes to WKPageUIClient structures to avoid build failures.
2011-01-20 Dirk Pranke <dpranke@chromium.org>
Reviewed by Eric Seidel.
nrwt: remove fs refs from printing, test_failures, test_expectations, text_diff
https://bugs.webkit.org/show_bug.cgi?id=52756
* Scripts/webkitpy/layout_tests/layout_package/printing.py:
* Scripts/webkitpy/layout_tests/layout_package/printing_unittest.py:
* Scripts/webkitpy/layout_tests/layout_package/test_expectations.py:
* Scripts/webkitpy/layout_tests/layout_package/test_expectations_unittest.py:
* Scripts/webkitpy/layout_tests/layout_package/test_failures.py:
* Scripts/webkitpy/layout_tests/test_types/text_diff.py:
2011-01-20 Mansi Mithal <mansi.mithal@nokia.com>
Reviewed by Antonio Gomes.
QtTestBrowser should have a seperate menu for Settings
https://bugs.webkit.org/show_bug.cgi?id=52817
Added a new menu item named "Settings"
and moved the "Enable interrupting js scripts" and "Enable js pop up windows" under the "Settings" menu
* QtTestBrowser/launcherwindow.cpp:
(LauncherWindow::createChrome):
2011-01-20 Sam Weinig <sam@webkit.org>
Reviewed by John Sullivan.
Update update-webkit-localizable-strings with new path to WebKit.
* Scripts/update-webkit-localizable-strings:
2011-01-20 James Robinson <jamesr@chromium.org>
Reviewed by Darin Fisher.
Implement mozilla's requestAnimationFrame API
https://bugs.webkit.org/show_bug.cgi?id=51218
Chromium DumpRenderTree support for window.webkitRequestAnimationFrame.
* DumpRenderTree/chromium/WebViewHost.cpp:
(invokeScheduleComposite):
(WebViewHost::scheduleAnimation):
(WebViewHost::paintInvalidatedRegion):
* DumpRenderTree/chromium/WebViewHost.h:
2011-01-20 Maciej Stachowiak <mjs@apple.com>
Reviewed by Adam Roben.
WebKitTestRunner needs to support layoutTestController.evaluateScriptInIsolatedWorld
https://bugs.webkit.org/show_bug.cgi?id=42327
* WebKitTestRunner/InjectedBundle/Bindings/CodeGeneratorTestRunner.pm: Add support
for methods that take their normal arguments but also a JSContextRef.
* WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl: IDL definition
for evaluateScriptInIsolatedWorld.
* WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
(WTR::InjectedBundlePage::didClearWindowForFrame): Set a magic variable only if
this call is for an isolated world.
* WebKitTestRunner/InjectedBundle/LayoutTestController.cpp:
(WTR::worldMap): Helper to create a world map.
(WTR::LayoutTestController::worldIDForWorld): Map from an ID to a world.
(WTR::LayoutTestController::evaluateScriptInIsolatedWorld): The newly
added LayoutTestController API.
* WebKitTestRunner/InjectedBundle/LayoutTestController.h:
2011-01-19 Adam Roben <aroben@apple.com>
Convert paths in environment variables back to msys-style inside commit-log-editor
When this script gets run from inside git commit, msys-style paths in the environment will
have been turned into Windows-style paths with forward slashes. This screws up functions
like File::Spec->rel2abs, which seem to rely on $PWD having an msys-style path. We convert
the paths back to msys-style before doing anything else.
Fixes <http://webkit.org/b/48527> commit-log-editor uses full paths for section headers when
using msysgit's Perl and multiple ChangeLogs have been edited
Reviewed by David Kilzer.
* Scripts/commit-log-editor: Call fixEnvironment before doing anything else.
(fixEnvironment): Added. When run in msys in conjunction with git (i.e., when invoked from
inside git commit), convert Windows-style paths in the environment back to msys-style paths.
2011-01-20 Zoltan Horvath <zoltan@webkit.org>
[Win] Unreviewed build fix after r76248.
* DumpRenderTree/ForwardingHeaders/wtf/FastAllocBase.h: Added.
2011-01-20 Zoltan Horvath <zoltan@webkit.org>
Reviewed by Csaba Osztrogonác.
Refactoring of the custom allocation framework
https://bugs.webkit.org/show_bug.cgi?id=49897
Inheriting from FastAllocBase can result in objects getting larger (bug #33896, #46589).
The modification replaces Noncopyable and FastAllocBase classes and these inherits with their
equivalent macro implementation at the necessary places.
2011-01-20 Yi Shen <yi.4.shen@nokia.com>
Unreviewed.
Adding myself to committers.py.
* Scripts/webkitpy/common/config/committers.py:
2011-01-20 Adam Barth <abarth@webkit.org>
Reviewed by Tony Chang.
Chromium builders don't delete unversioned directories when DEPS change
https://bugs.webkit.org/show_bug.cgi?id=52745
If we pass this option to gclient sync, gclient will delete directories
that are no longer part of DEPS instead of leaving them around to
confuse us later. Apparently, the downstream buildbots use this option
already.
* Scripts/update-webkit-chromium:
2011-01-19 Dirk Pranke <dpranke@chromium.org>
Unreviewed, build fix - I pulled this patch in to fix an issue
revealed by r76195 running on some new bots). Also fix a couple
of other issues revealed in testing.
nrwt: remove fs refs from layout_package/json*
https://bugs.webkit.org/show_bug.cgi?id=52754
* Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
* Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
* Scripts/webkitpy/layout_tests/layout_package/json_results_generator_unittest.py:
* Scripts/webkitpy/layout_tests/layout_package/test_runner.py:
* Scripts/webkitpy/layout_tests/port/base.py:
2011-01-19 Dirk Pranke <dpranke@chromium.org>
Unreviewed, build fix.
Fix a couple of minor bugs discovered while bringing up new bots
for testing.
* Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2011-01-19 James Robinson <jamesr@chromium.org>
Unreviewed, rolling out r76194.
http://trac.webkit.org/changeset/76194
https://bugs.webkit.org/show_bug.cgi?id=51218
Caused mysterious compile failure on the chromium win
build.webkit.org bots
* DumpRenderTree/chromium/WebViewHost.cpp:
(WebViewHost::paintInvalidatedRegion):
* DumpRenderTree/chromium/WebViewHost.h:
2011-01-19 Dirk Pranke <dpranke@chromium.org>
Reviewed by Mihai Parparita.
remove fs refs from test_runner, dump_render_tree_thread
https://bugs.webkit.org/show_bug.cgi?id=52753
* Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
* Scripts/webkitpy/layout_tests/layout_package/test_runner.py:
* Scripts/webkitpy/layout_tests/layout_package/test_runner_unittest.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2011-01-19 James Robinson <jamesr@chromium.org>
Reviewed by Darin Fisher.
Implement mozilla's requestAnimationFrame API
https://bugs.webkit.org/show_bug.cgi?id=51218
Chromium DumpRenderTree support for window.webkitRequestAnimationFrame.
* DumpRenderTree/chromium/WebViewHost.cpp:
(invokeScheduleComposite):
(WebViewHost::scheduleAnimation):
(WebViewHost::paintInvalidatedRegion):
* DumpRenderTree/chromium/WebViewHost.h:
2011-01-19 Dirk Pranke <dpranke@chromium.org>
Unreviewed, build fix.
Fix a few incorrect "self._filesystem" references introduced in
r76184.
* Scripts/webkitpy/layout_tests/port/chromium.py:
* Scripts/webkitpy/layout_tests/port/chromium_linux.py:
2011-01-19 Dirk Pranke <dpranke@chromium.org>
Reviewed by Mihai Parparita.
add a .sep property, abspath(), isabs(), mtime(), and
open_text_file_for_writing() to filesystem modules. Some of
these properties are not needed in this patch but will be needed
in subsequent patches (I'm doing this to avoid having to track
multiple versions of a single file).
Also, change most of the port/* modules to use the filesystem
objects instead of referencing the filesystem directly.
https://bugs.webkit.org/show_bug.cgi?id=52748
* Scripts/webkitpy/common/system/filesystem.py:
* Scripts/webkitpy/common/system/filesystem_mock.py:
* Scripts/webkitpy/layout_tests/port/__init__.py:
* Scripts/webkitpy/layout_tests/port/base.py:
* Scripts/webkitpy/layout_tests/port/base_unittest.py:
* Scripts/webkitpy/layout_tests/port/chromium.py:
* Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
* Scripts/webkitpy/layout_tests/port/config.py:
* Scripts/webkitpy/layout_tests/port/test.py:
* Scripts/webkitpy/layout_tests/port/test_files_unittest.py:
2011-01-19 Maciej Stachowiak <mjs@apple.com>
Reviewed by Anders Carlsson.
If resetting between tests times out, treat it as a WebProcess crash
https://bugs.webkit.org/show_bug.cgi?id=52757
* WebKitTestRunner/TestController.cpp:
(WTR::TestController::runTest):
2011-01-19 Alexey Proskuryakov <ap@apple.com>
Reviewed by Darin Adler.
https://bugs.webkit.org/show_bug.cgi?id=52739
Make it possible for a WebKit2 client to print headers and footers
Added dummy implementations for new printing UIClient calls.
* MiniBrowser/mac/BrowserWindowController.m:
(-[BrowserWindowController awakeFromNib]):
* WebKitTestRunner/TestController.cpp:
(WTR::createOtherPage):
(WTR::TestController::initialize):
2011-01-19 Dirk Pranke <dpranke@chromium.org>
Reviewed by Ojan Vafai.
Change webkitpy/layout_tests/port/test.py to use the
in-memory filesystem for cleaner unit testing. This
change allows us to kill a lot of code that was
specific to the test port, at the cost of being a
little less clear about how things would work if
you didn't have a filesystem.
https://bugs.webkit.org/show_bug.cgi?id=52605
* Scripts/webkitpy/layout_tests/port/test.py:
2011-01-19 Dirk Pranke <dpranke@chromium.org>
Unreviewed, build fix.
Fix test failure caused by committing attachment 79368 instead
of attachment 79461 on bug 52604.
* Scripts/webkitpy/common/system/filesystem_mock.py:
2011-01-19 Dirk Pranke <dpranke@chromium.org>
Reviewed by Mihai Parparita.
new-run-webkit-tests: remove use of os.walk, use mock filesystem for better
unit testing. os.walk() is too much of a hassle to implement on
top of the in-memory mock filesystem and adding the necessary
interface to files_under() gives clients a cleaner API anyway
(for this particular usage model).
https://bugs.webkit.org/show_bug.cgi?id=52691
* Scripts/webkitpy/common/system/filesystem.py:
* Scripts/webkitpy/common/system/filesystem_mock.py:
* Scripts/webkitpy/layout_tests/port/test_files.py:
* Scripts/webkitpy/layout_tests/port/test_files_unittest.py:
2011-01-19 Dirk Pranke <dpranke@chromium.org>
Reviewed by Tony Chang.
Change more modules in the layout-tests code to use the
filesystem wrapper for cleaner unit testing.
This patch also adds the glob() wrapper to the filesystem
abstraction.
https://bugs.webkit.org/show_bug.cgi?id=52604
* Scripts/webkitpy/layout_tests/port/test_files.py:
* Scripts/webkitpy/layout_tests/port/test_files_unittest.py:
* Scripts/webkitpy/layout_tests/test_types/test_type_base.py:
* Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
* Scripts/webkitpy/common/system/filesystem.py:
* Scripts/webkitpy/common/system/filesystem_mock.py:
2011-01-19 Levi Weintraub <leviw@chromium.org>
Unreviewed.
Updating my email addresses and irc nick.
* Scripts/webkitpy/common/config/committers.py:
2011-01-19 Tony Chang <tony@chromium.org>
Reviewed by Mihai Parparita.
[chromium] [linux] if check-sys-deps fails, output the failure reason
https://bugs.webkit.org/show_bug.cgi?id=52671
* Scripts/webkitpy/common/system/executive_mock.py: Add support for
error handler functions.
* Scripts/webkitpy/layout_tests/port/chromium.py:
output the error text from --check-sys-deps
* Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
2011-01-19 Aparna Nandyal <aparna.nand@wipro.com>
Reviewed by Andreas Kling.
[Qt] QtTestBrowser crashes when closing if Facebook is opened
https://bugs.webkit.org/show_bug.cgi?id=52554
QtNetworkAccessManager is created in a thread and set as
member variable in QWebPage. When the thread is destroyed
the object is still used. So fix will set the object to
NULL once the thread is destroyed.
* QtTestBrowser/launcherwindow.cpp:
(LauncherWindow::~LauncherWindow):
2011-01-18 Maciej Stachowiak <mjs@apple.com>
Rubber stamped by Csaba Osztrogonac
* WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp: Removed unncessary inculde of WKStringCF.h
2011-01-18 Dirk Pranke <dpranke@chromium.org>
Reviewed by Tony Chang.
new-run-webkit-tests: add a --build-directory command line argument
Official Google Chrome builds use a non-standard build directory
location. This patch adds a --build-directory argument that
allows that location to be specified on the command line. Only
Chromium-based builds will use this flag for now, but anyone can
in the future.
There are no unit tests for this since it's difficult to test
until mock filesystems are fully supported in the code. This was
tested by hand for now.
https://bugs.webkit.org/show_bug.cgi?id=52694
* Scripts/webkitpy/layout_tests/port/chromium_linux.py:
* Scripts/webkitpy/layout_tests/port/chromium_mac.py:
* Scripts/webkitpy/layout_tests/port/chromium_win.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2011-01-18 Maciej Stachowiak <mjs@apple.com>
Not reviewed. Bot fix.
Remove an inadvertently committed debugging print.
* DumpRenderTree/mac/LayoutTestControllerMac.mm:
(LayoutTestController::notifyDone):
2011-01-18 Maciej Stachowiak <mjs@apple.com>
Reviewed by Sam Weinig.
WebKitTestRunner should track loading more like DumpRenderTree
https://bugs.webkit.org/show_bug.cgi?id=52692
Change load tracking to track the current top loading frame, in the manner of DumpRenderTree.
This makes some tests that call notifyDone multiple times pass.
* WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
(WTR::InjectedBundle::InjectedBundle):
(WTR::InjectedBundle::done):
* WebKitTestRunner/InjectedBundle/InjectedBundle.h:
(WTR::InjectedBundle::topLoadingFrame):
(WTR::InjectedBundle::setTopLoadingFrame):
* WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
(WTR::InjectedBundlePage::InjectedBundlePage):
(WTR::InjectedBundlePage::stopLoading):
(WTR::InjectedBundlePage::didStartProvisionalLoadForFrame):
(WTR::InjectedBundlePage::didFailProvisionalLoadWithErrorForFrame):
(WTR::InjectedBundlePage::didFinishLoadForFrame):
(WTR::InjectedBundlePage::didFailLoadWithErrorForFrame):
* WebKitTestRunner/InjectedBundle/InjectedBundlePage.h:
* WebKitTestRunner/InjectedBundle/LayoutTestController.cpp:
(WTR::LayoutTestController::notifyDone):
* WebKitTestRunner/TestController.cpp:
(WTR::TestController::didReceiveMessageFromInjectedBundle):
2011-01-18 Mihai Parparita <mihaip@chromium.org>
Reviewed by Kent Tamura.
[Chromium] NRWT looking for DRT binary in the wrong path on Mac
https://bugs.webkit.org/show_bug.cgi?id=52678
Update path now that WebKit is inside Source.
* Scripts/webkitpy/layout_tests/port/chromium_mac.py:
2011-01-18 Robert Hogan <robert@webkit.org>
Reviewed by Simon Hausmann.
[Qt] http/tests/incremental/slow-utf8-text.pl fails
https://bugs.webkit.org/show_bug.cgi?id=44282
* DumpRenderTree/qt/DumpRenderTreeQt.cpp:
(WebCore::DumpRenderTree::dump):
2011-01-18 Dirk Pranke <dpranke@chromium.org>
Reviewed by Mihai Parparita.
Fix a small race condition that could cause the unit tests for
new-run-webkit-tests to fail in a multithreaded environment.
There was a small window in the code where one thread would
check to see if a thread had raised an exception, do something,
and then check to see if the thread was alive. If the other
thread raised an exception and exited in between these two
steps, the exception would be lost. Flipping the order of the
two checks fixes the race, since we will now always check for an
exception on the thread before exiting.
https://bugs.webkit.org/show_bug.cgi?id=51579
* Scripts/webkitpy/layout_tests/layout_package/message_broker.py:
2011-01-18 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
commit-queue dies when test archiving fails
https://bugs.webkit.org/show_bug.cgi?id=52617
I looked at the machine and it had 10 archives already
thus find_unused_name was returning None. I've upped
the limit to 100 (per bug) and tested the case where
find_unused_name returns None (making archive return None).
* Scripts/webkitpy/common/system/workspace.py:
* Scripts/webkitpy/common/system/workspace_unittest.py:
* Scripts/webkitpy/tool/bot/commitqueuetask.py:
* Scripts/webkitpy/tool/bot/commitqueuetask_unittest.py:
* Scripts/webkitpy/tool/commands/queues.py:
2011-01-18 Dirk Pranke <dpranke@chromium.org>
Reviewed by Ojan Vafai.
Update rebaseline-chromium-webkit-tests unit tests to use a mock
filesystem. Also fix a couple of "with" statements missed by
r76050.
https://bugs.webkit.org/show_bug.cgi?id=52487
* Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests_unittest.py:
2011-01-18 Sergio Villar Senin <svillar@igalia.com>
Reviewed by Martin Robinson.
[Gtk] Build fixes for gtk3 build
https://bugs.webkit.org/show_bug.cgi?id=52648
* DumpRenderTree/gtk/EventSender.cpp:
(beginDragWithFilesCallback):
2011-01-18 Dirk Pranke <dpranke@chromium.org>
Reviewed by Mihai Parparita.
Update rebaseline-chromium-webkit-tests to use filesystem objects
instead of direct references to os.path, shutil, tempfile, etc.
This patch doesn't change anything, but will allow subsequent
patches to change the unit tests to no longer use the real
filesystem.
This patch adds a bunch more methods to the filesystem object as
well.
https://bugs.webkit.org/show_bug.cgi?id=52482
* Scripts/webkitpy/common/system/filesystem.py:
* Scripts/webkitpy/common/system/filesystem_mock.py:
* Scripts/webkitpy/tool/commands/queues.py:
* Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
* Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests_unittest.py:
2011-01-18 Dirk Pranke <dpranke@chromium.org>
Reviewed by Mihai Parparita.
new-run-webkit-tests shouldn't crash when you try to run a
directory containing only skipped files.
https://bugs.webkit.org/show_bug.cgi?id=52478
* Scripts/webkitpy/layout_tests/layout_package/test_runner.py:
* Scripts/webkitpy/layout_tests/port/test.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2011-01-18 Adam Bergkvist <adam.bergkvist@ericsson.com> and Martin Robinson <mrobinson@igalia.com>
Reviewed by Gustavo Noronha Silva.
[GTK] The GTK+ EventSender needs support for beginDragWithFiles
https://bugs.webkit.org/show_bug.cgi?id=40833
* DumpRenderTree/gtk/EventSender.cpp: Add a beginDragWithFiles implementation for the
GTK+ event sender.
(dragWithFilesDragDataGetCallback): Added.
(dragWithFilesDragEndCallback): Added.
(beginDragWithFilesCallback): Added.
2011-01-18 Gyuyoung Kim <gyuyoung.kim@samsung.com>
Reviewed by Adam Barth.
ews: Add additional watcher to efl ews
https://bugs.webkit.org/show_bug.cgi?id=52613
Add gyuyoung.kim@samsung.com to watcher list of efl-ews.
* Scripts/webkitpy/tool/commands/earlywarningsystem.py:
2011-01-17 Tony Gentilcore <tonyg@chromium.org>
Reviewed by Alexey Proskuryakov.
Fix some headers with missing or misspelled #ifndef guards
https://bugs.webkit.org/show_bug.cgi?id=52545
* DumpRenderTree/chromium/TestEventPrinter.h:
* DumpRenderTree/chromium/WebPreferences.h:
2011-01-17 Jessie Berlin <jberlin@apple.com>
Reviewed by Anders Carlsson.
REGRESSION: Assertion failure in FrameLoader::continueLoadAfterWillSubmitForm() when
navigating back to an unreachable URL
https://bugs.webkit.org/show_bug.cgi?id=52388
Add a third parameter (unreachableURL) to queueLoadHTMLString.
* DumpRenderTree/LayoutTestController.cpp:
(queueLoadHTMLStringCallback):
(LayoutTestController::queueLoadHTMLString):
(LayoutTestController::queueLoadAlternateHTMLString):
* DumpRenderTree/LayoutTestController.h:
* DumpRenderTree/WorkQueueItem.h:
(LoadHTMLStringItem::LoadHTMLStringItem):
* DumpRenderTree/chromium/LayoutTestController.cpp:
(WorkItemLoadHTMLString::WorkItemLoadHTMLString):
(WorkItemLoadHTMLString::run):
(LayoutTestController::queueLoadHTMLString):
* DumpRenderTree/gtk/WorkQueueItemGtk.cpp:
(LoadHTMLStringItem::invoke):
* DumpRenderTree/mac/WorkQueueItemMac.mm:
(LoadHTMLStringItem::invoke):
* DumpRenderTree/win/WorkQueueItemWin.cpp:
(LoadHTMLStringItem::invoke):
2011-01-17 Dan Bernstein <mitz@apple.com>
Rubber-stamped by Mark Rowe.
Update xcodeproj svn:ignore to include xcuserdata.
* DumpRenderTree/DumpRenderTree.xcodeproj: Modified property svn:ignore.
* MiniBrowser/MiniBrowser.xcodeproj: Modified property svn:ignore.
* TestWebKitAPI/TestWebKitAPI.xcodeproj: Modified property svn:ignore.
* WebKitLauncher/WebKitLauncher.xcodeproj: Modified property svn:ignore.
* WebKitTestRunner/WebKitTestRunner.xcodeproj: Modified property svn:ignore.
2011-01-17 Csaba Osztrogonác <ossy@webkit.org>
Unreviewed fix after r75908.
Move WebKit into Source
https://bugs.webkit.org/show_bug.cgi?id=52530
* BuildSlaveSupport/build.webkit.org-config/master.cfg: Add Source to the path for Qt API tests.
2011-01-17 Csaba Osztrogonác <ossy@webkit.org>
Unreviewed.
Remove accidentally committed debug printing after r75908.
* Scripts/webkitdirs.pm:
2011-01-17 John Knottenbelt <jknotten@chromium.org>
Unreviewed.
Add myself to committers list
https://bugs.webkit.org/show_bug.cgi?id=52566
* Scripts/webkitpy/common/config/committers.py:
2011-01-17 John Knottenbelt <jknotten@chromium.org>
Reviewed by Jeremy Orlow.
GeolocationController should call stopUpdating on destruction
https://bugs.webkit.org/show_bug.cgi?id=52216
fast/dom/Geolocation/window-close-crash.html requires that a
Geolocation watch be started in a secondary window. Consequently,
we need to allow geolocation permission and provide a mock
geolocation position for the secondary window's
GeolocationClientMock.
* DumpRenderTree/chromium/LayoutTestController.cpp:
(LayoutTestController::setGeolocationPermission):
(LayoutTestController::setMockGeolocationPosition):
(LayoutTestController::setMockGeolocationError):
* DumpRenderTree/chromium/TestShell.h:
(TestShell::windowList):
2011-01-16 Adam Barth <abarth@webkit.org>
Update more include paths to reflect WebKit move.
* DumpRenderTree/chromium/EventSender.cpp:
(EventSender::keyDown):
(EventSender::needsShiftModifier):
* DumpRenderTree/chromium/WebThemeEngineDRTMac.h:
* DumpRenderTree/chromium/WebThemeEngineDRTMac.mm:
2011-01-16 Kent Tamura <tkent@chromium.org>
Unreviewed, Chromium build fix.
* Scripts/update-webkit-chromium: Webkit/chromium -> Source/WebKit/chromium
2011-01-16 Martin Robinson <mrobinson@igalia.com>
Build fix for GTK+ after source directory movement.
* GNUmakefile.am: Correct include paths.
2011-01-16 Adam Barth <abarth@webkit.org>
Teach svn-apply about WebKit move.
* Scripts/VCSUtils.pm:
2011-01-16 Adam Barth <abarth@webkit.org>
Rubber-stamped by Eric Seidel.
Move WebKit into Source
https://bugs.webkit.org/show_bug.cgi?id=52530
* DumpRenderTree/mac/InternalHeaders/WebKit/WebTypesInternal.h:
* DumpRenderTree/qt/DumpRenderTreeQt.cpp:
* DumpRenderTree/qt/DumpRenderTreeQt.h:
* DumpRenderTree/qt/GCControllerQt.cpp:
* DumpRenderTree/qt/LayoutTestControllerQt.cpp:
* DumpRenderTree/qt/PlainTextControllerQt.cpp:
* DumpRenderTree/qt/TextInputControllerQt.cpp:
* DumpRenderTree/wscript:
* QtTestBrowser/launcherwindow.h:
* Scripts/build-webkit:
* Scripts/webkitdirs.pm:
2011-01-15 Adam Barth <abarth@webkit.org>
Teach svn-apply about WebKit2 move.
* Scripts/VCSUtils.pm:
2011-01-15 Adam Barth <abarth@webkit.org>
Rubber-stamped by Eric Seidel.
Move WebKit2 into Source
https://bugs.webkit.org/show_bug.cgi?id=52438
* MiniBrowser/qt/MiniBrowser.pro:
* Scripts/build-webkit:
* Scripts/do-file-rename:
* Scripts/do-webcore-rename:
* Scripts/test-webkitpy:
* Scripts/webkitdirs.pm:
* Scripts/webkitpy/style/checker.py:
* WebKitTestRunner/DerivedSources.pro:
* WebKitTestRunner/InjectedBundle/qt/InjectedBundle.pro:
* WebKitTestRunner/qt/WebKitTestRunner.pro:
2011-01-15 Jochen Eisinger <jochen@chromium.org>
Reviewed by David Levin.
[chromium] remove obsolete createView method
https://bugs.webkit.org/show_bug.cgi?id=52437
* DumpRenderTree/chromium/WebViewHost.cpp:
(WebViewHost::createView):
* DumpRenderTree/chromium/WebViewHost.h:
2011-01-14 Brian Weinstein <bweinstein@apple.com>
Reviewed by Geoff Garen.
ShouldLoadResourceForFrame should use strings, not URLs.
https://bugs.webkit.org/show_bug.cgi?id=52476
* WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
(WTR::InjectedBundlePage::shouldLoadResourceForFrame): Update to use a WKStringRef.
* WebKitTestRunner/InjectedBundle/InjectedBundlePage.h:
2011-01-14 Tony Gentilcore <tonyg@chromium.org>
Reviewed by David Levin.
Add Tony Gentilcore as a reviewer
https://bugs.webkit.org/show_bug.cgi?id=52480
* Scripts/webkitpy/common/config/committers.py:
2011-01-14 David Levin <levin@chromium.org>
Reviewed by Eric Seidel.
check-webkit-style: _FunctionState should keep track of where the function name starts.
https://bugs.webkit.org/show_bug.cgi?id=52454
* Scripts/webkitpy/style/checkers/cpp.py: Added function_name_start_position to
_FunctionState and logic for setting it correctly.
* Scripts/webkitpy/style/checkers/cpp_unittest.py: Added tests.
2011-01-14 Vincent Scheib <scheib@chromium.org>
Reviewed by David Levin.
Add support to DumpRenderTree to use the GPU rather than software renderer
https://bugs.webkit.org/show_bug.cgi?id=52333
* DumpRenderTree/chromium/DumpRenderTree.cpp:
(main):
* Scripts/webkitpy/layout_tests/port/chromium.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2011-01-14 Tony Chang <tony@chromium.org>
Reviewed by Kent Tamura.
[chromium] update fonts.conf to fix 6 linux tests in drt
https://bugs.webkit.org/show_bug.cgi?id=52411
* DumpRenderTree/chromium/fonts.conf: Update to match the copy in
chromium's src/webkit/tools/test_shell/resources/fonts.conf
2011-01-14 Eric Seidel <eric@webkit.org>
Original patch from Peter Gal <galpater@inf.u-szeged.hu>
Reviewed by Adam Barth.
Fix when running Tools/Scripts/check-webkit-style without arguments.
https://bugs.webkit.org/show_bug.cgi?id=52261#c16
* Scripts/webkitpy/common/checkout/scm.py: Handle None for changed_files argument better in Git.create_patch method.
* Scripts/webkitpy/common/checkout/scm_unittest.py:
2011-01-04 Jochen Eisinger <jochen@chromium.org>
Reviewed by David Levin.
Store the tempdir object in a variable so it doesn't get deleted too early.
https://bugs.webkit.org/show_bug.cgi?id=51875
* Scripts/update-webkit-auxiliary-libs:
* Scripts/update-webkit-support-libs:
2011-01-14 Eric Seidel <eric@webkit.org>
Unreviewed. Make it work in python 2.5.
webkit-patch suggest-reviewers dies when ChangeLogs are missing
https://bugs.webkit.org/show_bug.cgi?id=49158
* Scripts/webkitpy/common/checkout/api_unittest.py:
2011-01-14 Sheriff Bot <webkit.review.bot@gmail.com>
Unreviewed, rolling out r75774.
http://trac.webkit.org/changeset/75774
https://bugs.webkit.org/show_bug.cgi?id=52431
gtk builds were broken (Requested by loislo2 on #webkit).
* Scripts/webkitpy/common/checkout/scm.py:
2011-01-14 David Levin <levin@chromium.org>
Reviewed by Shinichiro Hamaji.
check-webkit-style: _FunctionState should detect pure functions.
https://bugs.webkit.org/show_bug.cgi?id=52428
* Scripts/webkitpy/style/checkers/cpp.py: Added detection for if a function declaration
is a pure virtual function.
* Scripts/webkitpy/style/checkers/cpp_unittest.py: Added tests for verifying that pure
virtual functions are correctly detected.
2011-01-13 Ilya Tikhonovsky <loislo@chromium.org>
Reviewed by Yury Semikhatsky.
Just small fix for problem with check-webkit-style.
It just didn't work for this patch.
Old code can't concatenate a list of strings with None.
* Scripts/webkitpy/common/checkout/scm.py:
2011-01-13 David Levin <levin@chromium.org>
Reviewed by Shinichiro Hamaji.
check-webkit-style: _FunctionState should use Position to know where items are.
https://bugs.webkit.org/show_bug.cgi?id=52424
* Scripts/webkitpy/style/checkers/cpp.py:
(_FunctionState.*): Changed _FunctionState to use Position instead of line numbers.
(detect_functions): Changed to pass Position's to _FunctionState.begin and
did some minor clean-up.
(*): Other changes are simply about converting to use the Position's in_FunctionState
instead of line numbers.
* Scripts/webkitpy/style/checkers/cpp_unittest.py: Changed the test code
to verify the positions stored in _FunctionState.
2011-01-13 Eric Seidel <eric@webkit.org>
Reviewed by David Levin.
webkit-patch suggest-reviewers dies when ChangeLogs are missing
https://bugs.webkit.org/show_bug.cgi?id=49158
This is not the most elegant, but it is a very safe fix to this bug.
One advantage of catching ScriptError like this instead of adding a
new added_or_modified_files or fixing all changed_files callers
to use a more specific change_files variant, is that we catch
all kinds of ScriptErrors which might cause our (non-essential)
suggest-reviewers code to fail out. This should make passing
--suggest-reviewers to webkit-patch upload much more robust
and may even make it possible for us to make it default.
The root of the problem here is that SCM.changed_files includes
deleted ChangeLog paths (from moves, etc) which then when we ask
SVN/Git for the contents of the file at that revision, the command
errors out and Executive.run_command raises a ScriptError.
In the future we might fix this differently by making all current
callers of chagned_files use a more specific method for requesting
what types of changes they're interested in (adds, modifies, deletes, etc.)
* Scripts/webkitpy/common/checkout/api.py:
* Scripts/webkitpy/common/checkout/api_unittest.py:
2011-01-13 Dan Bernstein <mitz@apple.com>
Reviewed by Alexey Proskuryakov.
<rdar://problem/8827065> REGRESSION (r71884): Cross-origin XHR fails if willSendRequest changes the URL
https://bugs.webkit.org/show_bug.cgi?id=52419
Allow to further control the behavior of the willSendRequest delegate callback by making it
set a different URL in the new request.
* DumpRenderTree/LayoutTestController.cpp:
(addURLToRedirectCallback): Added.
(LayoutTestController::staticFunctions): Added addURLToRedirect.
(LayoutTestController::addURLToRedirect): Added. Adds the redirection to the map.
(LayoutTestController::redirectionDestinationForURL): Added this getter.
* DumpRenderTree/LayoutTestController.h:
* DumpRenderTree/mac/ResourceLoadDelegate.mm:
(-[ResourceLoadDelegate webView:resource:willSendRequest:redirectResponse:fromDataSource:]): Check
if there is a redirection for the request’s URL, established by addURLToRedirect(), and if so,
change the URL in the new request to the redirect destination.
2011-01-13 Eric Seidel <eric@webkit.org>
Reviewed by Mihai Parparita.
webkit-patch suggest-reviewers fails due to invalid utf8 in ChangeLog files
https://bugs.webkit.org/show_bug.cgi?id=52416
Example: webkit-patch suggest-reviewers -g ca3890fe74d94d85d6bfa48f9ea497b094d8e717
UnicodeDecodeError: 'utf8' codec can't decode bytes in position 536-538: invalid data
I also added FIXMEs about the changed_files returning deleted files problem
which is causing the other most common exception for suggest-reviewers
(and is probably breaking sheriff-bots blame detection).
* Scripts/webkitpy/common/checkout/api.py:
* Scripts/webkitpy/common/checkout/api_unittest.py:
* Scripts/webkitpy/common/checkout/scm.py:
2011-01-13 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
results archives from commit-queue are too large to upload
https://bugs.webkit.org/show_bug.cgi?id=52405
* Scripts/webkitpy/common/system/filesystem.py:
* Scripts/webkitpy/common/system/filesystem_mock.py:
* Scripts/webkitpy/tool/commands/queues.py:
2011-01-13 Alejandro G. Castro <alex@igalia.com>
Reviewed by Martin Robinson.
[gtk] Zero-sized font does not yet work
https://bugs.webkit.org/show_bug.cgi?id=49793
Set minimum-font-size to 0 for the DRT.
* DumpRenderTree/gtk/DumpRenderTree.cpp:
(resetDefaultsToConsistentValues):
2011-01-13 Csaba Osztrogonác <ossy@webkit.org>
Reviewed by Darin Adler.
Make old-run-webkit-tests more robust on buildbots
https://bugs.webkit.org/show_bug.cgi?id=52364
* BuildSlaveSupport/build.webkit.org-config/master.cfg: Pass --exit-after-n-failures 500 to old-run-webkit-tests script.
2011-01-13 Csaba Osztrogonác <ossy@webkit.org>
Reviewed by Darin Adler.
Make old-run-webkit-tests more robust on buildbots
https://bugs.webkit.org/show_bug.cgi?id=52364
* Scripts/old-run-webkit-tests: Ensure that testResultsDirectory is empty.
2011-01-13 Tony Chang <tony@chromium.org>
Reviewed by Kent Tamura.
[chromium] fix plugins/plugin-javascript-access.html on Linux and Mac
https://bugs.webkit.org/show_bug.cgi?id=52332
Forked Info.plist is not longer needed.
* DumpRenderTree/chromium/TestNetscapePlugIn/Info.plist: Removed.
2011-01-13 Mikhail Naganov <mnaganov@chromium.org>
Reviewed by Timothy Hatcher.
Web Inspector: Cleanup after r75613 -- make all UI-facing strings
to be passed directly to WebInspector.UIString
https://bugs.webkit.org/show_bug.cgi?id=52365
* Scripts/check-inspector-strings:
2011-01-13 Sheriff Bot <webkit.review.bot@gmail.com>
Unreviewed, rolling out r74881.
http://trac.webkit.org/changeset/74881
https://bugs.webkit.org/show_bug.cgi?id=52357
The EFL builder has been red for days (Requested by abarth on
#webkit).
* Scripts/webkitpy/common/net/buildbot/buildbot.py:
* Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
2011-01-12 Maciej Stachowiak <mjs@apple.com>
Reviewed by David Levin.
Add beforeunload support to WebKitTestRunner
https://bugs.webkit.org/show_bug.cgi?id=52353
This makes a number of currently failing tests pass.
* WebKitTestRunner/TestController.cpp:
(WTR::runBeforeUnloadConfirmPanel):
(WTR::createOtherPage):
(WTR::TestController::initialize):
2011-01-12 Maciej Stachowiak <mjs@apple.com>
Reviewed by Sam Weinig.
Detect, identify and recover from WebProcess crashes in run-webkit-tests
https://bugs.webkit.org/show_bug.cgi?id=52330
* Scripts/old-run-webkit-tests: Detect when the regression test tool reports
a web process crash.
* WebKitTestRunner/TestController.cpp:
(WTR::TestController::initialize): Set up a processDidCrash
callback.
(WTR::TestController::processDidCrash): Report that the Web process
crashed.
* WebKitTestRunner/TestController.h:
2011-01-12 David Levin <levin@chromium.org>
Reviewed by Shinichiro Hamaji.
check-webkit-style -- close_expression function doesn't work correctly.
https://bugs.webkit.org/show_bug.cgi?id=52272
* Scripts/webkitpy/style/checkers/cpp.py:
(Position.__str__): Added a way to convert it to a string which is useful in tests.
(Position.__cmp__): Added a way to compare Position which is useful in tests and
generally useful (for upcoming code changes).
(close_expression): Changed to use Position for input and output.
Also, fixed many bugs such as only working correctly for parenthesis,
not working correctly if given an offset in a line and not finding
the real closing element if there were multiple closing elements in
the same line.
(detect_functions): Adjusted due to the change in arguments for
close_expression.
* Scripts/webkitpy/style/checkers/cpp_unittest.py:
(CppStyleTestBase.assert_positions_equal): Added a way to verify that
two positions are the same.
(CppStyleTest.test_position): Added tests for the __str_ and __cmp__ methods.
(CppStyleTest.test_close_expression): Added tests to catch the issues
that were fixed.
2011-01-12 Sheriff Bot <webkit.review.bot@gmail.com>
Unreviewed, rolling out r75576.
http://trac.webkit.org/changeset/75576
https://bugs.webkit.org/show_bug.cgi?id=52304
broke rebaseline-chromium-webkit-test (Requested by thakis on
#webkit).
* Scripts/webkitpy/layout_tests/layout_package/printing_unittest.py:
* Scripts/webkitpy/layout_tests/layout_package/test_expectations.py:
* Scripts/webkitpy/layout_tests/layout_package/test_expectations_unittest.py:
* Scripts/webkitpy/layout_tests/layout_package/test_runner.py:
* Scripts/webkitpy/layout_tests/port/base.py:
* Scripts/webkitpy/layout_tests/port/chromium.py:
* Scripts/webkitpy/layout_tests/port/port_testcase.py:
* Scripts/webkitpy/layout_tests/port/test.py:
* Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
* Scripts/webkitpy/style/checkers/test_expectations.py:
* Scripts/webkitpy/style/checkers/test_expectations_unittest.py:
2011-01-12 Maciej Stachowiak <mjs@apple.com>
Reviewed by Darin Adler.
Add a no-build option to run-webkit-tests, to avoid building DRT/WKTR
https://bugs.webkit.org/show_bug.cgi?id=52307
* Scripts/old-run-webkit-tests:
2011-01-12 Koan-Sin Tan <koansin.tan@gmail.com>
Reviewed by Martin Robinson.
[Gtk] WebKitGtk+ doesn't build on Mac OS X 10.6
https://bugs.webkit.org/show_bug.cgi?id=50867
Fixed library extention for Mac in webkitdirs.pm and type conflict in
TestNetscapePlugin.cpp
* GNUmakefile.am: Don't use AppKit when compiling TestNetscapePlugin for GTK+
* Scripts/webkitdirs.pm:
2011-01-12 Mikhail Naganov <mnaganov@chromium.org>
Reviewed by Pavel Feldman.
Add check-inspector-strings script.
https://bugs.webkit.org/show_bug.cgi?id=52295
* Scripts/check-inspector-strings: Added.
2011-01-12 Adam Roben <aroben@apple.com>
Advertise the .testnetscape file extension on Windows
Fixes <http://webkit.org/b/52298>
plugins/no-mime-with-valid-extension.html fails on Windows
Reviewed by Dan Bernstein.
* DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.rc: Added a
FileExtents field to map the .testnetscape file extension to the
application/x-webkit-test-netscape MIME type, just as we do on Mac.
2011-01-12 Adam Barth <abarth@webkit.org>
Reviewed by Eric Seidel.
Make webkit-patch support subdirectories for SVN checkouts
https://bugs.webkit.org/show_bug.cgi?id=52261
os.relpath does not exist on Python 2.5, so we need to call our version
of the API.
* Scripts/webkitpy/common/checkout/scm.py:
2011-01-11 Maciej Stachowiak <mjs@apple.com>
Reviewed by Eric Seidel.
Close additional pages opened during a test once the test completes.
https://bugs.webkit.org/show_bug.cgi?id=52276
* WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
(WTR::InjectedBundle::done):
2011-01-11 David Levin <levin@chromium.org>
Reviewed by Eric Seidel.
Errors should be separated from the command line in bot messages.
https://bugs.webkit.org/show_bug.cgi?id=52275
* Scripts/webkitpy/common/system/executive.py:
(ScriptError.message_with_output): Added a newline to separate the command
line from the errors.
* Scripts/webkitpy/common/system/executive_unittest.py:
(ScriptErrorTest.test_message_with_output): Added tests for the various code
paths in message_with_output.
2011-01-11 Eric Seidel <eric@webkit.org>
Unreviewed.
commit-queue should know how to upload archived results (for test flakes or general failures)
https://bugs.webkit.org/show_bug.cgi?id=52048
The zips are mostly empty due to forgetting -r.
Expected diffs were not being pulled from the archive due
to the archive having longer paths than I realized.
* Scripts/webkitpy/common/system/workspace.py:
* Scripts/webkitpy/common/system/workspace_unittest.py:
* Scripts/webkitpy/tool/bot/flakytestreporter.py:
* Scripts/webkitpy/tool/bot/flakytestreporter_unittest.py:
2011-01-11 Dirk Pranke <dpranke@chromium.org>
Reviewed by Ojan Vafai.
The current modifier parsing code in test_expectations is
fragile and hard-coded, so it's not easy to understand the logic
or easily add new types of modifiers (like GPU vs. CPU testing
for graphics tests, or 32-bit vs. 64-bit differences).
This is the first of two patches that will add in more generic
support and then eliminate the GPU-specific test expectations
files for Chromium.
This patch adds two standalone objects for handling modifiers. The
rules for interpreting modifiers, precedence, and conflicts are
given in the docstring to the ModifierMatcher class, which
returns ModifierMatchResult objects.
This patch also adds routines to the Port interface and a
default set of values in the base object, in order to obtain the
values needed on a given test run. These values are then passed
to the expectation parser. This also allows us to clean up the
logic used to lint all of the different configurations in a
single test_expectations.txt file.
The next patch will merge in the separate GPU expectations file.
https://bugs.webkit.org/show_bug.cgi?id=51222
* Scripts/webkitpy/layout_tests/layout_package/printing_unittest.py:
* Scripts/webkitpy/layout_tests/layout_package/test_expectations.py:
* Scripts/webkitpy/layout_tests/layout_package/test_expectations_unittest.py:
* Scripts/webkitpy/layout_tests/port/base.py:
* Scripts/webkitpy/layout_tests/port/base_unittest.py:
* Scripts/webkitpy/layout_tests/port/chromium.py:
* Scripts/webkitpy/layout_tests/port/port_testcase.py:
* Scripts/webkitpy/layout_tests/port/test.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2011-01-11 Maciej Stachowiak <mjs@apple.com>
Reviewed by Adam Barth.
Make webkit-patch support subdirectories for SVN checkouts
https://bugs.webkit.org/show_bug.cgi?id=52261
* Scripts/webkitpy/common/checkout/scm.py:
* Scripts/webkitpy/style_references.py:
* Scripts/webkitpy/tool/main.py:
* Scripts/webkitpy/tool/steps/preparechangelog.py:
2011-01-11 Eric Seidel <eric@webkit.org>
Unreviewed.
commit-queue should know how to upload archived results (for test flakes or general failures)
https://bugs.webkit.org/show_bug.cgi?id=52048
I changed the API for archive_last_layout_test_results w/o updating the implementation. Oops.
This fixes an exception seen on the commit-queue when attempting to report flaky tests.
* Scripts/webkitpy/common/system/workspace.py:
* Scripts/webkitpy/common/system/workspace_unittest.py:
* Scripts/webkitpy/tool/commands/queues.py:
2011-01-11 Sam Weinig <sam@webkit.org>
Roll r75474 back in.
2011-01-11 Jocelyn Turcotte <jocelyn.turcotte@nokia.com>
Reviewed by Kenneth Rohde Christiansen.
[Qt] Remove uses of QFuture since it isn't supported on all platforms.
https://bugs.webkit.org/show_bug.cgi?id=51204
* QtTestBrowser/webpage.h:
(QtNAMThread::QtNAMThread):
(QtNAMThread::networkAccessManager):
(QtNAMThread::run):
2011-01-11 Peter Varga <pvarga@webkit.org>
Unreviewed.
Add myself to committers' list.
* Scripts/webkitpy/common/config/committers.py:
2011-01-11 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
commit-queue should know how to upload archived results (for test flakes or general failures)
https://bugs.webkit.org/show_bug.cgi?id=52048
Now the queue will always upload results. Either the entire zip, or just
the diffs.txt in the case of text failures.
This should make understanding flakes much easier, and paves the way
for having the EWS run layout tests (and upload failures).
In order to upload .zip files I had to teach bugzilla.py to autodetect
mime types from the filename. Since mimetypes.py doesn't include a mapping
for .patch files, I have it add one before calling guess_type.
We may find that always uploading the whole zip instead of just the -diffs.txt
file is preferable, but for now I'm keeping the old behavior because it makes
quickly understanding text failures easy.
* Scripts/webkitpy/common/net/bugzilla/bugzilla.py:
* Scripts/webkitpy/common/system/workspace.py: Added.
* Scripts/webkitpy/common/system/workspace_unittest.py: Added.
* Scripts/webkitpy/tool/bot/commitqueuetask.py:
* Scripts/webkitpy/tool/bot/commitqueuetask_unittest.py:
* Scripts/webkitpy/tool/bot/flakytestreporter.py:
* Scripts/webkitpy/tool/bot/flakytestreporter_unittest.py:
* Scripts/webkitpy/tool/commands/queues.py:
* Scripts/webkitpy/tool/commands/queues_unittest.py:
2011-01-10 Sheriff Bot <webkit.review.bot@gmail.com>
Unreviewed, rolling out r75464.
http://trac.webkit.org/changeset/75464
https://bugs.webkit.org/show_bug.cgi?id=52193
Followup rollout for 75474 (Requested by Ossy on #webkit).
* MiniBrowser/mac/BrowserWindowController.m:
(-[BrowserWindowController awakeFromNib]):
* MiniBrowser/win/BrowserView.cpp:
(BrowserView::create):
* WebKitTestRunner/TestController.cpp:
(WTR::createOtherPage):
(WTR::TestController::initialize):
2011-01-10 Laszlo Gombos <laszlo.1.gombos@nokia.com>
Reviewed by Csaba Osztrogonác.
[Qt] Baseline qt_minimal configuration
https://bugs.webkit.org/show_bug.cgi?id=51313
* DumpRenderTree/qt/DumpRenderTreeQt.cpp:
(WebCore::DumpRenderTree::resetToConsistentStateBeforeTesting): Add
QT_NO_UNDOSTACK guard.
* QtTestBrowser/launcherwindow.cpp:
(LauncherWindow::createChrome): Add QT_NO_SHORTCUT and QT_NO_PRINTER
guards.
(LauncherWindow::applyZoom): Add QT_NO_ANIMATION guard.
(LauncherWindow::screenshot): Add QT_NO_FILEDIALOG guard.
(LauncherWindow::selectElements): Add QT_NO_INPUTDIALOG guard.
(LauncherWindow::showUserAgentDialog): Add QT_NO_COMBOBOX guard.
* QtTestBrowser/launcherwindow.h:
* QtTestBrowser/locationedit.cpp: Add QT_NO_INPUTDIALOG guard.
* QtTestBrowser/locationedit.h: Ditto.
* QtTestBrowser/mainwindow.cpp:
(MainWindow::buildUI): Add QT_NO_INPUTDIALOG, QT_NO_SHORTCUT and
QT_NO_UNDOSTACK guards.
(MainWindow::setAddressUrl): Add QT_NO_INPUTDIALOG guard.
(MainWindow::changeLocation): Ditto.
(MainWindow::openFile): Add QT_NO_FILEDIALOG guard.
(MainWindow::openLocation): Add QT_NO_INPUTDIALOG guard.
* QtTestBrowser/webpage.cpp:
(WebPage::openUrlInDefaultBrowser): Add QT_NO_DESKTOPSERVICES guard.
(WebPage::authenticationRequired): Add QT_NO_LINEEDIT guard.
* QtTestBrowser/webview.cpp:
(WebViewGraphicsBased::animatedFlip): Add QT_NO_ANIMATION guard.
(WebViewGraphicsBased::animatedYFlip): Ditto.
2011-01-10 Sam Weinig <sam@webkit.org>
Fix build.
* MiniBrowser/mac/BrowserWindowController.m:
(-[BrowserWindowController awakeFromNib]):
* MiniBrowser/win/BrowserView.cpp:
(BrowserView::create):
* WebKitTestRunner/TestController.cpp:
(WTR::createOtherPage):
(WTR::TestController::initialize):
2011-01-10 Eric Seidel <eric@webkit.org>
Reviewed by Ojan Vafai.
style-queue messages are way too long for big patches
https://bugs.webkit.org/show_bug.cgi?id=52161
We definitely could build much fancier list-to-string-with-limit functions
but this should be sufficient for our needs at the moment.
* Scripts/webkitpy/common/system/executive.py:
* Scripts/webkitpy/common/system/executive_unittest.py:
2011-01-10 Tony Chang <tony@chromium.org>
Unreviewed, rolling out r75398.
http://trac.webkit.org/changeset/75398
https://bugs.webkit.org/show_bug.cgi?id=52008
Some layout tests are crashing on Chromium Win.
* DumpRenderTree/chromium/WebViewHost.cpp:
(WebViewHost::updatePaintRect):
* DumpRenderTree/chromium/WebViewHost.h:
2011-01-10 Adam Roben <aroben@apple.com>
Roll out r75392
It isn't needed now that r75289 has been rolled out.
* Scripts/run-javascriptcore-tests:
2011-01-10 Maciej Stachowiak <mjs@apple.com>
Reviewed by Anders Carlsson.
WebKitTestRunner uses crazy amounts of CPU, making layout tests slow and flaky
https://bugs.webkit.org/show_bug.cgi?id=52160
* WebKitTestRunner/mac/TestControllerMac.mm:
(WTR::TestController::platformRunUntil): Instead of busy-looping, wait for the timeout.
Anything that can make runUntil() conditions turn true will happen on a run loop cycle
anyway.
2011-01-10 Tony Chang <tony@chromium.org>
Reviewed by Mihai Parparita.
[chromium] fix for animated gif layout tests
https://bugs.webkit.org/show_bug.cgi?id=52008
* DumpRenderTree/chromium/WebViewHost.cpp: Schedule a paint when we update the paint rect
to trigger gif animations.
(WebViewHostPaintTask::WebViewHostPaintTask):
(WebViewHostPaintTask::runIfValid):
(WebViewHost::updatePaintRect):
* DumpRenderTree/chromium/WebViewHost.h:
(WebViewHost::taskList):
2011-01-10 Alejandro G. Castro <alex@igalia.com>
Reviewed by Xan Lopez.
[GTK] Fix gtk2 compilation for master
https://bugs.webkit.org/show_bug.cgi?id=51885
* DumpRenderTree/gtk/PixelDumpSupportGtk.cpp:
(createBitmapContextFromWebView): Replaced the old
gdk_drawable_get_size with gdk_pixmap_get_size.
2011-01-10 Patrick Gansterer <paroga@webkit.org>
Reviewed by Laszlo Gombos.
[WINCE] Reenable NPAPI after r75203
https://bugs.webkit.org/show_bug.cgi?id=52109
Remove the !isWinCE() check for ENABLE_NETSCAPE_PLUGIN_API and replace
it with !isEfl(), because EFL port does not support NPAPI (yet).
* Scripts/build-webkit:
2011-01-10 Adam Roben <aroben@apple.com>
Skip some asserting JS tests
See <http://webkit.org/b/52156>.
* Scripts/run-javascriptcore-tests:
2011-01-10 Martin Robinson <mrobinson@igalia.com>
Reviewed by Xan Lopez.
[GTK] build-webkit disable Blob support via arguments passed to automake
https://bugs.webkit.org/show_bug.cgi?id=52097
* Scripts/build-webkit: Enable Blob support when building WebKitGTK+ via
build-webkit.
2011-01-10 Adam Roben <aroben@apple.com>
Use the same name and description for TestNetscapePlugin on Windows and
Mac
Fixes <http://webkit.org/b/52151>
fast/frames/iframe-reparenting-plugins.html fails on Windows
Reviewed by Steve Falkenburg.
* DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.rc: Copied
the name and description from Mac's Info.plist.
2011-01-10 Siddharth Mathur <siddharth.mathur@nokia.com>
Reviewed by Laszlo Gombos.
[Qt][WK2] Guard gcc-specific syntax
Non-gcc based compliers choke on the "-include"
syntax for preinclude, place guard around it.
* WebKitTestRunner/InjectedBundle/qt/InjectedBundle.pro:
* WebKitTestRunner/qt/WebKitTestRunner.pro:
2011-01-10 Andreas Kling <kling@webkit.org>
Reviewed by Simon Hausmann.
[Qt] [WK2] MiniBrowser does not load pages opened in new windows
https://bugs.webkit.org/show_bug.cgi?id=52111
When getting a createNewPage() callback from QWKPage, create the
new page with the same QWKContext as the page that is passed to
the callback.
* MiniBrowser/qt/BrowserView.h:
* MiniBrowser/qt/BrowserWindow.cpp:
(newPageFunction):
(BrowserWindow::BrowserWindow):
* MiniBrowser/qt/BrowserWindow.h:
2011-01-10 Andreas Kling <kling@webkit.org>
Reviewed by Kenneth Rohde Christiansen.
[Qt] MiniBrowser: Remove superfluous createNewPage() callback
BrowserView set its own useless createNewPage() callback on QWKPage.
This was immediately overridden by BrowserWindow anyway, so remove
the BrowserView one.
* MiniBrowser/qt/BrowserView.cpp:
(BrowserView::BrowserView):
2011-01-10 John Knottenbelt <jknotten@chromium.org>
Reviewed by Jeremy Orlow.
[Chromium] Remove non-client-based Geolocation code
https://bugs.webkit.org/show_bug.cgi?id=50921
* DumpRenderTree/chromium/LayoutTestController.cpp:
(LayoutTestController::setGeolocationPermission):
(LayoutTestController::setMockGeolocationPosition):
(LayoutTestController::setMockGeolocationError):
* DumpRenderTree/chromium/WebViewHost.cpp:
(WebViewHost::geolocationClientMock):
(WebViewHost::reset):
* DumpRenderTree/chromium/WebViewHost.h:
2011-01-10 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
Commit bot tried to land r? cq+ patch
https://bugs.webkit.org/show_bug.cgi?id=36638
It should now be possible to post a patch with r? and cq+ and the
commit-queue will correctly wait for the r+ before attempting to land the patch.
(This was a feature requested by several users.)
One caveat: We don't actually reject r-, cq+ patches from the queue
(we just ignore them), but I think that's OK for now.
I also noticed that setting r- once a commit-queue node had grabbed
a patch would not properly cancel the landing. That's fixed and
tested in this patch.
* Scripts/webkitpy/tool/bot/commitqueuetask.py:
* Scripts/webkitpy/tool/bot/commitqueuetask_unittest.py:
* Scripts/webkitpy/tool/bot/feeders.py:
* Scripts/webkitpy/tool/bot/feeders_unittest.py:
2011-01-08 Benjamin Poulain <benjamin.poulain@nokia.com>
Reviewed by Kenneth Rohde Christiansen.
[Qt] QtWebKit crashes when using XMLHttpRequest from the unload event
https://bugs.webkit.org/show_bug.cgi?id=51934
When accessing the network from the destructor of QWebPage, the network access manager
was already deleted because the object WebPage was already destructed.
This solve the problem by linking the lifetime of the network access manager to WebPage's QObject.
The object is now destructed in the destructor of QObject.
* QtTestBrowser/webpage.cpp:
(WebPage::WebPage):
(WebPage::setQnamThreaded):
* QtTestBrowser/webpage.h:
(QtNAMThread::QtNAMThread):
2011-01-08 Eric Seidel <eric@webkit.org>
Reviewed by Ojan Vafai.
test_failures.py should not mention TestShell (long live DumpRenderTree!)
https://bugs.webkit.org/show_bug.cgi?id=52067
I ran test-webkitpy, but didn't bother adding a unit test for this string change.
https://bugs.webkit.org/show_bug.cgi?id=51138#c3 is an example of why this is needed.
* Scripts/webkitpy/layout_tests/layout_package/test_failures.py:
2011-01-08 Eric Seidel <eric@webkit.org>
Reviewed by Mihai Parparita.
commit-queue hangs when the ChangeLog diff is bad
https://bugs.webkit.org/show_bug.cgi?id=52072
* Scripts/webkitpy/tool/steps/validatechangelogs.py:
* Scripts/webkitpy/tool/steps/validatechangelogs_unittest.py:
2011-01-08 Adam Barth <abarth@webkit.org>
Teach svn-apply how to re-write WebCore to Source/WebCore.
* Scripts/VCSUtils.pm:
2011-01-07 Adam Barth <abarth@webkit.org>
Rubber-stamped by Eric Seidel.
Move WebCore to Source
https://bugs.webkit.org/show_bug.cgi?id=52050
Update the tools to understand WebCore's new location.
* DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro:
* DumpRenderTree/wscript:
* GNUmakefile.am:
* Scripts/build-webkit:
* Scripts/do-file-rename:
* Scripts/do-webcore-rename:
* Scripts/generate-qt-inspector-resource:
* Scripts/prepare-ChangeLog:
* Scripts/run-bindings-tests:
* Scripts/update-iexploder-cssproperties:
* Scripts/update-sources-list.py:
* Scripts/update-webkit-localizable-strings:
* Scripts/webkitdirs.pm:
* Scripts/webkitpy/common/config/build.py:
* Scripts/webkitpy/common/config/build_unittest.py:
* Scripts/webkitpy/layout_tests/deduplicate_tests_unittest.py:
* Scripts/webkitpy/style/checker.py:
* Scripts/webkitpy/style/checker_unittest.py:
* WebKitTestRunner/DerivedSources.pro:
* WebKitTestRunner/InjectedBundle/qt/InjectedBundle.pro:
* iExploder/iexploder-1.7.2/tools/update_html_tags_from_sources.sh:
* wx/browser/wscript:
* wx/build/settings.py:
2011-01-07 Maciej Stachowiak <mjs@apple.com>
Reviewed by Sam Weinig.
Make WebKitTestRunner rest between loads more robust to avoid test crashes
https://bugs.webkit.org/show_bug.cgi?id=52086
* WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
(WTR::InjectedBundle::didReceiveMessage):
* WebKitTestRunner/InjectedBundle/LayoutTestController.cpp:
(WTR::LayoutTestController::notifyDone):
* WebKitTestRunner/TestController.cpp:
(WTR::TestController::resetStateToConsistentValues):
(WTR::TestController::run):
2011-01-07 James Robinson <jamesr@chromium.org>
Revert "Implement mozilla's animationTime property"
https://bugs.webkit.org/show_bug.cgi?id=51952
This approach isn't quite right.
* DumpRenderTree/chromium/WebViewHost.cpp:
(WebViewHost::paintInvalidatedRegion):
2011-01-07 Adam Roben <aroben@apple.com>
Build master fix
* BuildSlaveSupport/build.webkit.org-config/master.cfg:
(RunWebKitTests.start): Don't try to concatenate a list and a string.
2011-01-07 Jedrzej Nowacki <jedrzej.nowacki@nokia.com>
Reviewed by Csaba Osztrogonác.
run-qtwebkit-tests should be able to kill a testsuite.
Add new option in the script that can setup a timeout for a test.
If the test execution takes more then specified time then the test
would be terminated.
[Qt] run-qtwebkit-tests needs timeout
https://bugs.webkit.org/show_bug.cgi?id=51894
* BuildSlaveSupport/build.webkit.org-config/master.cfg:
* Scripts/run-qtwebkit-tests:
2011-01-07 Csaba Osztrogonác <ossy@webkit.org>
Unreviewed fix after r75233.
Based on idea from Eric Seidel.
Add remote zip file handling to webkitpy.
https://bugs.webkit.org/show_bug.cgi?id=50901
* Scripts/webkitpy/common/system/fileset.py: Make python 2.5 happy.
2011-01-07 Shane Stephens <shanestephens@google.com>
Reviewed by Eric Seidel.
svn-apply: should support git binary delta diffs
https://bugs.webkit.org/show_bug.cgi?id=38864
* Scripts/VCSUtils.pm:
* Scripts/svn-apply:
2011-01-06 Julie Parent <jparent@chromium.org>
Reviewed by Eric Seidel.
Test Result Server not removing old results.
https://bugs.webkit.org/show_bug.cgi?id=52012
* TestResultServer/model/jsonresults.py:
Fix off-by-one: Truncate if >= num_runs, not just >.
* TestResultServer/model/jsonresults_unittest.py:
Add unit test to test that merging in a new result with the same value
as the last result causes previous result to fall off.
2011-01-06 James Kozianski <koz@chromium.org>
Reviewed by Mihai Parparita.
Add classes to provide a consistent interface to a set of files.
https://bugs.webkit.org/show_bug.cgi?id=50901
These classes allow us to write code that is agnostic to whether a
particular set of files resides in a local directory or in a zip file
on a remote machine.
* Scripts/webkitpy/common/system/directoryfileset.py: Added.
* Scripts/webkitpy/common/system/directoryfileset_unittest.py: Added.
* Scripts/webkitpy/common/system/fileset.py: Added.
* Scripts/webkitpy/common/system/filesystem.py:
* Scripts/webkitpy/common/system/filesystem_mock.py:
* Scripts/webkitpy/common/system/zipfileset.py: Added.
* Scripts/webkitpy/common/system/zipfileset_unittest.py: Added.
2011-01-06 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
NetworkTransaction should log the url it is retrying
https://bugs.webkit.org/show_bug.cgi?id=52043
* Scripts/webkitpy/common/net/networktransaction.py:
* Scripts/webkitpy/common/net/networktransaction_unittest.py:
2011-01-06 Steve Falkenburg <sfalken@apple.com>
Windows build fix.
Import os before using it.
* BuildSlaveSupport/build.webkit.org-config/master.cfg:
2011-01-06 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
StyleQueue fails to report errors when patches have many files
https://bugs.webkit.org/show_bug.cgi?id=52042
This is one of at least two errors preventing style error reporting at the moment.
When we moved to passing the list of files to check explicitly,
we started to exceed the 500 character TEXT field limit for our AppEngine model object.
We could change the type of the QueueStatus.message type, or we could
just teach statusserver.py how to clamp values (and warn when it does).
I chose the latter path.
* Scripts/webkitpy/common/net/statusserver.py:
2011-01-06 Steve Falkenburg <sfalken@apple.com>
Reviewed by Stephanie Lewis.
Fix syntax error in script.
Windows bots need to archive/unarchive builds to/from configuration-specific directories
https://bugs.webkit.org/show_bug.cgi?id=51996
* BuildSlaveSupport/build.webkit.org-config/master.cfg:
2011-01-06 Laszlo Gombos <laszlo.1.gombos@nokia.com>
Unreviewed, fixing WinCE build after r75196.
WinCE port does not (yet) support NPAPI.
* Scripts/build-webkit:
2011-01-06 Xan Lopez <xlopez@igalia.com>
Reviewed by Darin Adler.
File::Spec->rel2abs in sunspider-compare-results mangles parameters
https://bugs.webkit.org/show_bug.cgi?id=52015
* Scripts/sunspider-compare-results: call rel2abs on our arguments
after the platform has been detected, otherwise the script will
mangle things like '--gtk' into '/a/path/--gtk', breaking the
detection.
2011-01-06 Laszlo Gombos <laszlo.1.gombos@nokia.com>
Reviewed by Eric Seidel.
Add ENABLE_NETSCAPE_PLUGIN_API, ENABLE_ORIENTATION_EVENTS and ENABLE_TOUCH_EVENTS flags to build-webkit
https://bugs.webkit.org/show_bug.cgi?id=51346
* Scripts/build-webkit:
2011-01-06 Steve Falkenburg <sfalken@apple.com>
Reviewed by Adam Roben.
Windows bots need to archive/unarchive builds to/from configuration-specific directories
https://bugs.webkit.org/show_bug.cgi?id=51996
After the configuration directory change, we were archiving the incorrect directory for
build product, so we were running tests on the wrong binaries.
This change archives the correct directory, and also unarchives it into a configuration-
specific directory.
* BuildSlaveSupport/build.webkit.org-config/master.cfg: Use a configuration-specific directory on Windows.
* BuildSlaveSupport/built-product-archive: Unarchive/archive to/from proper directory.
2011-01-06 Eric Seidel <eric@webkit.org>
Reviewed by Mihai Parparita.
commit-queue mentions "Text diff mismatch" 4 times instead of once per failure
https://bugs.webkit.org/show_bug.cgi?id=52002
See an example of this here:
https://bugs.webkit.org/show_bug.cgi?id=51314#c2
It's due to the fact that we pass around TestFailure() instance
objects and yet expect them to all compare to be the same.
This is bad. We should either have shared instances of these values
or we should just pass around the class names. There are already hacks
in place to make TestFailure objects pass equality tests, I just added
a __hash__ implementation so they can be used in sets as expected.
* Scripts/webkitpy/common/net/layouttestresults_unittest.py:
* Scripts/webkitpy/layout_tests/layout_package/test_failures.py:
2011-01-06 Evan Martin <evan@chromium.org>
Reviewed by Tony Chang.
[gdb] handle inaccessible memory and UChar*
https://bugs.webkit.org/show_bug.cgi?id=52003
To pretty-print a UChar*, we just grope around in the memory
looking for a terminating NUL. We need to handle the exception
that can occur when we poke into an invalid location.
Additionally, the logic for choosing which pretty-printer would
early return before hitting the point where we'd use the UChar*
printer. (Did this ever work? How did it regress?)
* gdb/webkit.py:
2011-01-05 Brian Weinstein <bweinstein@apple.com>
Reviewed by Ada Chan.
WebKit2: Should be able to call into injected bundle to ask if we should allow resource loads
https://bugs.webkit.org/show_bug.cgi?id=51969
Add shouldLoadResourceForFrame to WTR::InjectedBundlePage.
* WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
(WTR::InjectedBundlePage::InjectedBundlePage):
(WTR::InjectedBundlePage::shouldLoadResourceForFrame): Stub implementation.
* WebKitTestRunner/InjectedBundle/InjectedBundlePage.h:
2011-01-06 James Robinson <jamesr@chromium.org>
Reviewed by Simon Fraser.
Implement mozilla's animationTime property
https://bugs.webkit.org/show_bug.cgi?id=51952
Chromium DRT support for webkitAnimationTime.
* DumpRenderTree/chromium/WebViewHost.cpp:
(WebViewHost::paintInvalidatedRegion):
2011-01-06 Julie Parent <jparent@chromium.org>
Reviewed by Tony Chang.
Test Result Server always truncates number of results to JSON_RESULTS_MAX_BUILDS
https://bugs.webkit.org/show_bug.cgi?id=51217
* TestResultServer/model/jsonresults.py:
Pass num_runs through to _remove_items_over_max_number_of_builds, and use the value.
Adds missing documentation.
* TestResultServer/model/jsonresults_unittest.py:
Add unit test to test that truncation happens at smaller value than JSON_RESULTS_MAX_BUILDS.
Update test_merge to take a number of builds to truncate at.
2011-01-06 Adam Barth <abarth@webkit.org>
Reviewed by Eric Seidel.
The EWS should log outside of the working directory
https://bugs.webkit.org/show_bug.cgi?id=51986
This patch prepare us for using git clean -x -d -f in the EWS wrapper
script, which will actually clean the working copy back to a pristine
state instead of leaving untracked directories and ignored files.
* Scripts/webkitpy/tool/commands/queues.py:
* Scripts/webkitpy/tool/commands/queues_unittest.py:
2011-01-05 Steve Falkenburg <sfalken@apple.com>
Windows build fix.
Back out attempted change to built-product-archive.
* BuildSlaveSupport/built-product-archive:
2011-01-05 Steve Falkenburg <sfalken@apple.com>
Windows build fix.
Include configuration name when unpacking a build archive.
Needed, since we split Windows build results into configuration-specific directories
under WebKitOutputDir.
* BuildSlaveSupport/built-product-archive:
2011-01-05 Steve Falkenburg <sfalken@apple.com>
Reviewed by Darin Adler.
Debug and Release builds on Windows clobber each other
https://bugs.webkit.org/show_bug.cgi?id=49185
Changes the structure of WebKitBuild build products directory so we
completely separate each build configuration into independent directories.
Although we previously had per-configuration directories for obj, this change adds
per-configuration directories for bin, lib, obj, and include. Each configuration's
build products are stored within a directory inside of WebKitBuild.
Most use of $(WebKitOutputDir) in the build files has been replaced by $(ConfigurationBuildDir),
defined in common.vsprops to be $(WebKitOutputDir)\$(ConfigurationName).
* DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePluginCommon.vsprops:
* DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePluginPostBuild.cmd:
* DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePluginPreBuild.cmd:
* DumpRenderTree/win/DumpRenderTreeCommon.vsprops:
* DumpRenderTree/win/DumpRenderTreePostBuild.cmd:
* DumpRenderTree/win/DumpRenderTreePreBuild.cmd:
* DumpRenderTree/win/ImageDiffCommon.vsprops:
* DumpRenderTree/win/ImageDiffPostBuild.cmd:
* DumpRenderTree/win/ImageDiffPreBuild.cmd:
* FindSafari/FindSafari.vcproj:
* MiniBrowser/Configurations/MiniBrowserCommon.vsprops:
* MiniBrowser/MiniBrowserPostBuild.cmd:
* MiniBrowser/MiniBrowserPreBuild.cmd:
* Scripts/webkitdirs.pm:
* TestWebKitAPI/Configurations/TestWebKitAPICommon.vsprops:
* TestWebKitAPI/win/TestWebKitAPIGeneratedCommon.vsprops:
* TestWebKitAPI/win/TestWebKitAPIPostBuild.cmd:
* TestWebKitAPI/win/TestWebKitAPIPreBuild.cmd:
* TestWebKitAPI/win/copy-resources.cmd:
* WebKitAPITest/WebKitAPITestCommon.vsprops:
* WebKitAPITest/WebKitAPITestPostBuild.cmd:
* WebKitAPITest/WebKitAPITestPreBuild.cmd:
* WebKitTestRunner/Configurations/InjectedBundleCommon.vsprops:
* WebKitTestRunner/Configurations/WebKitTestRunnerCommon.vsprops:
* WebKitTestRunner/win/InjectedBundle.vcproj:
* WebKitTestRunner/win/InjectedBundleGeneratedCommon.vsprops:
* WebKitTestRunner/win/InjectedBundlePostBuild.cmd:
* WebKitTestRunner/win/InjectedBundlePreBuild.cmd:
* WebKitTestRunner/win/WebKitTestRunnerPostBuild.cmd:
* WebKitTestRunner/win/WebKitTestRunnerPreBuild.cmd:
* WinLauncher/WinLauncherCommon.vsprops:
* WinLauncher/WinLauncherPostBuild.cmd:
* WinLauncher/WinLauncherPreBuild.cmd:
2011-01-05 Jan Erik Hanssen <jhanssen@sencha.com>
Reviewed by Ariya Hidayat.
[Qt] TestNetscapePlugin doesn't link on OS X
https://bugs.webkit.org/show_bug.cgi?id=51948
Only add a dependency on libX11 on X11 systems
* DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro:
2011-01-05 Jeff Miller <jeffm@apple.com>
Reviewed by Adam Roben.
check-webkit-style should allow unnecessary parameter names in WebKit2 APIs because we're matching CF's header style
https://bugs.webkit.org/show_bug.cgi?id=51970
* Scripts/webkitpy/style/checker.py:
Add -readability/naming for WebKit2 C API directories.
2011-01-05 Eric Seidel <eric@webkit.org>
Unreviewed, just fixing one more log error from my previous python patch.
Move LayoutTestResults over to new-run-webkit-tests TestResult architecture
https://bugs.webkit.org/show_bug.cgi?id=51802
webkit-patch failure-reason was logging about parsing errors
due to my misunderstanding of how BeautifulSoup.findAll worked.
Fixed, and added a unit test.
* Scripts/webkitpy/common/net/layouttestresults.py:
* Scripts/webkitpy/common/net/layouttestresults_unittest.py:
2011-01-05 Benjamin Poulain <benjamin.poulain@nokia.com>
Reviewed by Kenneth Rohde Christiansen.
[Qt] [WK2] Add the possibility to open window from the WebProcess in Minibrowser
https://bugs.webkit.org/show_bug.cgi?id=51951
Set a function to QWKPage::setCreateNewPageFunction() in Minibrowser so there is a way
to create new window from WebKit.
* MiniBrowser/qt/BrowserWindow.cpp:
(newPageFunction):
(BrowserWindow::BrowserWindow):
(BrowserWindow::page): Added to avoid the indirect access to QWKPage spread in the class
(BrowserWindow::newWindow):
(BrowserWindow::updateUserAgentList):
(BrowserWindow::showUserAgentDialog):
* MiniBrowser/qt/BrowserWindow.h: Add the global variable backingStoreTypeForNewWindow to
define the type of the backing store used for new windows.
* MiniBrowser/qt/main.cpp:
(main):
2011-01-05 Eric Seidel <eric@webkit.org>
Reviewed by Ojan Vafai.
commit-queue should report failure type when reporting flaky tests
https://bugs.webkit.org/show_bug.cgi?id=51900
This patch was a ridiculous amount of plumbing. In the end I like
the NRWT-style TestResult class even less. But now we're passing
a list of TestResult objects from the queue to the FlakyTestReporter.
Thus the FlakyTestReporter can more easily report what type of failure
occurred.
In the process, I found that I was not alone in finding TestResult
cumbersome to use. A bunch of code was trying to create a TestResult
object with default values. However since the constructor didn't make
this easy, some places were getting it wrong (including the TestResult unit test)!
I've fixed the TestResult constructor to have default values for non-essential
arguments.
* Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
* Scripts/webkitpy/common/net/layouttestresults.py:
* Scripts/webkitpy/common/net/layouttestresults_unittest.py:
* Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
* Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
* Scripts/webkitpy/layout_tests/layout_package/printing_unittest.py:
* Scripts/webkitpy/layout_tests/layout_package/test_results.py:
* Scripts/webkitpy/layout_tests/layout_package/test_results_unittest.py:
* Scripts/webkitpy/layout_tests/layout_package/test_runner.py:
* Scripts/webkitpy/tool/bot/commitqueuetask.py:
* Scripts/webkitpy/tool/bot/commitqueuetask_unittest.py:
* Scripts/webkitpy/tool/bot/flakytestreporter.py:
* Scripts/webkitpy/tool/bot/flakytestreporter_unittest.py:
* Scripts/webkitpy/tool/commands/queues.py:
* Scripts/webkitpy/tool/commands/queues_unittest.py:
2011-01-05 Eric Seidel <eric@webkit.org>
Unreviewed.
Move LayoutTestResults over to new-run-webkit-tests TestResult architecture
https://bugs.webkit.org/show_bug.cgi?id=51802
Add another unit test to cover a previously missing import.
* Scripts/webkitpy/common/net/layouttestresults.py:
* Scripts/webkitpy/common/net/layouttestresults_unittest.py:
2011-01-05 Mihai Parparita <mihaip@chromium.org>
Reviewed by David Levin.
Add mihaip@chromium.org as a reviewer
https://bugs.webkit.org/show_bug.cgi?id=51966
Add myself as a reviewer.
* Scripts/webkitpy/common/config/committers.py:
2011-01-05 James Simonsen <simonjam@chromium.org>
Unreviewed. Adding myself to the list of committers.
* Scripts/webkitpy/common/config/committers.py:
2011-01-05 Eric Seidel <eric@webkit.org>
Unreviewed.
Move LayoutTestResults over to new-run-webkit-tests TestResult architecture
https://bugs.webkit.org/show_bug.cgi?id=51802
Fix two exceptions seen on the bots after my original commit.
* Scripts/webkitpy/common/net/layouttestresults.py:
* Scripts/webkitpy/common/net/layouttestresults_unittest.py:
2011-01-05 Kundu Suchismita <suchi.kundu@nokia.com>
Reviewed by Andreas Kling.
[Qt]Add local storage settings to QtTestBrowser command line arguments
https://bugs.webkit.org/show_bug.cgi?id=51750
* QtTestBrowser/main.cpp:
(LauncherApplication::handleUserOptions):
Local storage settings can be enable form command line arguments for
QtTestBrowser.
2011-01-05 Chris Fleizach <cfleizach@apple.com>
Reviewed by Martin Robinson.
GTK: AX: DRT needs to use correct root object method.
https://bugs.webkit.org/show_bug.cgi?id=51911
* DumpRenderTree/gtk/AccessibilityControllerGtk.cpp:
(AccessibilityController::rootElement):
2011-01-05 Gabor Rapcsanyi <rgabor@inf.u-szeged.hu>
Rubber-stamped by Csaba Osztrogonác.
Based on idea from Péter Gál.
* Scripts/webkitpy/tool/bot/sheriff.py: Remove unnecessary apostrophes around revision numbers.
2011-01-05 Gabor Rapcsanyi <rgabor@inf.u-szeged.hu>
Reviewed by Adam Barth.
sheriff-bot should be able to do multi-revision rollouts
https://bugs.webkit.org/show_bug.cgi?id=51176
* Scripts/webkitpy/tool/bot/irc_command.py:
* Scripts/webkitpy/tool/bot/sheriff.py:
* Scripts/webkitpy/tool/bot/sheriffircbot_unittest.py:
2011-01-05 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
queues.webkit.org should have per-bot status pages
https://bugs.webkit.org/show_bug.cgi?id=51906
Just adds a /queue-status/QUEUE_NAME/bots/BOT_ID page to show
statuses from a single queue. Now that we have many bots servicing
some queues (commit-queue, win-ews, etc.) its easy to have the messages
you care about scroll off the end of the 15-message limit on the main page.
Eventually we should probably rename /queue-status to /queue or /queues,
but that's fodder for another patch.
* QueueStatusServer/handlers/queuestatus.py:
* QueueStatusServer/index.yaml:
* QueueStatusServer/main.py:
* QueueStatusServer/templates/includes/singlequeuestatus.html:
- This probably should use a custom filter instead of hard-coding
the URL scheme here, but I couldn't figure out how to easily
create such a filter. Most filters work with "strings" so we can't
pass the status object. We could add a method to the status
object and call that, but that seemed a bit strange too.
* QueueStatusServer/templates/queuestatus.html:
2011-01-05 Carlos Garcia Campos <cgarcia@igalia.com>
Unreviewed. Adding myself to the list of committers.
* Scripts/webkitpy/common/config/committers.py:
2011-01-04 Koan-Sin Tan <koansin.tan@gmail.com>
Reviewed by David Levin.
check-webkit-style treated some macros with parentheses after #elif as function calls
https://bugs.webkit.org/show_bug.cgi?id=51695
Ingore function call space checking in any preprocessor directives
(things starting with #). Change search() to match() because
preprocessor directives are supposed to be in the beginning of lines.
* Scripts/webkitpy/style/checkers/cpp.py:
* Scripts/webkitpy/style/checkers/cpp_unittest.py: make sure there is no false positives for #elif cases
2011-01-04 Søren Gjesse <sgjesse@chromium.org>
Reviewed by Tony Chang.
Change the --multi-load test option to --stress-opt
https://bugs.webkit.org/show_bug.cgi?id=50751
* DumpRenderTree/chromium/DumpRenderTree.cpp:
(runTest):
(main):
* DumpRenderTree/chromium/TestShell.cpp:
(TestShell::TestShell):
* DumpRenderTree/chromium/TestShell.h:
(TestShell::stressOpt):
(TestShell::setStressOpt):
(TestShell::stressDeopt):
(TestShell::setStressDeopt):
(TestShell::javaScriptFlags):
(TestShell::setJavaScriptFlags):
* Scripts/webkitpy/layout_tests/port/chromium.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2011-01-04 Chris Fleizach <cfleizach@apple.com>
Reviewed by Sam Weinig.
WK2: Support Accessibility
https://bugs.webkit.org/show_bug.cgi?id=51859
Use rootObject() method to get top of accessibility tree.
* DumpRenderTree/mac/AccessibilityControllerMac.mm:
(AccessibilityController::focusedElement):
(AccessibilityController::rootElement):
2011-01-03 Martin Robinson <mrobinson@igalia.com>
Reviewed by Darin Adler.
Remove the last non-GObject usage of PlatformRefPtr and move the code to GRefPtr
https://bugs.webkit.org/show_bug.cgi?id=51846
* DumpRenderTree/gtk/AccessibilityUIElementGtk.cpp: Change usage of PlatformRefPtr
back to GRefPtr.
2011-01-04 Zhe Su <suzhe@chromium.org>
Reviewed by Kenneth Russell.
Fix insertText, setMarkedText and unmarkText methods of
TextInputController to call corresponding methods of WebKit::WebView
rather than WebKit::WebFrame. This change matches the behavior of
chromium browser.
insertText corresponds to WebView::confirmComposition(text)
setMarkedText corresponds to WebView::setComposition(...)
unmarkText corresponds to WebView::confirmComposition()
https://bugs.webkit.org/show_bug.cgi?id=51693
* DumpRenderTree/chromium/TextInputController.cpp:
(TextInputController::insertText):
(TextInputController::setMarkedText):
(TextInputController::unmarkText):
2011-01-04 Dihan Wickremasuriya <dihan.wickremasuriya@nokia.com>
Reviewed by Laszlo Gombos.
[Qt] [Symbian] Do not remove 0 byte sized files from productDir
On Symbian productDir points to the source directory. Removing all the
empty files from the productDir directory corrupts the svn repository.
* Scripts/build-webkit:
2011-01-03 Yi Shen <yi.4.shen@nokia.com>
Reviewed by Adam Barth.
[Qt] Add SelectAll option to the context menu for the editor
https://bugs.webkit.org/show_bug.cgi?id=50049
Set shortcut for the SelectAll action.
* QtTestBrowser/mainwindow.cpp:
(MainWindow::buildUI):
2011-01-03 David Levin <levin@chromium.org>
Reviewed by Eric Seidel.
check-webkit-style shouldn't consider "value" automatically to be a meaningless name.
https://bugs.webkit.org/show_bug.cgi?id=51842
* Scripts/webkitpy/style/checkers/cpp.py: Remove the check for value.
* Scripts/webkitpy/style/checkers/cpp_unittest.py: Remove the test for value.
2011-01-03 Patrick Gansterer <paroga@webkit.org>
Reviewed by David Kilzer.
[WINCE] Adds a build slave.
https://bugs.webkit.org/show_bug.cgi?id=50523
* BuildSlaveSupport/build.webkit.org-config/config.json: Add an
entry for a release build of the WinCE port.
* BuildSlaveSupport/build.webkit.org-config/master.cfg: Add platform
flag to build the WinCE port.
2011-01-03 Patrick Gansterer <paroga@webkit.org>
Reviewed by David Kilzer.
Add WinCE support to build-webkit
https://bugs.webkit.org/show_bug.cgi?id=51642
* Scripts/build-webkit:
* Scripts/webkitdirs.pm:
2011-01-03 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
Move LayoutTestResults over to new-run-webkit-tests TestResult architecture
https://bugs.webkit.org/show_bug.cgi?id=51802
I'm not the biggest fan of the test_failures or test_results classes,
but it's better to have one shared set of classes, than separate ones
for new vs. old run-webkit-test result handling.
This moves the ORWT results class "LayoutTestResults" over to using
TestResult and TestFailure classes, making it easy for us to
make all our sheriff-bot and other webkitpy code NRWT ready.
This also makes it a trivial patch to generate results.json information
from ORWT results.html files (for flaky test analysis, etc.) as well
as making it a one-liner to report test failure types when the
commit-queue sees flaky tests.
This patch tried not to add new functionality, but only to replace
the guts of LayoutTestResults, while adding unit tests and hoping
not to break anything.
I also moved callers which assumed User.prompt* were static/class methods
to using them as instance methods (since we'll eventually want to make them such).
In the process of re-writing things, I broke the rebaseline command, so I wrote
a unit test to catch my breakage were I do do so again in the future.
* Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
* Scripts/webkitpy/common/net/layouttestresults.py:
* Scripts/webkitpy/common/net/layouttestresults_unittest.py:
* Scripts/webkitpy/layout_tests/layout_package/test_failures.py:
* Scripts/webkitpy/layout_tests/layout_package/test_results.py:
* Scripts/webkitpy/tool/commands/queries.py:
* Scripts/webkitpy/tool/commands/rebaseline.py:
* Scripts/webkitpy/tool/commands/rebaseline_unittest.py:
* Scripts/webkitpy/tool/mocktool.py:
2011-01-03 Adam Barth <abarth@webkit.org>
Reviewed by Simon Fraser.
webkit-patch shouldn't waste time cleaning a working directory that's already clean
https://bugs.webkit.org/show_bug.cgi?id=51840
* Scripts/webkitpy/common/checkout/scm.py:
2011-01-03 Adam Barth <abarth@webkit.org>
Reviewed by Simon Fraser.
webkit-patch should only hide update output if --quiet
https://bugs.webkit.org/show_bug.cgi?id=51838
As requested by smfr.
* Scripts/webkitpy/tool/steps/update.py:
2011-01-03 Adam Barth <abarth@webkit.org>
Reviewed by David Levin.
commit-queue shouldn't reject patches twice if two bots process the same patch
https://bugs.webkit.org/show_bug.cgi?id=51805
Before raising an error, we first check whether the patch is still in
the commit-queue. This check is still racy, of course, but the time
window is much smaller.
* Scripts/webkitpy/tool/bot/commitqueuetask.py:
2011-01-03 Sam Weinig <sam@webkit.org>
Reviewed by Anders Carlsson.
Make run-api-tests less chatty.
https://bugs.webkit.org/show_bug.cgi?id=51831
- Make script quiet by default and add --verbose option (replacing --quiet).
- When not verbose, pipe stdout and stderr to devnull.
* Scripts/run-api-tests:
2011-01-03 Pratik Solanki <psolanki@apple.com>
Unreviewed. Adding myself to committers.py.
* Scripts/webkitpy/common/config/committers.py:
2011-01-03 Koan-Sin Tan <koansin.tan@gmail.com>
Reviewed by Xan Lopez.
"Tool/Scripts/run-launcher --gtk" stopped working after 74855
https://bugs.webkit.org/show_bug.cgi?id=51806
Originally, the $libraryName is 'JavaScriptCore'. When building release
one, the $libraryDir was 'WebKitBuild/Release/JavaScriptCore/../.libs/'
which doesn't exist anymore ('WebKitBuild/Release/JavaScriptCore'
is 'WebKitBuild/Release/Source/JavaScriptCore' now), and
'WebKitBuild/Release/Source/JavaScriptCore/../../.libs' looks weird,
so make it 'WebKitBuild/Release/.libs'
* Scripts/webkitdirs.pm:
2011-01-02 Patrick Gansterer <paroga@webkit.org>
Reviewed by Adam Barth.
Make EFL a core builder
https://bugs.webkit.org/show_bug.cgi?id=51804
* Scripts/webkitpy/common/net/buildbot/buildbot.py:
* Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
2011-01-02 Csaba Osztrogonác <ossy@webkit.org>
[Qt] Unreviewed fix after r74855 and r74875.
* BuildSlaveSupport/built-product-archive:
- Create directories explicitly.
- Pass shell=True to subprocess.call().
2011-01-02 Csaba Osztrogonác <ossy@webkit.org>
[Qt] Unreviewed fix after r74855.
Move JavaScriptCore to Source
https://bugs.webkit.org/show_bug.cgi?id=51604
* BuildSlaveSupport/built-product-archive: JavaScriptCore -> Source/JavaScriptCore
2011-01-02 Csaba Osztrogonác <ossy@webkit.org>
[Qt] Unreviewed run-javasriptcore-tests fix after r74855.
Move JavaScriptCore to Source
https://bugs.webkit.org/show_bug.cgi?id=51604
* Scripts/webkitdirs.pm:
2011-01-02 Robert Hogan <robert@webkit.org>
Unreviewed, build fix.
[Qt] Fix DRT build after 74855
* DumpRenderTree/qt/DumpRenderTree.pro:
2011-01-01 Adam Barth <abarth@webkit.org>
Remove the assumption from the Qt and Gtk builds that every project is
in the root directory.
* Scripts/build-webkit:
2011-01-01 Adam Barth <abarth@webkit.org>
Reviewed by Eric Seidel.
Teach svn-apply how to apply patches even though JavaScriptCore has moved
https://bugs.webkit.org/show_bug.cgi?id=51796
After this change, svn-apply will magically apply patches to
JavaScriptCore/foo to Source/JavaScriptCore/foo.
* Scripts/VCSUtils.pm:
2011-01-01 Adam Barth <abarth@webkit.org>
Reviewed by Eric Seidel.
Move JavaScriptCore to Source
https://bugs.webkit.org/show_bug.cgi?id=51604
Update references to JavaScriptCore to point to the new location.
* BuildSlaveSupport/build.webkit.org-config/master.cfg:
* DumpRenderTree/qt/DumpRenderTree.pro:
* DumpRenderTree/qt/ImageDiff.pro:
* DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro:
* DumpRenderTree/wscript:
* Scripts/build-jsc:
* Scripts/build-webkit:
* Scripts/do-file-rename:
* Scripts/do-webcore-rename:
* Scripts/run-javascriptcore-tests:
* Scripts/update-javascriptcore-test-results:
* Scripts/webkitdirs.pm:
* Scripts/webkitpy/common/config/build_unittest.py:
* Scripts/webkitpy/style/checker.py:
* Scripts/webkitpy/style/checker_unittest.py:
* Scripts/webkitpy/style/checkers/cpp_unittest.py:
* WebKitTestRunner/InjectedBundle/qt/InjectedBundle.pro:
* WebKitTestRunner/qt/WebKitTestRunner.pro:
* wx/build/settings.py:
2011-01-01 Adam Barth <abarth@webkit.org>
Reviewed by Eric Seidel.
Move Sources to Source
https://bugs.webkit.org/show_bug.cgi?id=51794
Update scripts to point to the new location.
* Scripts/build-webkit:
* Scripts/do-file-rename:
* Scripts/do-webcore-rename:
* Scripts/webkitpy/common/config/build.py:
* Scripts/webkitpy/common/config/build_unittest.py:
2010-12-31 Kent Tamura <tkent@chromium.org>
Reviewed by Eric Seidel.
[DRT/Chromium] Enable mock spell checker on non-OSX
https://bugs.webkit.org/show_bug.cgi?id=51401
* DumpRenderTree/chromium/WebViewHost.cpp:
(WebViewHost::spellCheck):
2010-12-31 Adam Barth <abarth@webkit.org>
Reviewed by Ariya Hidayat.
Remove Tools/Scripts/wkstyle
https://bugs.webkit.org/show_bug.cgi?id=51774
This script appears to not have been touched in a while and seems to
have been replaced by check-webkit-style.
* Scripts/wkstyle: Removed.
2010-12-31 Ilya Tikhonovsky <loislo@chromium.org>
Unreviewed. Coding style fix.
* DumpRenderTree/chromium/TestShell.h:
2010-12-31 Ilya Tikhonovsky <loislo@chromium.org>
Reviewed by Pavel Feldman.
Web Inspector: [Chromium] Inspector's tests are very slow on Win(Dbg).
It was a problem with Chromium DRT. DevToolsAgent object instance was
attached not only to the inspected page but to the DevTools window too.
As result all the inspector files were interpreted as a content of inspected page
and were transfered to the DevTools scripts panel for debugging etc.
https://bugs.webkit.org/show_bug.cgi?id=51735
* DumpRenderTree/chromium/TestShell.cpp:
(TestShell::TestShell):
(TestShell::bindJSObjectsToWindow):
(TestShell::createNewWindow):
* DumpRenderTree/chromium/TestShell.h:
2010-12-31 Adam Barth <abarth@webkit.org>
Rubber-stamped by Eric Seidel.
Move PageLoadTests to PerformanceTests/PageLoad
https://bugs.webkit.org/show_bug.cgi?id=51771
Update references to PageLoadTests to point to the new location.
* Scripts/run-pageloadtest:
* Scripts/webkitpy/common/config/build.py:
2010-12-31 Adam Barth <abarth@webkit.org>
Rubber-stamped by Eric Seidel.
Move SunSpider into PerformanceTests
https://bugs.webkit.org/show_bug.cgi?id=51769
Update these scripts to point to the new location.
* Scripts/run-sunspider:
* Scripts/sunspider-compare-results:
2010-12-30 Mihai Parparita <mihaip@chromium.org>
Reviewed by Kent Tamura.
[Chromium] Add WebThemeEngineDRTMac so that Chromium DRT scrollbar rendering can match the Mac port's
https://bugs.webkit.org/show_bug.cgi?id=51728
Add implementation of the Mac WebThemeEngine that uses an NSScroller
to render top-level scrollbars. This makes them match the Mac port's
use of an NSScrollView, which means that we'll be able to share more
pixel baselines.
The new rendering code will not be activated until the
USE_WEB_THEME_ENGINE_TO_PAINT_THUMB #define is fliped in
ScrollbarThemeChromiumMac.
* DumpRenderTree/DumpRenderTree.gypi:
* DumpRenderTree/chromium/TestShellMac.mm:
(platformInit):
* DumpRenderTree/chromium/WebThemeEngineDRTMac.h: Added.
* DumpRenderTree/chromium/WebThemeEngineDRTMac.mm: Added.
(+[FakeActiveWindow alwaysActiveWindow]):
(+[FakeActiveWindow alwaysInactiveWindow]):
(-[FakeActiveWindow initWithActiveControls:]):
(-[FakeActiveWindow _hasActiveControls]):
(WebThemeEngineDRTMac::paintScrollbarThumb):
(stateToHIEnableState):
(WebThemeEngineDRTMac::paintHIThemeScrollbarThumb):
(WebThemeEngineDRTMac::paintNSScrollerScrollbarThumb):
2010-12-30 Konstantin Tokarev <annulen@yandex.ru>
Reviewed by David Kilzer.
[Qt] Don't build wtf/TCSystemAlloc.cpp if --system-malloc option is
used
https://bugs.webkit.org/show_bug.cgi?id=51672
* DumpRenderTree/qt/DumpRenderTree.pro: Replaced USE_SYSTEM_MALLOC
with USE_SYSTEM_MALLOC=1
* WebKitTestRunner/qt/WebKitTestRunner.pro: Replaced USE_SYSTEM_MALLOC
with USE_SYSTEM_MALLOC=1
2010-12-30 Laszlo Gombos <laszlo.1.gombos@nokia.com>
Reviewed by David Kilzer.
[Qt] [Symbian] Fix build-webkit script for Symbian
https://bugs.webkit.org/show_bug.cgi?id=51509
Set the OUTPUT_DIR for Symbian to be the same as the source
directory.
* Scripts/webkitdirs.pm:
2010-12-30 Martin Robinson <mrobinson@igalia.com>
Reviewed by Darin Adler.
check-webkit-style should ignore NULL usage in calls to gtk_widget_style_get
https://bugs.webkit.org/show_bug.cgi?id=51758
Add a check-webkit-style exception for gtk_widget_style_get and NULL usage.
* Scripts/webkitpy/style/checkers/cpp.py: Add the exception.
* Scripts/webkitpy/style/checkers/cpp_unittest.py: Add some tests for this.
2010-12-29 Patrick Gansterer <paroga@webkit.org>
Unreviewed WinCE buildfix.
* WinCELauncher/main.cpp: Add missing include.
2010-12-29 Daniel Bates <dbates@rim.com>
Reviewed by Darin Adler.
svn-apply updates date of wrong change log entry for a change log diff that
contains two consecutive entries with the same author and date
https://bugs.webkit.org/show_bug.cgi?id=46061
Fixes an issue where the date of the wrong change log entry may be modified.
Moreover, changes fixChangeLogPatch() to move entries inserted earlier
in a ChangeLog file to the top of the file.
Currently, fixChangeLogPatch() explicitly bails out and returns an unchanged
diff when it detects that the diff inserts a change log entry earlier in a
ChangeLog. It is unusual to land a patch that has such a deliberate ChangeLog
change. With the advent of the commit-queue this functionality of bailing out
and hence landing the patch as-is is harmful to the accuracy of the ChangeLog.
Instead, we should always move the change log entry to the top of the ChangeLog file.
A side-effect of this change is that setChangeLogDateAndReviewer() now updates
the date line of the correct change log entry in a ChangeLog diff.
* Scripts/VCSUtils.pm: Modified fixChangeLogPatch() to move entries inserted earlier to the top.
* Scripts/webkitperl/VCSUtils_unittest/fixChangeLogPatch.pl:
Updated the following unit tests now that we move entries inserted earlier to the top:
- "fixChangeLogPatch: New entry inserted in middle."
(formerly named "fixChangeLogPatch: [no change] New entry inserted in middle.")
- "fixChangeLogPatch: New entry inserted earlier in the file, but after an entry with the same author and date."
(formerly named "fixChangeLogPatch: [no change] New entry inserted earlier in the file, but after an entry with the same author and date.")
* Scripts/webkitperl/VCSUtils_unittest/fixChangeLogPatchThenSetChangeLogDateAndReviewer.pl: Added.
2010-12-29 Konstantin Tokarev <annulen@yandex.ru>
Reviewed by Eric Seidel.
[Qt] Fixed compatibility with gold linker on X11 platforms
https://bugs.webkit.org/show_bug.cgi?id=51700
* DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro:
Linked TestNetscapePlugin with libX11 on Unix platforms
2010-12-29 Ademar de Souza Reis Jr <ademar.reis@openbossa.org>
Reviewed by Darin Adler.
Circular dependency in webkitpy.common.checkout.changelog module
https://bugs.webkit.org/show_bug.cgi?id=50475
Remove automatic import of api.Checkout module when any checkout/
submodule is imported (e.g.: when checkout.scm.Git is imported).
* Scripts/webkitpy/common/checkout/__init__.py:
2010-12-29 Martin Robinson <mrobinson@igalia.com>
Reviewed by Ojan Vafai.
prepare-ChangeLog should support the -g option for specifying the git-commit
https://bugs.webkit.org/show_bug.cgi?id=51708
Add a -g shortcut for the --git-commit prepare-ChangeLog option. This matches
webkit-patch, though the exact meaning of the two options remain different.
* Scripts/prepare-ChangeLog: Add -g.
2010-12-29 Kent Hansen <kent.hansen@nokia.com>
Reviewed by Simon Hausmann.
[Qt] Fix compilation with Qt in namespace
https://bugs.webkit.org/show_bug.cgi?id=51701
* QtTestBrowser/webview.h:
2010-12-28 Daniel Bates <dbates@rim.com>
Reviewed by Sam Weinig.
Substitute // MARK: for compiler-specific #pragma mark
https://bugs.webkit.org/show_bug.cgi?id=51657
Fix future compilation warnings about "#pragma mark" on GTK+ bots by
substituting "// MARK:" for "#pragma mark", which provides
analogous code-bookmarking functionality under Xcode.
Also, for consistency, we should substitute "// MARK:" for compiler-
specific "#pragma mark" in the source files for the Mac port.
* DumpRenderTree/AccessibilityTextMarker.cpp:
* DumpRenderTree/cf/WebArchiveDumpSupport.h:
* DumpRenderTree/mac/AccessibilityTextMarkerMac.mm:
* DumpRenderTree/mac/ObjCController.m:
* MiniBrowser/mac/AppDelegate.m:
* MiniBrowser/mac/BrowserWindowController.m:
2010-12-28 Xan Lopez <xlopez@igalia.com>
Reviewed by Eric Seidel.
Add --no-build option to run-sunspider
https://bugs.webkit.org/show_bug.cgi?id=51658
* Scripts/run-sunspider: add --no-build flag to skip building
phase.
2010-12-27 Daniel Bates <dbates@rim.com>
Rubber-stamped by Martin Robinson.
Append the directory Tools/DumpRenderTree to the list of directories that
do-webcore-rename should search so that it can rename DRT files and
source code.
* Scripts/do-webcore-rename:
2010-12-27 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
build-webkit should detect missing Java SDK and explain how to install it
https://bugs.webkit.org/show_bug.cgi?id=51651
* Scripts/build-webkit:
2010-12-26 Martin Robinson <mrobinson@igalia.com>
Reviewed by Daniel Bates.
[GTK] Add support for build-jsc
https://bugs.webkit.org/show_bug.cgi?id=51625
Add GTK+ support to build-jsc now that building the jsc executable as
a standalone target is easy.
* Scripts/build-jsc: Call buildGtkProject for the GTK+ case now.
* Scripts/build-webkit: Pass in the proper target name to buildGtkProject.
* Scripts/webkitdirs.pm: Actually process the project parameter in buildGtkProject
and pass it along to buildAutotoolsProject.
2010-12-27 David Levin <levin@chromium.org>
Reviewed by Eric Seidel.
check-webkit-style check for meaningless variable names in function declarations.
https://bugs.webkit.org/show_bug.cgi?id=51523
* Scripts/webkitpy/style/checker.py: Exempted JavaScriptCore/jit/JITStubs.cpp
from the new check and whitespace/parens because the syntax is unusual and
produced a fair number of positives for these checks.
* Scripts/webkitpy/style/checkers/cpp.py:
(_convert_to_lower_with_underscores): Used as a canonical form for type names
and parameter names when determining if the parameter name is useless.
(_create_acronym): Used to check for redundant variable names in cases like "ExceptionCode ec"
(Parameter.lower_with_underscores_name): Gives back the parameter name in a lower_with_underscore
format.
(_check_parameter_name_against_text): Checks to see if the parameter name is in the
text or an acronym of it.
(check_function_definition): Checks function definitions for meaningless variable names.
(process_line): Added call to check_function_definition.
* Scripts/webkitpy/style/checkers/cpp_unittest.py:
(CppFunctionsTest.test_convert_to_lower_with_underscores): Test for _convert_to_lower_with_underscores.
(CppFunctionsTest.test_create_acronym): Test for _create_acronym.
(CppFunctionsTest.test_check_parameter_against_text): Test for _check_parameter_against_text.
(CppStyleTestBase.perform_single_line_lint): Removed the parameter name check
because when only checking a snippet, there are a lot of bogus functions.
(CppStyleTestBase.perform_multi_line_lint): Ditto and removed a bogus filter
that I put there previously and just noticed.
(WebKitStyleTest.test_parameter_names): Tests for the functionality -- both
check_function_definition and process_line.
2010-12-27 Carlos Garcia Campos <cgarcia@igalia.com>
Reviewed by Martin Robinson.
[GTK] Disable flash plugin in GtkLauncher when using gtk3
https://bugs.webkit.org/show_bug.cgi?id=51591
Flash plugin uses gtk2 that is incompatible with gtk3 making
webkit crash when it's built with gtk3.
* GtkLauncher/main.c:
(main):
2010-12-25 Adam Barth <abarth@webkit.org>
Reviewed by Eric Seidel.
Move ANGLE to Sources/ThirdParty
https://bugs.webkit.org/show_bug.cgi?id=51605
* Scripts/build-webkit:
- Point to the new ANGLE location.
2010-12-26 David Levin <levin@chromium.org>
Reviewed by Eric Seidel.
check-webkit-style should be able to parse function declaration parameters.
https://bugs.webkit.org/show_bug.cgi?id=51451
* Scripts/webkitpy/style/checkers/cpp.py:
(Position.__init__): Holds simple position information (row, column).
(Parameter.__init__): Holds information about a parameter.
(SingleLineView.__init): Converts multiple lines into a single line for simpler searches.
(SingleLineView.convert_column_to_row): Returns the original row given a column.
(create_skeleton_parameters): Simplifies a parameter list for easier searching.
(find_parameter_name_index): Finds where the parameter name is.
(parameter_list): Generates the list of parameters for a function.
(_FunctionState.begin): Added information to allow determining the parameters
on demand.
(_FunctionState.get_parameter_list): Returns a tuple of function parameters.
(detect_functions): Improve function detection for operator functions and
determine where the parameters end and pass that to _FunctionState.begin.
* Scripts/webkitpy/style/checkers/cpp_unittest.py:
(CppFunctionsTest.test_parameter): Verifies Parameter functionality.
(CppFunctionsTest.test_single_line_view): Verifies SingleLineView functionality.
(CppFunctionsTest.test_create_skeleton_parameters): Verifies create_skeleton_parameters.
(CppFunctionsTest.test_find_parameter_name_index): Verifies find_parameter_name_index.
(CppFunctionsTest.test_parameter_list): Does some minimal verification for parameter list.
Much more thorough verification is done as part of FunctionDetectionTest.test_parameter_list.
(FunctionDetectionTest.perform_function_detection): Added support for verifying
the parameters found.
(FunctionDetectionTest.test_function_declaration_detection): Added more function
detection test to verify that we catch the operator functions.
(FunctionDetectionTest.test_ignore_macros): Verify that function detection ignores macros.
(FunctionDetectionTest.test_parameter_list): Added tests to verify the parameter parsing.
(CheckForFunctionLengthsTest.test_function_length_check_definition_severity1_for_bad_test_doesnt_break): Removed
because the error. The test is about the bad function name. Fixing the name makes the test exactly like
test_function_length_check_definition_severity1_for_test.
2010-12-25 Patrick Gansterer <paroga@webkit.org>
Unreviewed WinCE buildfix after r74334.
* CMakeListsWinCE.txt:
2010-12-24 Sheriff Bot <webkit.review.bot@gmail.com>
Unreviewed, rolling out r74632.
http://trac.webkit.org/changeset/74632
https://bugs.webkit.org/show_bug.cgi?id=51600
makes test-webkitpy hang forever on GTK (Requested by philn-tp
on #webkit).
* Scripts/webkitpy/layout_tests/layout_package/message_broker.py:
* Scripts/webkitpy/layout_tests/layout_package/message_broker_unittest.py:
2010-12-24 Ryuan Choi <ryuan.choi@samsung.com>
Reviewed by Kenneth Rohde Christiansen.
[EFL] Screen doesn't be rendered when changing size of EWebLauncher
https://bugs.webkit.org/show_bug.cgi?id=43528
Remove on_resize because ewk_view_fixed_layout_size_set should not be
called on callback of EVAS_CALLBACK_RESIZE. As it change size of
FrameView, rendering was ignored when _ewk_view_smart_calculate was
called.
* EWebLauncher/main.c:
(browserCreate):
2010-12-24 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
webkit-patch (or a pre-commit hook) needs to prevent bad ChangeLog changes
https://bugs.webkit.org/show_bug.cgi?id=28291
This is a start. At least now webkit-patch will prompt when your ChangeLog looks questionable.
We could do more advanced things, like parsing the ChangeLog (with changelog.py) and comparing that
to strings with find in the diff.
Since non-interactive always returns the default, this should cause patches with bad changelogs to fail on the commit-queue.
* Scripts/webkitpy/common/checkout/api.py:
* Scripts/webkitpy/common/checkout/diff_parser.py:
* Scripts/webkitpy/tool/steps/abstractstep.py:
* Scripts/webkitpy/tool/steps/cleanworkingdirectory.py:
* Scripts/webkitpy/tool/steps/validatechangelogs.py: Copied from Tools/Scripts/webkitpy/tool/steps/validatereviewer.py.
* Scripts/webkitpy/tool/steps/validatechangelogs_unittest.py: Copied from Tools/Scripts/webkitpy/tool/steps/cleanworkingdirectory.py.
* Scripts/webkitpy/tool/steps/validatereviewer.py:
2010-12-24 Dirk Pranke <dpranke@chromium.org>
Reviewed by Kenneth Russell.
Modify new-run-webkit-tests to shut up and exit when all the threads
are wedged so that test runs complete faster and we don't end up
w/ enormous log files.
https://bugs.webkit.org/show_bug.cgi?id=51572
* Scripts/webkitpy/layout_tests/layout_package/message_broker.py:
* Scripts/webkitpy/layout_tests/layout_package/message_broker_unittest.py:
2010-12-24 Eric Seidel <eric@webkit.org>
Reviewed by David Levin.
commit-queue should better explain how it will re-open bugs and follow duplicate chains
https://bugs.webkit.org/show_bug.cgi?id=51549
I've had multiple people ask me about the commit-queue's behavior
regarding reporting flaky tests. I figured it would be best to just
have the queue explain itself in the bugs it files.
* Scripts/webkitpy/tool/bot/flakytestreporter.py:
* Scripts/webkitpy/tool/bot/flakytestreporter_unittest.py:
2010-12-24 Shinichiro Hamaji <hamaji@chromium.org>
Reviewed by Eric Seidel.
check-webkit-style: should ignore "wrong" namings in JavaScriptCore/qt
https://bugs.webkit.org/show_bug.cgi?id=51467
* Scripts/webkitpy/style/checker.py:
* Scripts/webkitpy/style/checker_unittest.py:
2010-12-23 Adam Barth <abarth@webkit.org>
Rubber-stamped by Eric Seidel.
Move JavaScriptGlue into Sources
https://bugs.webkit.org/show_bug.cgi?id=51583
These tools refer to JavaScriptGlue and need to be updated to its new
location.
* Scripts/build-webkit:
* Scripts/do-file-rename:
* Scripts/do-webcore-rename:
* Scripts/webkitpy/common/config/build.py:
* Scripts/webkitpy/common/config/build_unittest.py:
2010-12-23 Lucas Forschler <lforschler@apple.com>
Reviewed by Eric Seidel.
Bug 51539 - update iexploder scripts to point to new location
* Scripts/run-iexploder-tests:
* Scripts/update-iexploder-cssproperties:
2010-12-23 Evan Martin <evan@chromium.org>
Reviewed by Ryosuke Niwa.
[chromium] LayoutTestController warning in float->int conversion
https://bugs.webkit.org/show_bug.cgi?id=51553
Rather than reading an int, converting to a float, then getting a compiler
warning on passing the float to something that expects an int, instead
just pass around ints.
No tests, fixes a compiler warning.
* DumpRenderTree/chromium/LayoutTestController.cpp:
(parsePageSizeParameters):
(LayoutTestController::pageNumberForElementById):
(LayoutTestController::numberOfPages):
2010-12-23 Lucas Forschler <lforschler@apple.com>
Reviewed by Darin Adler.
Bug 51537 - iexploder needs cleanup
* iExploder/htdocs: Removed.
* iExploder/htdocs/config.rb: Removed.
* iExploder/htdocs/cssproperties.in: Removed.
* iExploder/htdocs/cssvalues.in: Removed.
* iExploder/htdocs/htmlattrs.in: Removed.
* iExploder/htdocs/htmltags.in: Removed.
* iExploder/htdocs/htmlvalues.in: Removed.
* iExploder/htdocs/iexploder.cgi: Removed.
* iExploder/htdocs/iexploder.rb: Removed.
* iExploder/htdocs/index.html: Removed.
* iExploder/htdocs/webserver.rb: Removed.
* iExploder/iexploder-1.7.2/output: Removed.
* iExploder/tools: Removed.
* iExploder/tools/lasthit.rb: Removed.
* iExploder/tools/osx_last_crash.rb: Removed.
* iExploder/tools/showtest.rb: Removed.
2010-12-23 Lucas De Marchi <lucas.demarchi@profusion.mobi>
Reviewed by Kenneth Rohde Christiansen.
[EFL] Make 'single' the default backing store in EWebLauncher
https://bugs.webkit.org/show_bug.cgi?id=51534
'single' is the simplest backing store and the one intended to be the
default in the test browser. By mistake the tiled backing store was
left as default in r72617.
* EWebLauncher/main.c:
(browserCreate): ditto.
2010-12-23 Lucas De Marchi <lucas.demarchi@profusion.mobi>
Reviewed by Kenneth Rohde Christiansen.
[EFL] Subject: [webkit 4/4] [EFL] Remove non-sense workaround
https://bugs.webkit.org/show_bug.cgi?id=51533
There's no point in concatenating env("HOME") with "blah", so remove
option from EWebLauncher.
* EWebLauncher/main.c:
(main): ditto.
2010-12-23 Ilya Tikhonovsky <loislo@chromium.org>
Reviewed by Eric Seidel.
Web Inspector: chromium: inspector's tests are quite slow especially at windows(Debug).
Almost all inspector's tests become flaky on chromium Win(dbg) after switching to DRT.
I just discovered that we have two additional GC runs per each WebViewHost.
Test_shell had no such calls.
These gc runs eat 15% of cpu time.
https://bugs.webkit.org/show_bug.cgi?id=51479
* DumpRenderTree/chromium/WebViewHost.cpp:
(WebViewHost::~WebViewHost):
2010-12-22 Dirk Pranke <dpranke@chromium.org>
Reviewed by Ojan Vafai.
This change splits out the TestRunner class into its own file.
When we fork the code for message passing, we will create a new
TestRunner, and this change will allow that to be more modular.
While we're at it, split out TestInput and ResultSummary into
their own files, to reduce the size of run_webkit_tests to
something more managable.
https://bugs.webkit.org/show_bug.cgi?id=51092
* Scripts/webkitpy/layout_tests/layout_package/printing_unittest.py:
* Scripts/webkitpy/layout_tests/layout_package/result_summary.py:
* Scripts/webkitpy/layout_tests/layout_package/test_runner.py:
* Scripts/webkitpy/layout_tests/layout_package/test_runner_unittest.py:
* Scripts/webkitpy/layout_tests/layout_package/test_input.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2010-12-22 Dirk Pranke <dpranke@chromium.org>
Reviewed by Ojan Vafai.
nrwt multiprocessing - start over, prepare to fork the code
This code cleans up the signatures and implementation of the
TestRunner class so we can easily fork it to run either the
stable implementation or the new, unstable message-passing
implementation. The two variants will have different
implementations of the run_tests() method. We will switch
between the two based on the setting for the '--worker-model'
switch. We rename the two currently valid values to 'old-inline'
and 'old-threads'.
https://bugs.webkit.org/show_bug.cgi?id=51081
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2010-12-22 Victor Wang <victorw@chromium.org>
Reviewed by Darin Fisher.
[Chromium] Update chromium archive test result url in rebaseline script.
https://bugs.webkit.org/show_bug.cgi?id=51503
* Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
2010-12-22 Lucas Forschler <lforschler@apple.com>
Reviewed by Alice Liu.
<rdar://problem/8633222> record-memory-win needs to record memory used by the webprocess.
Verified on Chrome, Safari, and IE.
With this change, we will record memory from a parent browser window and all child processes.
* record-memory-win/main.cpp:
(ProcessArgs):
(PrintUsage):
(getMemoryInfo):
(printProcessInfo):
(evalProcesses):
(UseImage):
(QueryContinuously):
(ElapsedTime):
2010-12-22 Lucas Forschler <lforschler@apple.com>
Unreviewed rollout r74489, because it was missing changelog.
* record-memory-win/main.cpp:
2010-12-22 Ryosuke Niwa <rniwa@webkit.org>
Adding myself as a reviewer.
* Scripts/webkitpy/common/config/committers.py:
2010-12-22 Andrew Scherkus <scherkus@chromium.org>
Unreviewed. Adding myself to committers.py.
* Scripts/webkitpy/common/config/committers.py:
2010-12-22 Ariya Hidayat <ariya@sencha.com>
Reviewed by Kenneth Rohde Christiansen.
TestNetscapePlugIn should use #if defined
https://bugs.webkit.org/show_bug.cgi?id=51471
* DumpRenderTree/TestNetscapePlugIn/main.cpp:
(NP_Initialize):
(NPP_New):
2010-12-22 Ilya Tikhonovsky <loislo@chromium.org>
Reviewed by Yury Semikhatsky.
Web Inspector: [chromium] DRT inspector layout tests are flaky in debug.
DevTools window should be closed explicitly because it has custom deinitialization code.
https://bugs.webkit.org/show_bug.cgi?id=50722
* DumpRenderTree/chromium/TestShell.cpp:
(TestShell::closeRemainingWindows):
2010-12-22 David Levin <levin@chromium.org>
Reviewed by Shinichiro Hamaji.
check-webkit-style shouldn't complaint about underscores in variables in objective C files.
https://bugs.webkit.org/show_bug.cgi?id=51452
* Scripts/webkitpy/style/checkers/cpp.py:
(_FileState.__init__): Added the information to determine if a file is C or Objective C.
Using the file extension if possible but falling back to the file contents if we have a header file.
(_FileState.is_objective_c): Determine if we have an Objective C by examining the file contents if needed.
(_FileState.is_c_or_objective_c):
(check_using_std): Changed to using _FileState to determine the file type.
(check_max_min_macros): Ditto.
(check_for_null): Ditto.
(check_style): Changed the parameters to various calls since they now need _FileState
to determine the file type.
(check_language): Added the file_state parameter so it could be passed
to check_identifier_name_in_declaration.
(check_identifier_name_in_declaration): Don't warn about underscores in variables if
this is an Objective C file.
(_process_lines): Added information for the _FileState constructor (and moved the
call to a place that had the information).
* Scripts/webkitpy/style/checkers/cpp_unittest.py:
(CppFunctionsTest.test_is_c_or_objective_c): Changed the tests to use FileState and exercise
its functionality.
(WebKitStyleTest.test_names): Add tests for underscores in Objective C files.
2010-12-21 Andy Estes <aestes@apple.com>
Reviewed by Mark Rowe.
Do not build non-Intel architectures for WebKit2-related projects.
https://bugs.webkit.org/show_bug.cgi?id=51440
* MiniBrowser/Configurations/Base.xcconfig: Only build for i386 and x86_64.
* TestWebKitAPI/Configurations/Base.xcconfig: Ditto.
* WebKitTestRunner/Configurations/Base.xcconfig: Ditto.
2010-12-21 Mihai Parparita <mihaip@chromium.org>
Reviewed by Darin Fisher.
[Chromium] Rename WebThemeEngine/ControlDRT to WebThemeEngine/ControlDRTWin
https://bugs.webkit.org/show_bug.cgi?id=51406
Rename WebThemeControlDRT to WebThemeControlDRTWin and WebThemeEngineDRT
to WebThemeEngineDRTWin and update references everywhere.
* DumpRenderTree/DumpRenderTree.gypi:
* DumpRenderTree/chromium/TestShellWin.cpp:
* DumpRenderTree/chromium/WebThemeControlDRTWin.cpp: Renamed from Tools/DumpRenderTree/chromium/WebThemeControlDRT.cpp.
* DumpRenderTree/chromium/WebThemeControlDRTWin.h: Renamed from Tools/DumpRenderTree/chromium/WebThemeControlDRT.h.
* DumpRenderTree/chromium/WebThemeEngineDRTWin.cpp: Renamed from Tools/DumpRenderTree/chromium/WebThemeEngineDRT.cpp.
* DumpRenderTree/chromium/WebThemeEngineDRTWin.h: Renamed from Tools/DumpRenderTree/chromium/WebThemeEngineDRT.h.
2010-12-21 Anders Carlsson <andersca@apple.com>
Reviewed by John Sullivan.
Clicking missing plug-in text does not show a sheet
https://bugs.webkit.org/show_bug.cgi?id=51403
Update for changes to the WebPageUIClient.
* MiniBrowser/mac/BrowserWindowController.m:
(-[BrowserWindowController awakeFromNib]):
* WebKitTestRunner/TestController.cpp:
(WTR::createOtherPage):
(WTR::TestController::initialize):
2010-12-21 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
commit-queue will report constant failures as flaky if other tests flake
https://bugs.webkit.org/show_bug.cgi?id=51272
This patch just removes functionality and adds testing.
Previously we attempted to report flaky tests when we had
two different tests fail in a row. However, since we stop
running the tests at the first failure, our code was wrong in
trying to determine flakiness from the incomplete runs.
Originally I posted an alternate patch:
https://bug-51272-attachments.webkit.org/attachment.cgi?id=77078
which fixed our flaky logic in this case, however it was decided
that that patch would be too difficult to maintain, so now
I'm just removing the broken logic.
This will dramatically cut-down on our flaky-test false positives
at the (small) cost of the queues being unable to report
any flakiness if the tree is very flaky. (With at least one test
flaking on every run, we'll never report failures anymore.) I think
this is a tradeoff worth making.
* Scripts/webkitpy/tool/bot/commitqueuetask.py:
* Scripts/webkitpy/tool/bot/commitqueuetask_unittest.py:
2010-12-20 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
commit-queue wrongly rejects patches when it can't update itself
https://bugs.webkit.org/show_bug.cgi?id=46636
* Scripts/webkitpy/tool/bot/commitqueuetask.py:
* Scripts/webkitpy/tool/bot/commitqueuetask_unittest.py:
* Scripts/webkitpy/tool/commands/queues_unittest.py:
2010-12-20 Adam Barth <abarth@webkit.org>
Move web sites to Websites directory
https://bugs.webkit.org/show_bug.cgi?id=51323
Update references to BugsSite to point to the new location.
This patch was never officially reviewed (because it was too large to
upload to bugs.webkit.org), but it was discussed on webkit-dev and Mark
Rowe gave me the green light.
* Scripts/old-run-webkit-tests:
* Scripts/webkitpy/common/config/build.py:
* Scripts/webkitpy/common/config/build_unittest.py:
* Scripts/webkitpy/common/prettypatch.py:
* Scripts/webkitpy/layout_tests/port/base.py:
* Scripts/webkitpy/style/main.py:
2010-12-20 Adrienne Walker <enne@google.com>
Unreviewed. Adding myself to the list of committers.
* webkitpy/common/config/committers.py:
2010-12-20 Mark Rowe <mrowe@apple.com>
Reviewed by Dan Bernstein.
<rdar://problem/8757601> Don't install header files in to the Resources directory.
* Scripts/check-for-inappropriate-files-in-framework: Don't allow .h files in the
Resources directory.
2010-12-20 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
commit-queue should include bot id when attaching failure diffs
https://bugs.webkit.org/show_bug.cgi?id=51280
This is a tiny change to include the bot id in the name of the attachment.
Most of this diff is just changing the unit test expectations
now that I hid the comment printing when the comment is None.
* Scripts/webkitpy/tool/bot/flakytestreporter.py:
* Scripts/webkitpy/tool/bot/flakytestreporter_unittest.py:
* Scripts/webkitpy/tool/commands/upload_unittest.py:
* Scripts/webkitpy/tool/mocktool.py:
2010-12-20 Jeff Miller <jeffm@apple.com>
Reviewed by Dan Bernstein.
Tools/vcbin/midl.exe needs to be rebuilt
https://bugs.webkit.org/show_bug.cgi?id=51347
* vcbin/midl.exe: Rebuilt.
2010-12-20 David Levin <levin@chromium.org>
Reviewed by Ariya Hidayat.
.gitignore and webkit-tools-completion.sh have references to WebKitTools that should be Tools.
https://bugs.webkit.org/show_bug.cgi?id=51343
Follow up from the WebKitTools -> Tools rename.
* Scripts/webkit-tools-completion.sh:
2010-12-20 Cosmin Truta <ctruta@chromium.org>
Reviewed by James Robinson.
new-run-webkit-tests ignores trailing EOL differences in text tests
https://bugs.webkit.org/show_bug.cgi?id=36983
Changed the handling of new-line characters within new-run-webkit-tests
to match old-run-webkit-tests. Differences in leading and trailing empty
lines in text expectation files are no longer ignored.
* Scripts/webkitpy/layout_tests/port/base.py:
* Scripts/webkitpy/layout_tests/port/test.py:
Added unit tests. Removed old duplicate unit test entries.
* Scripts/webkitpy/layout_tests/test_types/text_diff.py:
2010-12-20 David Levin <levin@chromium.org>
Reviewed by Shinichiro Hamaji.
check-webkit-style should detect function declarations (and trivial functions).
https://bugs.webkit.org/show_bug.cgi?id=51303
* Scripts/webkitpy/style/checkers/cpp.py:
(_FunctionState.begin): Add is_declaration and changed the line count
start to begin at -1 (which will keep the results consistent, since
the starting line number passed in is one less in this change).
(detect_functions): changed function detection to now catch trivial
functions and declarations.
(check_pass_ptr_usage): Don't check for Pass*Ptr on the first line
of the function as this may look at return values (when processing
a declaration).
* Scripts/webkitpy/style/checkers/cpp_unittest.py:
(FunctionDetectionTest.perform_function_detection): Basic mechanics
of testing the function detection.
(FunctionDetectionTest.test_basic_function_detection): Test a simple
function.
(FunctionDetectionTest.test_function_declaration_detection): Test a
declaration.
(FunctionDetectionTest.test_non_functions): A test case for a case
that caused the code to fail due to the { being in quotes.
(PassPtrTest.test_pass_ref_ptr_return_value): Added some more test
cases to help catch false alarms for return values.
(PassPtrTest.test_pass_ref_ptr_member_variable): Ensure that
we don't get false alarms for member variables either.
2010-12-20 Ryuan Choi <ryuan.choi@samsung.com>
Reviewed by Antonio Gomes.
[CMAKE] Rename WEBKITTOOLS_DIR to TOOLS_DIR
https://bugs.webkit.org/show_bug.cgi?id=51319
* CMakeListsEfl.txt:
2010-12-18 Adam Barth <abarth@webkit.org>
Reviewed by Sam Weinig.
Move WebKitExamplePlugins to Examples
https://bugs.webkit.org/show_bug.cgi?id=51291
* Scripts/webkitpy/common/config/build.py:
2010-12-17 Csaba Osztrogonác <ossy@webkit.org>
Unreviewed Qt buildfix after r74301.
Rename WebKitTools to Tools
https://bugs.webkit.org/show_bug.cgi?id=49861
* MiniBrowser/DerivedSources.pro:
* MiniBrowser/qt/MiniBrowser.pro:
2010-12-17 Dan Bernstein <mitz@apple.com>
Reviewed by Simon Fraser.
Rename WebKitTools to Tools
https://bugs.webkit.org/show_bug.cgi?id=49861
* BuildSlaveSupport/build.webkit.org-config/master.cfg:
* CodeCoverage/README:
* DumpRenderTree/qt/DumpRenderTreeQt.cpp:
(WebCore::DumpRenderTree::initializeFonts):
* EWSTools/start-commit-queue.sh:
* EWSTools/start-queue.sh:
* GNUmakefile.am:
* MIDLWrapper/MIDLWrapper.cpp:
(wmain):
* MiniBrowser/qt/MiniBrowser.pro:
* Scripts/build-api-tests:
* Scripts/build-dumprendertree:
* Scripts/build-webkit:
* Scripts/build-webkittestrunner:
* Scripts/generate-coverage-data:
* Scripts/old-run-webkit-tests:
* Scripts/run-api-tests:
* Scripts/run-iexploder-tests:
* Scripts/run-javascriptcore-tests:
* Scripts/run-mangleme-tests:
* Scripts/run-sunspider:
* Scripts/run-webkit-websocketserver:
* Scripts/sunspider-compare-results:
* Scripts/test-webkitperl:
* Scripts/test-webkitpy:
* Scripts/update-iexploder-cssproperties:
* Scripts/update-webkit:
* Scripts/update-webkit-localizable-strings:
* Scripts/webkitdirs.pm:
* Scripts/webkitpy/common/checkout/scm.py:
* Scripts/webkitpy/common/checkout/scm_unittest.py:
* Scripts/webkitpy/common/config/committervalidator.py:
* Scripts/webkitpy/common/config/committervalidator_unittest.py:
* Scripts/webkitpy/common/config/ports.py:
* Scripts/webkitpy/common/config/ports_unittest.py:
* Scripts/webkitpy/common/system/logutils_unittest.py:
* Scripts/webkitpy/common/system/ospath_unittest.py:
* Scripts/webkitpy/layout_tests/port/chromium.py:
* Scripts/webkitpy/layout_tests/port/config.py:
* Scripts/webkitpy/layout_tests/port/config_unittest.py:
* Scripts/webkitpy/layout_tests/port/http_server.py:
* Scripts/webkitpy/layout_tests/port/test.py:
* Scripts/webkitpy/style/checker.py:
* Scripts/webkitpy/style/checker_unittest.py:
* Scripts/webkitpy/tool/bot/feeders_unittest.py:
* Scripts/webkitpy/tool/commands/queues_unittest.py:
* Scripts/webkitpy/tool/steps/steps_unittest.py:
* WebKitTestRunner/DerivedSources.pro:
* WebKitTestRunner/InjectedBundle/qt/ActivateFontsQt.cpp:
(WTR::activateFonts):
* WebKitTestRunner/InjectedBundle/qt/InjectedBundle.pro:
* wx/build/build_utils.py:
2010-12-16 Antonio Gomes <agomes@rim.com>
Reviewed by Martin Robinson.
Settings::editingBehaviorType() incorrectly returns mac for the gtk build when running tests
https://bugs.webkit.org/show_bug.cgi?id=51163
Since GTK+'s default editing behavior was changed to UNIX in r70975 (see webkitwebsettings.cpp),
GTK+'s DRT should also reset to UNIX after each test execution.
* DumpRenderTree/gtk/DumpRenderTree.cpp:
(resetDefaultsToConsistentValues):
2010-12-17 David Levin <levin@chromium.org>
Reviewed by Eric Seidel.
check-webkit-style should understand WTF #include guards
https://bugs.webkit.org/show_bug.cgi?id=44911
* Scripts/webkitpy/style/checkers/cpp.py:
(get_header_guard_cpp_variable): modify to suggest the WTF style
of header guard when appropriate.
(check_for_header_guard): handle multiple return values from
get_header_guard_cpp_variable
* Scripts/webkitpy/style/checkers/cpp_unittest.py:
(CppStyleTest.test_build_header_guard): Added tests for the WTF
header style.
2010-12-16 David Levin <levin@chromium.org>
Reviewed by Eric Seidel.
test-webkitpy: unittest for the xml.py checker displays a decprecation warning.
https://bugs.webkit.org/show_bug.cgi?id=51210
The error only shows up when using python 2.6 or later.
* Scripts/webkitpy/style/checkers/xml.py: Replace the usage of
the decprecated field error,message with something equivalent.
2010-12-16 Mihai Parparita <mihaip@chromium.org>
Reviewed by Eric Seidel.
Add --exit-after-n-failures/crashes to NRWT
https://bugs.webkit.org/show_bug.cgi?id=51160
Abort test run (in a similar way to how control-C is handled) when
--exit-after-n-failures/crashes-or-timeouts are passed and we've reached
that number of unexpected failures/crashes/timeouts.
* Scripts/webkitpy/layout_tests/port/test.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2010-12-16 David Levin <levin@chromium.org>
Reviewed by Shinichiro Hamaji.
check-webkit-style unit tests has some duplicate boilerplate code.
https://bugs.webkit.org/show_bug.cgi?id=49519
* Scripts/webkitpy/style/checkers/cpp.py:
(update_include_state): Replaced the "io" parameter with the global
configuration _unit_test_config. This allowed not calling into
functions at a low level and also not plumbing through the injection
information through many levels of code.
(check_for_include_what_you_use): Ditto.
(process_file_data): Added the ability to set up the unit test config
to allow for injection.
* Scripts/webkitpy/style/checkers/cpp_unittest.py:
(ErrorCollector.__init__): Added support for having a filter for errors.
(ErrorCollector.__call__): Ditto.
(CppStyleTestBase.process_file_data): Added the ability to set unit_test_config.
(CppStyleTestBase.perform_lint): Consolidated logic for the perform functions.
(CppStyleTestBase.perform_single_line_lint): Replace specific calls to
functions in the cpp.py with generic processing and a filter that
indicates what errors should be kept.
(CppStyleTestBase.perform_multi_line_lint): Ditto.
(CppStyleTestBase.perform_language_rules_check): Ditto.
(CppStyleTestBase.perform_function_lengths_check): Ditto.
(CppStyleTestBase.perform_pass_ptr_check): Ditto.
(CppStyleTestBase.perform_include_what_you_use): Ditto.
(CppStyleTest.test_multi_line_comments): Added another
error message which applies to the test case.
(CppStyleTest.test_spacing_for_binary_ops): Fixed test
to not have config.h, since it is processed as a header file.
(CppStyleTest.test_static_or_global_stlstrings): Fixed variable name
style and indentation in checked code.
(OrderOfIncludesTest.test_check_preprocessor_in_include_section):
Fixed line number.
(NoNonVirtualDestructorsTest.test_multi_line_declaration_with_error):
Ditto.
2010-12-15 Sheriff Bot <webkit.review.bot@gmail.com>
Unreviewed, rolling out r74136.
http://trac.webkit.org/changeset/74136
https://bugs.webkit.org/show_bug.cgi?id=51135
r74136 breaks chromium canary bots because some tests are not
rebaselined correctly to resolve EOL differences (Requested by
jianli on #webkit).
* Scripts/webkitpy/layout_tests/port/base.py:
* Scripts/webkitpy/layout_tests/port/test.py:
* Scripts/webkitpy/layout_tests/test_types/text_diff.py:
2010-12-15 Adam Roben <aroben@apple.com>
Teach check-webkit-style to check .vcproj and .vsprops files for XML
syntax errors
Fixes <http://webkit.org/b/51103> check-webkit-style should check for
XML syntax errors in .vcproj/.vsprops files
Reviewed by Dave Levin.
* Scripts/webkitpy/style/checker.py: Added lists of file extensions
that should be treated as XML and that should be allowed to contain
carriage returns. (These lists happen to be identical currently.)
(FileType): Added a new XML type.
(CheckerDispatcher.should_check_and_strip_carriage_returns): Added.
Just does a simple file extension check.
(CheckerDispatcher._file_type): Added a case for XML files.
(CheckerDispatcher._create_checker): Ditto. We use XMLChecker for XML
files (surprise!).
(StyleProcessor.process): Ask the dispatcher whether we should pass the
lines through the carriage checker.
* Scripts/webkitpy/style/checker_unittest.py:
(CheckerDispatcherCarriageReturnTest.test_should_check_and_strip_carriage_returns):
Added. Checks a few file names to see if carriage returns are allowed
or not.
(CheckerDispatcherDispatchTest.assert_checker_xml): Added. Similar to
other assert_checker_* functions.
(CheckerDispatcherDispatchTest.test_xml_paths): Added. Similar to other
test_*_paths functions.
(CheckerDispatcherDispatchTest.test_xml_paths): Added. Similar to other
test_*_paths functions.
(CheckerDispatcherDispatchTest.test_none_paths): Removed the vcproj
file from this test case, as vcproj files now have a type.
(StyleProcessor_CodeCoverageTest.MockDispatcher.should_check_and_strip_carriage_returns):
Added. Similar to the other should_* functions.
(StyleProcessor_CodeCoverageTest.test_process__carriage_returns_not_stripped):
Added. Checks that carriage returns aren't checked for or stripped for
allowed files.
* Scripts/webkitpy/style/checkers/xml.py: Added.
(XMLChecker.__init__): Simple init method.
(XMLChecker.check): Pass each line through the expat parser, and record
a style error for any errors thrown by the parser.
* Scripts/webkitpy/style/checkers/xml_unittest.py: Added.
(XMLCheckerTest.assert_no_error): Checks that the given XML does not
produce a style error.
(XMLCheckerTest.assert_error): Checks that the given XML produces an
error of the given category on the given line.
(XMLCheckerTest.mock_handle_style_error): Does nothing. Used for
checking that the XMLChecker constructor works properly.
(XMLCheckerTest.test_conflict_marker): Tests that conflict markers
cause a style error (see, e.g., r73887).
(XMLCheckerTest.test_extra_closing_tag): Tests that extra closing tags
cause a style error (see, e.g., r73773).
(XMLCheckerTest.test_init): Tests that the XMLChecker constructor works
properly.
(XMLCheckerTest.test_missing_closing_tag): Tests that missing closing
tags cause a style error (see, e.g., r72795).
(XMLCheckerTest.test_no_error): Tests that valid XML does not cause a
style error.
2010-12-15 Lucas Forschler <lforschler@apple.com>
Reviewed by Stephanie Lewis.
https://bugs.webkit.org/show_bug.cgi?id=51117
Add a new leopard test bot
really really commit to trunk this time.
* BuildSlaveSupport/build.webkit.org-config/config.json:
2010-12-15 Sam Weinig <sam@webkit.org>
Reviewed by Anders Carlsson.
WebKit2: Can't add files to an <input type=file>
https://bugs.webkit.org/show_bug.cgi?id=51087
* MiniBrowser/mac/BrowserWindowController.m:
(runOpenPanel):
(-[BrowserWindowController awakeFromNib]):
Add simple implementation of runOpenPanel callback.
* MiniBrowser/win/BrowserView.cpp:
(BrowserView::create):
* WebKitTestRunner/TestController.cpp:
(WTR::createOtherPage):
(WTR::TestController::initialize):
Stub out runOpenPanel callback.
2010-12-14 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
commit-queue should upload failure diffs when tests flake
https://bugs.webkit.org/show_bug.cgi?id=51051
To make this testable I needed to pipe FileSystem down onto tool.
We've wanted it there for a long time anyway.
This patch is kinda a big hack. But we don't have a nice
way to read results.html files. I think this will need further
revision before this code actually feels clean.
As part of testing this change, I had to make MockBugzilla.create_bug
actually return an id (like it should) which required updating
a few other unit test results (for the better).
The results_matching_keys change in layouttestresults/rebasline
was an alternate path which I decided not to use in the end, but
I left the change as it seemed an improvement.
* Scripts/webkitpy/common/net/bugzilla/bugzilla.py:
* Scripts/webkitpy/common/net/layouttestresults.py:
* Scripts/webkitpy/tool/bot/flakytestreporter.py:
* Scripts/webkitpy/tool/bot/flakytestreporter_unittest.py:
* Scripts/webkitpy/tool/commands/queues.py:
* Scripts/webkitpy/tool/commands/rebaseline.py:
* Scripts/webkitpy/tool/main.py:
* Scripts/webkitpy/tool/mocktool.py:
2010-12-15 Cosmin Truta <ctruta@chromium.org>
Reviewed by James Robinson.
new-run-webkit-tests ignores trailing EOL differences in text tests
https://bugs.webkit.org/show_bug.cgi?id=36983
Changed the handling of new-line characters within new-run-webkit-tests
to match old-run-webkit-tests. Differences in leading and trailing empty
lines in text expectation files are no longer ignored.
* Scripts/webkitpy/layout_tests/port/base.py:
* Scripts/webkitpy/layout_tests/port/test.py:
Added unit tests. Removed old duplicate unit test entries.
* Scripts/webkitpy/layout_tests/test_types/text_diff.py:
2010-12-15 Ojan Vafai <ojan@chromium.org>
Reviewed by Adam Barth.
make status-bubble white-space:nowrap so we can measure it's width without wrapping
https://bugs.webkit.org/show_bug.cgi?id=51149
* QueueStatusServer/templates/statusbubble.html:
2010-12-15 Ojan Vafai <ojan@chromium.org>
Reviewed by Adam Barth.
have the statusbubble postMessage it's metrics so that embedders can properly size the iframe
https://bugs.webkit.org/show_bug.cgi?id=51125
* QueueStatusServer/templates/statusbubble.html:
2010-12-15 Darin Adler <darin@apple.com>
* Scripts/webkitpy/common/net/buildbot: Added property svn:ignore.
2010-12-15 Sheriff Bot <webkit.review.bot@gmail.com>
Unreviewed, rolling out r74117.
http://trac.webkit.org/changeset/74117
https://bugs.webkit.org/show_bug.cgi?id=51113
This broke the GTK1 build. (Requested by mrobinson on
#webkit).
* Scripts/webkitdirs.pm:
2010-12-15 Amruth Raj <amruthraj@motorola.com> and Ravi Kasibhatla <ravi.kasibhatla@motorola.com>
Reviewed by Martin Robinson.
Change generate-forwarding-headers.pl for GTK port usage
(https://bugs.webkit.org/show_bug.cgi?id=37369)
* Scripts/webkitdirs.pm: Added changes to build webkit2 for GTK port using build-webkit script.
2010-12-14 Kent Tamura <tkent@chromium.org>
Reviewed by Jian Li.
[DRT/Chromium] Remove another unnecessary error message
https://bugs.webkit.org/show_bug.cgi?id=51083
* DumpRenderTree/chromium/WebViewHost.cpp:
(WebViewHost::didInvalidateRect):
2010-12-14 Lucas Forschler <lforschler@apple.com>
Reviewed by Stephanie Lewis.
Add a new Leopard Debug Test WK2 Bot
* BuildSlaveSupport/build.webkit.org-config/config.json:
2010-12-14 Ojan Vafai <ojan@chromium.org>
Fix python unittests after http://trac.webkit.org/changeset/74070.
* Scripts/webkitpy/style/checkers/test_expectations_unittest.py:
2010-12-14 Kent Tamura <tkent@chromium.org>
Reviewed by Jian Li.
[DRT/Chromium] Remove a unnecessary error message
https://bugs.webkit.org/show_bug.cgi?id=51069
* DumpRenderTree/chromium/WebViewHost.cpp:
(WebViewHost::paintInvalidatedRegion):
2010-12-14 Benjamin Kalman <kalman@chromium.org>
Reviewed by Ojan Vafai.
Using BUG/BUGWK in test_expectations is error prone, should use BUGCR/BUGWK
https://bugs.webkit.org/show_bug.cgi?id=48926
* Scripts/webkitpy/layout_tests/layout_package/test_expectations.py:
Add presubmit check that BUG isn't used, either BUGCR/BUGWK/BUGV8_.
2010-12-14 Laszlo Gombos <laszlo.1.gombos@nokia.com>
Reviewed by Eric Seidel.
[Qt] [Symbian] Do not use pkg-config on Symbian as it is not supported
https://bugs.webkit.org/show_bug.cgi?id=50231
Guard CONFIG+=link_pkgconfig with !symbian.
* DumpRenderTree/qt/DumpRenderTree.pro:
* WebKitTestRunner/InjectedBundle/qt/InjectedBundle.pro:
* WebKitTestRunner/qt/WebKitTestRunner.pro:
2010-12-14 Philippe Normand <pnormand@igalia.com>
Reviewed by Ojan Vafai.
[new-run-webkit-tests] expectations parsing is slow
https://bugs.webkit.org/show_bug.cgi?id=50635
Avoid expensive iteration of all the tests when checking if a test
file is to be skipped or not.
* Scripts/webkitpy/layout_tests/layout_package/test_expectations.py:
* Scripts/webkitpy/layout_tests/layout_package/test_expectations_unittest.py:
* Scripts/webkitpy/layout_tests/port/test.py:
2010-12-14 Mario Sanchez Prada <msanchez@igalia.com>
Reviewed by Xan Lopez.
[Gtk] Implement STATE_FOCUSED, STATE_FOCUSABLE, and corresponding events for text objects
https://bugs.webkit.org/show_bug.cgi?id=27048
Add support in DRT for checking whether an accessibility UI
element is focusable and/or focused. Implemented for GTK.
* DumpRenderTree/AccessibilityUIElement.cpp:
(getIsFocusedCallback): New.
(getIsFocusableCallback): New.
(AccessibilityUIElement::getJSClass): Add the new available
callbacks for isFocused and isFocusable.
* DumpRenderTree/AccessibilityUIElement.h:
* DumpRenderTree/gtk/AccessibilityUIElementGtk.cpp:
(AccessibilityUIElement::isFocused): New, implemented by checking
whether the related AtkState value is in the object's state set.
(AccessibilityUIElement::isFocusable): Ditto.
* DumpRenderTree/mac/AccessibilityUIElementMac.mm:
(AccessibilityUIElement::isFocused): New, dummy implementation.
(AccessibilityUIElement::isFocusable): Ditto.
* DumpRenderTree/win/AccessibilityUIElementWin.cpp:
(AccessibilityUIElement::isFocused): Ditto.
(AccessibilityUIElement::isFocusable): Ditto.
2010-12-14 Eric Seidel <eric@webkit.org>
Reviewed by Ojan Vafai.
webkit-patch should warn users when they're using a 32-bit git on a 64-bit system
https://bugs.webkit.org/show_bug.cgi?id=50715
This patch makes webkit-patch print the following:
Warning: This machine is 64-bit, but the git binary (/usr/local/git/bin/git) does not support 64-bit.
Install a 64-bit git for better performance, see:
https://lists.webkit.org/pipermail/webkit-dev/2010-December/015249.html
I wrote this mostly because I have approximately 8 machines that I use
and making sure each one is using a good Git install seemed folly.
webkit-patch makes a lot of git calls, so using a fast git can shave
several seconds in every invocation. See the webkit-dev thread for more info.
This message will print twice during 'webkit-patch upload',
once from webkit-patch and once from check-webkit-style.
Unfortunately there is no good way to test this due to how machine-dependent
the code is. I considered writing a test for the log message, but it seemed not worth it.
* Scripts/webkitpy/common/checkout/scm.py:
2010-12-13 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
FlakyTestReporter doesn't understand bots running from multiple email addresses
https://bugs.webkit.org/show_bug.cgi?id=50960
This explains at least one of the dupes of:
https://bugs.webkit.org/show_bug.cgi?id=50863
that we saw filed by the commit-queue this morning.
I think the other one was explained by my previous fix to result counting code.
Since this is really hard to test with a unit test, instead I
create a new (possibly useful in the future) command
which given a layout test path will return you
the one bug which our tools would assume it the flaky test bug.
If some other script wants to use bug-for-test we'll
need to extend it with some options like --create-if-missing or similar.
* Scripts/webkitpy/common/net/bugzilla/bug.py:
* Scripts/webkitpy/common/net/bugzilla/bugzilla.py:
* Scripts/webkitpy/tool/bot/flakytestreporter.py:
* Scripts/webkitpy/tool/commands/__init__.py:
* Scripts/webkitpy/tool/commands/bugfortest.py: Added.
2010-12-13 Mihai Parparita <mihaip@chromium.org>
Reviewed by James Robinson.
Add option to build-webkit to not build WebKit2
https://bugs.webkit.org/show_bug.cgi?id=50988
Add support for a --no-webkit2 option to build-webkit. It seems to save
~1 minute from clean builds and ~10 seconds from no-op builds.
* Scripts/build-webkit:
2010-12-13 Lucas Forschler <lforschler@apple.com>
Reviewed by Dave Kilzer.
Update iexploder from 1.3.2 to 1.7.2
Move 1.3.2 into its own subdir
Add 1.7.2 into its own subdir
update scripts to point to new 1.3.2 location (avoid breaking existing scripts)
* Scripts/run-iexploder-tests:
* Scripts/update-iexploder-cssproperties:
* iExploder/CHANGELOG.txt: Removed.
* iExploder/LICENSE.txt: Removed.
* iExploder/README.txt: Removed.
* iExploder/htdocs: Removed.
* iExploder/iexploder-1.3.2: Added.
* iExploder/iexploder-1.3.2/CHANGELOG.txt: Copied from iExploder/CHANGELOG.txt.
* iExploder/iexploder-1.3.2/LICENSE.txt: Copied from iExploder/LICENSE.txt.
* iExploder/iexploder-1.3.2/README.txt: Copied from iExploder/README.txt.
* iExploder/iexploder-1.3.2/htdocs: Added.
* iExploder/iexploder-1.3.2/htdocs/config.rb: Copied from iExploder/htdocs/config.rb.
* iExploder/iexploder-1.3.2/htdocs/cssproperties.in: Copied from iExploder/htdocs/cssproperties.in.
* iExploder/iexploder-1.3.2/htdocs/cssvalues.in: Copied from iExploder/htdocs/cssvalues.in.
* iExploder/iexploder-1.3.2/htdocs/htmlattrs.in: Copied from iExploder/htdocs/htmlattrs.in.
* iExploder/iexploder-1.3.2/htdocs/htmltags.in: Copied from iExploder/htdocs/htmltags.in.
* iExploder/iexploder-1.3.2/htdocs/htmlvalues.in: Copied from iExploder/htdocs/htmlvalues.in.
* iExploder/iexploder-1.3.2/htdocs/iexploder.cgi: Copied from iExploder/htdocs/iexploder.cgi.
* iExploder/iexploder-1.3.2/htdocs/iexploder.rb: Copied from iExploder/htdocs/iexploder.rb.
* iExploder/iexploder-1.3.2/htdocs/index.html: Copied from iExploder/htdocs/index.html.
* iExploder/iexploder-1.3.2/htdocs/webserver.rb: Copied from iExploder/htdocs/webserver.rb.
* iExploder/iexploder-1.3.2/tools: Added.
* iExploder/iexploder-1.3.2/tools/lasthit.rb: Copied from iExploder/tools/lasthit.rb.
* iExploder/iexploder-1.3.2/tools/osx_last_crash.rb: Copied from iExploder/tools/osx_last_crash.rb.
* iExploder/iexploder-1.3.2/tools/showtest.rb: Copied from iExploder/tools/showtest.rb.
* iExploder/iexploder-1.7.2: Added.
* iExploder/iexploder-1.7.2/ChangeLog.txt: Added.
* iExploder/iexploder-1.7.2/LICENSE.txt: Added.
* iExploder/iexploder-1.7.2/README.txt: Added.
* iExploder/iexploder-1.7.2/output: Added.
* iExploder/iexploder-1.7.2/src: Added.
* iExploder/iexploder-1.7.2/src/browser_harness.rb: Added.
* iExploder/iexploder-1.7.2/src/config.yaml: Added.
* iExploder/iexploder-1.7.2/src/css-atrules: Added.
* iExploder/iexploder-1.7.2/src/css-atrules/mozilla: Added.
* iExploder/iexploder-1.7.2/src/css-atrules/webkit: Added.
* iExploder/iexploder-1.7.2/src/css-properties: Added.
* iExploder/iexploder-1.7.2/src/css-properties/dillo: Added.
* iExploder/iexploder-1.7.2/src/css-properties/gtkhtml: Added.
* iExploder/iexploder-1.7.2/src/css-properties/internet_explorer6: Added.
* iExploder/iexploder-1.7.2/src/css-properties/mozilla: Added.
* iExploder/iexploder-1.7.2/src/css-properties/webkit: Added.
* iExploder/iexploder-1.7.2/src/css-pseudo: Added.
* iExploder/iexploder-1.7.2/src/css-pseudo/mozilla: Added.
* iExploder/iexploder-1.7.2/src/css-pseudo/webkit: Added.
* iExploder/iexploder-1.7.2/src/css-values: Added.
* iExploder/iexploder-1.7.2/src/css-values/dillo: Added.
* iExploder/iexploder-1.7.2/src/css-values/gtkhtml: Added.
* iExploder/iexploder-1.7.2/src/css-values/mozilla: Added.
* iExploder/iexploder-1.7.2/src/css-values/other: Added.
* iExploder/iexploder-1.7.2/src/css-values/webkit: Added.
* iExploder/iexploder-1.7.2/src/headers: Added.
* iExploder/iexploder-1.7.2/src/headers/dillo: Added.
* iExploder/iexploder-1.7.2/src/headers/gtkhtml: Added.
* iExploder/iexploder-1.7.2/src/headers/mozilla: Added.
* iExploder/iexploder-1.7.2/src/headers/webkit: Added.
* iExploder/iexploder-1.7.2/src/html-attrs: Added.
* iExploder/iexploder-1.7.2/src/html-attrs/dillo: Added.
* iExploder/iexploder-1.7.2/src/html-attrs/gtkhtml: Added.
* iExploder/iexploder-1.7.2/src/html-attrs/internet_explorer6: Added.
* iExploder/iexploder-1.7.2/src/html-attrs/mozilla: Added.
* iExploder/iexploder-1.7.2/src/html-attrs/other: Added.
* iExploder/iexploder-1.7.2/src/html-attrs/webkit: Added.
* iExploder/iexploder-1.7.2/src/html-tags: Added.
* iExploder/iexploder-1.7.2/src/html-tags/dillo: Added.
* iExploder/iexploder-1.7.2/src/html-tags/gtkhtml: Added.
* iExploder/iexploder-1.7.2/src/html-tags/mozilla: Added.
* iExploder/iexploder-1.7.2/src/html-tags/other: Added.
* iExploder/iexploder-1.7.2/src/html-tags/webkit: Added.
* iExploder/iexploder-1.7.2/src/html-values: Added.
* iExploder/iexploder-1.7.2/src/html-values/dillo: Added.
* iExploder/iexploder-1.7.2/src/html-values/gtkhtml: Added.
* iExploder/iexploder-1.7.2/src/html-values/mozilla: Added.
* iExploder/iexploder-1.7.2/src/html-values/other: Added.
* iExploder/iexploder-1.7.2/src/html-values/webkit: Added.
* iExploder/iexploder-1.7.2/src/iexploder.cgi: Added.
* iExploder/iexploder-1.7.2/src/iexploder.rb: Added.
* iExploder/iexploder-1.7.2/src/index.html: Added.
* iExploder/iexploder-1.7.2/src/media: Added.
* iExploder/iexploder-1.7.2/src/media/blank.ogg: Added.
* iExploder/iexploder-1.7.2/src/media/blank.snd: Added.
* iExploder/iexploder-1.7.2/src/media/blank.wav: Added.
* iExploder/iexploder-1.7.2/src/media/bug.bmp: Added.
* iExploder/iexploder-1.7.2/src/media/bug.gif: Added.
* iExploder/iexploder-1.7.2/src/media/bug.ico: Added.
* iExploder/iexploder-1.7.2/src/media/bug.jng: Added.
* iExploder/iexploder-1.7.2/src/media/bug.jpg: Added.
* iExploder/iexploder-1.7.2/src/media/bug.png: Added.
* iExploder/iexploder-1.7.2/src/media/bug.svg: Added.
* iExploder/iexploder-1.7.2/src/media/bug.tiff: Added.
* iExploder/iexploder-1.7.2/src/media/bug.xbm: Added.
* iExploder/iexploder-1.7.2/src/media/bug.xpm: Added.
* iExploder/iexploder-1.7.2/src/mime-types: Added.
* iExploder/iexploder-1.7.2/src/mime-types/dillo: Added.
* iExploder/iexploder-1.7.2/src/mime-types/mozilla: Added.
* iExploder/iexploder-1.7.2/src/mime-types/webkit: Added.
* iExploder/iexploder-1.7.2/src/protocols: Added.
* iExploder/iexploder-1.7.2/src/protocols/dillo: Added.
* iExploder/iexploder-1.7.2/src/protocols/gtkhtml: Added.
* iExploder/iexploder-1.7.2/src/protocols/mozilla: Added.
* iExploder/iexploder-1.7.2/src/protocols/webkit: Added.
* iExploder/iexploder-1.7.2/src/scanner.rb: Added.
* iExploder/iexploder-1.7.2/src/version.rb: Added.
* iExploder/iexploder-1.7.2/src/webserver.rb: Added.
* iExploder/iexploder-1.7.2/testcases: Added.
* iExploder/iexploder-1.7.2/testcases/testcase-Linux_x86_64_rv2.0b6pre_Gecko-20100904_Firefox-4.0b6pre-TEST-8375-1_59.html: Added.
* iExploder/iexploder-1.7.2/testcases/testcase-Opera-9.80_Linux_x86_64_en_Presto-2.6.30_Version-10.61-16704-3_108,3.html: Added.
* iExploder/iexploder-1.7.2/testcases/testcase-U_Intel_Mac_OS_X_10_6_4_en-US_AppleWebKit-534.8_Chrome-7.0.529.0-TEST-611-3_36,9.html: Added.
* iExploder/iexploder-1.7.2/testcases/testcase-U_Intel_Mac_OS_X_10_6_4_en-US_AppleWebKit-534.8_hrome-7.0.529.0-TEST-55313622206-3_6,0.html: Added.
* iExploder/iexploder-1.7.2/testcases/testcase-U_Linux_x86_64_en-US_AppleWebKit-534.6_Chrome-7.0.503.1-TEST-120813-8_72,56,24,8,0.html: Added.
* iExploder/iexploder-1.7.2/testcases/testcase-U_Linux_x86_64_en-US_AppleWebKit-534.6_Chrome-7.0.503.1-TEST-121240-3_81,3.html: Added.
* iExploder/iexploder-1.7.2/testcases/testcase-U_Linux_x86_64_en-US_AppleWebKit-534.7_Chrome-7.0.513.0-4800-5_80,65,15.html: Added.
* iExploder/iexploder-1.7.2/testcases/testcase-U_Linux_x86_64_en-US_AppleWebKit-534.7_Chrome-7.0.513.0-TEST-23583190347-3_15,12.html: Added.
* iExploder/iexploder-1.7.2/tools: Added.
* iExploder/iexploder-1.7.2/tools/lasthit.rb: Added.
* iExploder/iexploder-1.7.2/tools/osx_last_crash.rb: Added.
* iExploder/iexploder-1.7.2/tools/release_src.sh: Added.
* iExploder/iexploder-1.7.2/tools/update_html_tags_from_sources.sh: Added.
* iExploder/tools: Removed.
2010-12-13 Sam Weinig <sam@webkit.org>
Reviewed by Anders Carlsson.
Remove associated page concept from WKView constructor
https://bugs.webkit.org/show_bug.cgi?id=50983
Step 1: Remove associate page constructors from WKView and make
all views use the shared namespace for the context.
* TestWebKitAPI/PlatformWebView.h:
* TestWebKitAPI/mac/PlatformWebViewMac.mm:
* TestWebKitAPI/win/PlatformWebViewWin.cpp:
* WebKitTestRunner/PlatformWebView.h:
* WebKitTestRunner/TestController.cpp:
(WTR::createOtherPage):
* WebKitTestRunner/mac/PlatformWebViewMac.mm:
* WebKitTestRunner/qt/PlatformWebViewQt.cpp:
* WebKitTestRunner/win/PlatformWebViewWin.cpp:
2010-12-13 Sam Weinig <sam@webkit.org>
Reviewed by Anders Carlsson.
Change the WebKit2 public API so there is no explicit WKPageNamespace object
https://bugs.webkit.org/show_bug.cgi?id=50898
* MiniBrowser/mac/AppDelegate.h:
* MiniBrowser/mac/AppDelegate.m:
* MiniBrowser/mac/BrowserWindowController.h:
* MiniBrowser/mac/BrowserWindowController.m:
* MiniBrowser/win/BrowserView.cpp:
Convert MiniBrowser to not use PageNamespaces.
* TestWebKitAPI/PlatformWebView.h:
* TestWebKitAPI/Tests/WebKit2/DocumentStartUserScriptAlertCrash.cpp:
* TestWebKitAPI/Tests/WebKit2/EvaluateJavaScript.cpp:
* TestWebKitAPI/Tests/WebKit2/FailedLoad.cpp:
* TestWebKitAPI/Tests/WebKit2/Find.cpp:
* TestWebKitAPI/Tests/WebKit2/FrameMIMETypeHTML.cpp:
* TestWebKitAPI/Tests/WebKit2/FrameMIMETypePNG.cpp:
* TestWebKitAPI/Tests/WebKit2/InjectedBundleBasic.cpp:
* TestWebKitAPI/Tests/WebKit2/PageLoadBasic.cpp:
* TestWebKitAPI/Tests/WebKit2/PageLoadDidChangeLocationWithinPageForFrame.cpp:
* TestWebKitAPI/Tests/WebKit2/PreventEmptyUserAgent.cpp:
* TestWebKitAPI/Tests/WebKit2/SpacebarScrolling.cpp:
* TestWebKitAPI/Tests/WebKit2/win/AltKeyGeneratesWMSysCommand.cpp:
* TestWebKitAPI/Tests/WebKit2/win/WMCloseCallsUIClientClose.cpp:
* TestWebKitAPI/mac/PlatformWebViewMac.mm:
* TestWebKitAPI/win/PlatformWebViewWin.cpp:
Convert TestWebKitAPI to not use PageNamespaces.
* WebKitTestRunner/PlatformWebView.h:
* WebKitTestRunner/TestController.cpp:
* WebKitTestRunner/TestController.h:
* WebKitTestRunner/mac/PlatformWebViewMac.mm:
* WebKitTestRunner/qt/PlatformWebViewQt.cpp:
* WebKitTestRunner/win/PlatformWebViewWin.cpp:
Convert WebKitTestRunner to not use PageNamespaces.
2010-12-13 Gabor Rapcsanyi <rgabor@inf.u-szeged.hu>
Reviewed by Eric Seidel.
webkit-patch rollout should be able to do multi-revision rollouts
https://bugs.webkit.org/show_bug.cgi?id=33336
Make it possible to pass more than one revision to webkit-patch.
* Scripts/webkitpy/common/checkout/api.py:
* Scripts/webkitpy/common/checkout/changelog.py:
* Scripts/webkitpy/common/checkout/changelog_unittest.py:
* Scripts/webkitpy/tool/commands/abstractsequencedcommand.py:
* Scripts/webkitpy/tool/commands/download.py:
* Scripts/webkitpy/tool/commands/download_unittest.py:
* Scripts/webkitpy/tool/mocktool.py:
* Scripts/webkitpy/tool/steps/preparechangelogforrevert.py:
* Scripts/webkitpy/tool/steps/revertrevision.py:
2010-12-13 Eric Seidel <eric@webkit.org>
Unreviewed.
Teach webkit-patch how to search bugzilla
https://bugs.webkit.org/show_bug.cgi?id=50500
webkit-patch bug-search fails when there is only one result.
Turns out we needed a bit more logic in our result count parsing code.
* Scripts/webkitpy/common/net/bugzilla/bugzilla.py:
* Scripts/webkitpy/common/net/bugzilla/bugzilla_unittest.py:
2010-12-13 Antti Koivisto <antti@apple.com>
Reviewed by Alexey Proskuryakov.
Add setSerializeHTTPLoads function to allow testing resource load order on OS X.
https://bugs.webkit.org/show_bug.cgi?id=50758
* DumpRenderTree/LayoutTestController.cpp:
(setSerializeHTTPLoadsCallback):
(LayoutTestController::staticFunctions):
* DumpRenderTree/LayoutTestController.h:
* DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
(LayoutTestController::setSerializeHTTPLoads):
* DumpRenderTree/mac/DumpRenderTree.mm:
(resetDefaultsToConsistentValues):
* DumpRenderTree/mac/LayoutTestControllerMac.mm:
(LayoutTestController::setSerializeHTTPLoads):
* DumpRenderTree/win/LayoutTestControllerWin.cpp:
(LayoutTestController::setSerializeHTTPLoads):
2010-12-13 Dimitri Glazkov <dglazkov@chromium.org>
Reviewed by Ojan Vafai.
Remove "use-drt" option from test step command line, since the option was removed in r73748.
https://bugs.webkit.org/show_bug.cgi?id=50936
* BuildSlaveSupport/build.webkit.org-config/master.cfg: Removed the option.
2010-12-13 Patrick Gansterer <paroga@webkit.org>
Reviewed by Csaba Osztrogonác.
Add revlink to BuildBot status
https://bugs.webkit.org/show_bug.cgi?id=50914
This adds a hyperlink to trac changeset on the revision numbers.
* BuildSlaveSupport/build.webkit.org-config/master.cfg:
2010-12-13 Andras Becsi <abecsi@webkit.org>
Reviewed by Csaba Osztrogonác.
[Qt][WK2] Fix build if WebKitTools are not available
https://bugs.webkit.org/show_bug.cgi?id=50242
* MiniBrowser/qt/MiniBrowser.pro: Correct the comment.
* Scripts/webkitdirs.pm: Added WebKitTools/MiniBrowser to $subdirs.
* MiniBrowser/DerivedSources.pro: Added.
2010-12-13 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
EWS Queues should remove orpahned pyc files before starting
https://bugs.webkit.org/show_bug.cgi?id=50904
A previous check added a platform.py which caused
EWS bots on some platforms to hit an import exception.
The platform.py file was removed in a subsequent commit, but
the EWS bots in question stayed stuck. This change will
fix those bots once restarted (and prevent this in the future).
I also went ahead and unified start-queue and start-commit-queue
since they were nearly identical. I also added bot_id support
to both force all EWSes to have bot ids as well as get rid of the one
remaining reason why I had my own copies of these scripts.
Hopefully this will be enough to get all the other bot admins
(ahem, Adam, ahem) to move to using these checked in copies as well.
* EWSTools/start-commit-queue.sh: Removed.
* EWSTools/start-queue.sh:
2010-10-28 MORITA Hajime <morrita@google.com>
Reviewed by Ojan Vafai.
spellcheck does not check pasted text
https://bugs.webkit.org/show_bug.cgi?id=40092
Added LayoutTestController::setAsynchronousSpellCheckingEnabled()
to control the setting.
* DumpRenderTree/LayoutTestController.cpp:
(setAsynchronousSpellCheckingEnabledCallback):
(LayoutTestController::staticFunctions):
* DumpRenderTree/LayoutTestController.h:
* DumpRenderTree/chromium/LayoutTestController.h:
* DumpRenderTree/chromium/LayoutTestController.cpp:
(LayoutTestController::LayoutTestController):
(LayoutTestController::setAsynchronousSpellCheckingEnabled):
* DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
(LayoutTestController::setAsynchronousSpellCheckingEnabled):
* DumpRenderTree/mac/DumpRenderTree.mm:
(resetDefaultsToConsistentValues):
* DumpRenderTree/mac/LayoutTestControllerMac.mm:
(LayoutTestController::setAsynchronousSpellCheckingEnabled):
* DumpRenderTree/qt/LayoutTestControllerQt.h:
* DumpRenderTree/qt/LayoutTestControllerQt.cpp:
(LayoutTestController::setAsynchronousSpellCheckingEnabled):
* DumpRenderTree/win/LayoutTestControllerWin.cpp:
(LayoutTestController::setAsynchronousSpellCheckingEnabled):
* DumpRenderTree/wx/LayoutTestControllerWx.cpp:
(LayoutTestController::setAsynchronousSpellCheckingEnabled):
2010-12-12 Sam Weinig <sam@webkit.org>
Reviewed by Anders Carlsson.
Fix failing API test. It turns out that a pop-state event
is sent before every fragment navigation, so we have to test
for it in addition in PageLoadDidChangeLocationWithinPageForFrame.
* TestWebKitAPI/Tests/WebKit2/PageLoadDidChangeLocationWithinPageForFrame.cpp:
(TestWebKitAPI::didSameDocumentNavigationForFrame):
2010-12-12 Alejandro G. Castro <alex@igalia.com>
Reviewed by Eric Seidel.
[GTK] Add new-run-webkit-tests support to gtk
https://bugs.webkit.org/show_bug.cgi?id=50681
Adding the basic support to run the new-run-webkit-tests.
* Scripts/webkitpy/layout_tests/port/gtk.py:
2010-10-11 Diego Gonzalez <diegohcg@webkit.org>
Reviewed by Kenneth Rohde Christiansen.
[Qt] Mock DeviceOrientation client for DRT
https://bugs.webkit.org/show_bug.cgi?id=47490
* DumpRenderTree/qt/DumpRenderTreeQt.cpp:
(WebCore::DumpRenderTree::DumpRenderTree):
(WebCore::DumpRenderTree::~DumpRenderTree):
* DumpRenderTree/qt/LayoutTestControllerQt.cpp:
(LayoutTestController::setMockDeviceOrientation):
2010-12-11 Philippe Normand <pnormand@igalia.com>
Reviewed by Gustavo Noronha Silva.
[GTK] launcher: disable the Mozilla-style fullscreen API
https://bugs.webkit.org/show_bug.cgi?id=50874
Disabling the fullscreen API until its implementation for GTK is
mature enough. The webview setting for it is already FALSE by
default.
* GtkLauncher/main.c:
(createWindow):
2010-12-10 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
commit-queue flaky test messages show cryptic version information for mac os x
https://bugs.webkit.org/show_bug.cgi?id=50864
Turns out platform.platform() returns kernel version information
which isn't helpful, and just plain confusing on Mac
(OS X 10.6.5 uses Darwin Kernel 10.5.0).
So I've updated PlatformInfo.display_name() to special case
mac. I also found a bad use of sys.platform in the process
and fixed that. (sys.platform always returns 'darwin' on mac).
* Scripts/webkitpy/common/system/platforminfo.py:
* Scripts/webkitpy/common/system/user.py:
* Scripts/webkitpy/tool/bot/flakytestreporter.py:
* Scripts/webkitpy/tool/bot/flakytestreporter_unittest.py:
* Scripts/webkitpy/tool/commands/queues_unittest.py:
2010-12-10 Eric Seidel <eric@webkit.org>
Unreviewed.
Exception seen while reporting flaky test with commit-queue.
Just a missing include.
* Scripts/webkitpy/tool/bot/flakytestreporter.py:
* Scripts/webkitpy/tool/bot/flakytestreporter_unittest.py:
2010-12-10 Kinuko Yasuda <kinuko@chromium.org>
Reviewed by Eric Seidel.
[Chromium] Remove old JSONResultsGenerator script that existed for backward-compatibility
https://bugs.webkit.org/show_bug.cgi?id=50796
Also updating the test code to use JSONResultsGeneratorBase and
to improve test coverage for incremental cases.
* Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
* Scripts/webkitpy/layout_tests/layout_package/json_results_generator_unittest.py:
2010-12-10 Joseph Pecoraro <joepeck@webkit.org>
Reviewed by Darin Adler.
prepare-ChangeLog --help doesn't mention --bug shorthand -b
https://bugs.webkit.org/show_bug.cgi?id=50835
* Scripts/prepare-ChangeLog: be more explicit about -b and match file style.
2010-12-10 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
Bugs created by the commit-queue should all block on a master bug
https://bugs.webkit.org/show_bug.cgi?id=50857
This makes all bugs created by the commit-queue block on:
https://bugs.webkit.org/show_bug.cgi?id=50856
In the process of testing this, I found that the existing
create_bug code was wrong. I also found that existing
unit tests for create-rollout used invalid values
for options.blocks. I fixed both issues and tested.
* Scripts/webkitpy/tool/bot/flakytestreporter.py:
* Scripts/webkitpy/tool/bot/flakytestreporter_unittest.py:
* Scripts/webkitpy/tool/commands/commandtest.py:
* Scripts/webkitpy/tool/commands/download_unittest.py:
* Scripts/webkitpy/tool/commands/upload_unittest.py:
* Scripts/webkitpy/tool/mocktool.py:
* Scripts/webkitpy/tool/steps/options.py:
2010-12-10 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
Teach webkitpy how to follow duplicate chains when posting comments on flake bugs
https://bugs.webkit.org/show_bug.cgi?id=50853
I also discovered when doing this that the code was posting
the comment on the wrong bug, but that's fixed here too.
* Scripts/webkitpy/common/net/bugzilla/bug.py:
* Scripts/webkitpy/common/net/bugzilla/bugzilla.py:
* Scripts/webkitpy/tool/bot/flakytestreporter.py:
* Scripts/webkitpy/tool/bot/flakytestreporter_unittest.py:
* Scripts/webkitpy/tool/commands/queues_unittest.py:
* Scripts/webkitpy/tool/mocktool.py:
2010-12-10 Anders Carlsson <andersca@apple.com>
Reviewed by John Sullivan.
Make WKContextGetStatistics gather global statistics
https://bugs.webkit.org/show_bug.cgi?id=50850
Remove the code that gets the statistics and fills in the per context
statistics information for now.
* MiniBrowser/mac/BrowserStatisticsWindowController.m:
(-[BrowserStatisticsWindowController refreshStatistics:]):
2010-12-10 Eric Seidel <eric@webkit.org>
Unreviewed.
commit-queue should report port/platform information when commenting on flaky test bugs
https://bugs.webkit.org/show_bug.cgi?id=50839
Renamed platform.py to platforminfo.py. This broke import platform in executive.py
which was only used by new run webkit tests (and clearly not unit tested).
* Scripts/webkitpy/common/system/platforminfo.py: Renamed from WebKitTools/Scripts/webkitpy/common/system/platform.py.
* Scripts/webkitpy/tool/main.py:
* Scripts/webkitpy/tool/mocktool.py:
2010-12-10 Eric Seidel <eric@webkit.org>
Reviewed by Tony Chang.
commit-queue should report port/platform information when commenting on flaky test bugs
https://bugs.webkit.org/show_bug.cgi?id=50839
This was a suggestion from Tony Chang this morning.
I added a platform.py class so I could easily mock the platform call,
but that may not be the final solution for this mocking.
We'll try it and see.
* Scripts/webkitpy/common/system/platform.py: Added.
* Scripts/webkitpy/tool/bot/flakytestreporter.py:
* Scripts/webkitpy/tool/bot/flakytestreporter_unittest.py:
* Scripts/webkitpy/tool/commands/queues_unittest.py:
* Scripts/webkitpy/tool/main.py:
* Scripts/webkitpy/tool/mocktool.py:
2010-12-10 Krithigassree Sambamurthy <krithigassree.sambamurthy@nokia.com>
Reviewed by Joseph Pecoraro.
Bug 43455 - [Qt]: Implement Application Cache Quotas
https://bugs.webkit.org/show_bug.cgi?id=43455
Introduce functions to allow new appcache layout test origin-quota.html
to work correctly under Qt.
* DumpRenderTree/qt/DumpRenderTreeQt.cpp:
(WebCore::DumpRenderTree::DumpRenderTree):
(WebCore::DumpRenderTree::dumpApplicationCacheQuota):
* DumpRenderTree/qt/DumpRenderTreeQt.h:
Introduce functions to allow new appcache layout test origin-quota.html
to be successful under qt. Changes required because LayoutTestControllerQt
does not inherit from LayoutController.
* DumpRenderTree/qt/LayoutTestControllerQt.cpp:
(LayoutTestController::reset):
(LayoutTestController::clearAllApplicationCaches):
(LayoutTestController::setApplicationCacheOriginQuota):
* DumpRenderTree/qt/LayoutTestControllerQt.h:
(LayoutTestController::shouldDumpApplicationCacheDelegateCallbacks):
(LayoutTestController::dumpApplicationCacheDelegateCallbacks):
2010-12-10 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
Move buildbot.py into its own module so we can split it out into one-file-per-class
https://bugs.webkit.org/show_bug.cgi?id=50806
We're adding more buildbot logic these days, so it makes sense
to give buildbot its own module.
* Scripts/webkitpy/common/net/buildbot/__init__.py: Added.
* Scripts/webkitpy/common/net/buildbot/buildbot.py: Renamed from WebKitTools/Scripts/webkitpy/common/net/buildbot.py.
* Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py: Renamed from WebKitTools/Scripts/webkitpy/common/net/buildbot_unittest.py.
* Scripts/webkitpy/common/net/failuremap.py:
* Scripts/webkitpy/common/net/regressionwindow.py:
* Scripts/webkitpy/tool/commands/rebaseline.py:
2010-12-10 Tony Chang <tony@chromium.org>
Reviewed by Eric Seidel.
Use FileSystem::remove when cleaning up http lock files
https://bugs.webkit.org/show_bug.cgi?id=50830
* Scripts/webkitpy/layout_tests/port/http_lock.py:
2010-12-10 Joone Hur <joone@kldp.org>
Reviewed by Martin Robinson.
[GTK] DRT needs layoutTestController.setCacheModel
https://bugs.webkit.org/show_bug.cgi?id=50705
webkit_set_cache_mode() is called to set the cache model.
* DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
(LayoutTestController::setCacheModel): Set the cache model.
2010-12-10 Tony Chang <tony@chromium.org>
Reviewed by Ojan Vafai.
[chromium] remove --use-drt and add --use-test-shell
https://bugs.webkit.org/show_bug.cgi?id=50701
* Scripts/webkitpy/common/config/ports.py:
* Scripts/webkitpy/common/config/ports_unittest.py:
* Scripts/webkitpy/layout_tests/port/chromium.py:
* Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
* Scripts/webkitpy/layout_tests/port/chromium_gpu_unittest.py:
* Scripts/webkitpy/layout_tests/port/chromium_linux.py:
* Scripts/webkitpy/layout_tests/port/chromium_mac.py:
* Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
* Scripts/webkitpy/layout_tests/port/chromium_win.py:
* Scripts/webkitpy/layout_tests/port/webkit.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
* Scripts/webkitpy/style/checkers/test_expectations.py:
* Scripts/webkitpy/tool/commands/queries.py:
2010-12-10 John Knottenbelt <jknotten@chromium.org>
Reviewed by Steve Block.
[Chromium] Implement mocks for client-based geolocation
https://bugs.webkit.org/show_bug.cgi?id=46895
* DumpRenderTree/chromium/LayoutTestController.cpp:
(LayoutTestController::setGeolocationPermission):
(LayoutTestController::setMockGeolocationPosition):
(LayoutTestController::setMockGeolocationError):
* DumpRenderTree/chromium/LayoutTestController.h:
* DumpRenderTree/chromium/WebViewHost.cpp:
(WebViewHost::geolocationClient):
(WebViewHost::geolocationClientMock):
(WebViewHost::reset):
* DumpRenderTree/chromium/WebViewHost.h:
2010-12-10 Alejandro G. Castro <alex@igalia.com>
Reviewed by Martin Robinson.
[GTK] Add hasSpellingMarker support to the DRT
https://bugs.webkit.org/show_bug.cgi?id=50739
Implemented the hasSpellingMarker to support spelling tests.
* DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
(LayoutTestController::hasSpellingMarker):
2010-12-10 Eric Seidel <eric@webkit.org>
Reviewed by Ojan Vafai.
webkit-patch: not possible to use build-and-test with local commits
https://bugs.webkit.org/show_bug.cgi?id=33378
Make --no-clean not even check if we have local commits.
It's unclear to me why the code was originally written this way.
I was unable to dig up a reason from svn history.
* Scripts/webkitpy/tool/steps/cleanworkingdirectory.py:
* Scripts/webkitpy/tool/steps/cleanworkingdirectory_unittest.py: Copied from WebKitTools/Scripts/webkitpy/tool/steps/cleanworkingdirectory.py.
2010-12-10 Mario Sanchez Prada <msanchez@igalia.com>
Reviewed by Martin Robinson.
[GTK] AX: implement isEnabled in DRT
https://bugs.webkit.org/show_bug.cgi?id=50814
Implement AccessibilityUIElement::isEnabled() for GTK.
* DumpRenderTree/gtk/AccessibilityUIElementGtk.cpp:
(checkElementState): New function to refactor common code.
(AccessibilityUIElement::isEnabled): Implement this by checking
the current state of the AtkObject.
(AccessibilityUIElement::isSelected): Use the new
checkElementState function.
2010-12-10 Jocelyn Turcotte <jocelyn.turcotte@nokia.com>
Reviewed by Kenneth Rohde Christiansen.
[Qt][WK2] Add the QWKContext API class to wrap the WebContext.
https://bugs.webkit.org/show_bug.cgi?id=50750
* MiniBrowser/qt/BrowserView.cpp:
(BrowserView::BrowserView):
* MiniBrowser/qt/BrowserView.h:
* WebKitTestRunner/qt/PlatformWebViewQt.cpp:
(WTR::WebView::WebView):
2010-12-10 Joone Hur <joone@kldp.org>
Reviewed by Xan Lopez.
[GTK] Unskipping some test cases in userscripts dependent on addUserStyleSheet
https://bugs.webkit.org/show_bug.cgi?id=50808
Unskipping the following test cases:
userscripts/mixed-case-stylesheet.html
userscripts/simple-stylesheet.html
userscripts/user-style-all-frames.html
* DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
(LayoutTestController::addUserStyleSheet):
2010-12-10 Joone Hur <joone@kldp.org>
Reviewed by Eric Seidel.
[GTK] Add DRT support for pageProperty, isPageBoxVisible, pageSizeAndMarginsInPixels, and addUserStyleSheet
https://bugs.webkit.org/show_bug.cgi?id=50783
This patch allows to unskip the following test cases:
printing/page-rule-selection.html
printing/page-format-data.html
* DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
(LayoutTestController::pageProperty): Added.
(LayoutTestController::isPageBoxVisible): Ditto.
(LayoutTestController::pageSizeAndMarginsInPixels): Ditto.
(LayoutTestController::addUserStyleSheet): Ditto.
2010-12-10 Jocelyn Turcotte <jocelyn.turcotte@nokia.com>
Reviewed by Kenneth Rohde Christiansen.
[Qt] Make QtTestBrowser spawn QNAM in a secondary thread.
A menu allows to switch QNAM back to the same thread.
* QtTestBrowser/launcherwindow.cpp:
(LauncherWindow::initializeView):
(LauncherWindow::createChrome):
(LauncherWindow::toggleThreadedQnam):
* QtTestBrowser/launcherwindow.h:
(WindowOptions::WindowOptions):
* QtTestBrowser/webpage.cpp:
(WebPage::WebPage):
(WebPage::setQnamThreaded):
* QtTestBrowser/webpage.h:
(QnamThread::QnamThread):
(QnamThread::~QnamThread):
(QnamThread::networkAccessManager):
(QnamThread::run):
2010-12-10 Martin Robinson <mrobinson@igalia.com>
Reviewed by Xan Lopez.
[Gtk] style="font-family: courier" makes text disappear
https://bugs.webkit.org/show_bug.cgi?id=47452
Add a test that verifies that fonts without valid charmaps are
never selected.
* DumpRenderTree/gtk/DumpRenderTree.cpp:
(initializeFonts): Initialize DRT with our font (derived from Ahem)
that has no valid charmap.
* DumpRenderTree/gtk/fonts/FontWithNoValidEncoding.fon: Added.
* DumpRenderTree/gtk/fonts/fonts.conf: Updated settings to override users'
settings which may disable selection of bitmap fonts.
2010-12-10 Eric Seidel <eric@webkit.org>
Unreviewed.
The commit-queue should file bugs about flaky tests it encounters
https://bugs.webkit.org/show_bug.cgi?id=50803
create_bug_for_flaky_test was insufficiently tested, thus buggy.
Now it's tested.
* Scripts/webkitpy/tool/bot/flakytestreporter.py:
* Scripts/webkitpy/tool/bot/flakytestreporter_unittest.py:
2010-12-10 Eric Seidel <eric@webkit.org>
Unreviewed. Somehow this last-minute edit didn't quite get committed.
The commit-queue should file bugs about flaky tests it encounters
https://bugs.webkit.org/show_bug.cgi?id=50803
* Scripts/webkitpy/tool/bot/flakytestreporter.py:
2010-12-09 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
The commit-queue should file bugs about flaky tests it encounters
https://bugs.webkit.org/show_bug.cgi?id=50803
This change got a bit big. I also added a new config.urls file
because I needed to share the view_source_url code with committervalidator.py.
This adds a new class FlakyTestReporter which holds all the logic about
reporting flaky tests to bugzilla.
Right now this code knows how to look up bugs for flaky tests.
If it can't find a bug filed from the commit-queue, it will open a new
one, ccing the relevant people and adding information about the failure.
It is not yet smart enough to chase down duplicate chains, or to include
the actual failure diff. But those can be added in later iterations.
* Scripts/webkitpy/common/checkout/api.py:
* Scripts/webkitpy/common/checkout/changelog.py:
* Scripts/webkitpy/common/checkout/commitinfo.py:
* Scripts/webkitpy/common/config/committervalidator.py:
* Scripts/webkitpy/common/config/urls.py: Copied from WebKitTools/Scripts/webkitpy/tool/comments.py.
* Scripts/webkitpy/common/net/bugzilla/bugzilla.py:
* Scripts/webkitpy/tool/bot/flakytestreporter.py: Added.
* Scripts/webkitpy/tool/bot/flakytestreporter_unittest.py: Copied from WebKitTools/Scripts/webkitpy/tool/comments.py.
* Scripts/webkitpy/tool/bot/irc_command.py:
* Scripts/webkitpy/tool/bot/sheriff.py:
* Scripts/webkitpy/tool/commands/download.py:
* Scripts/webkitpy/tool/commands/queues.py:
* Scripts/webkitpy/tool/commands/queues_unittest.py:
* Scripts/webkitpy/tool/comments.py:
* Scripts/webkitpy/tool/mocktool.py:
* Scripts/webkitpy/tool/steps/commit.py:
2010-12-09 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
Teach webkit-patch how to search bugzilla
https://bugs.webkit.org/show_bug.cgi?id=50500
This is a step towards teaching webkitpy how to file
new bugs for flaky tests and update them when new flakes occur.
* Scripts/webkitpy/common/net/bugzilla/bugzilla.py:
* Scripts/webkitpy/common/net/bugzilla/bugzilla_unittest.py:
* Scripts/webkitpy/tool/commands/__init__.py:
* Scripts/webkitpy/tool/commands/bugsearch.py: Added.
2010-12-09 Adam Barth <abarth@webkit.org>
Reviewed by Ojan Vafai.
commit-queue errors out due to local commits
https://bugs.webkit.org/show_bug.cgi?id=50766
We seem to be tripping over having local commits in the working copy a
lot recently. I don't quite fully understand what the issue is, but
this patch attempts to solve the problem by adding an explicit clean
step before processing a patch. Previously, we did the cleaning as
part of the apply-attachment step, so this might not actually fix the
problem.
* Scripts/webkitpy/tool/bot/commitqueuetask.py:
* Scripts/webkitpy/tool/bot/commitqueuetask_unittest.py:
* Scripts/webkitpy/tool/commands/queues_unittest.py:
2010-12-09 Dirk Pranke <dpranke@chromium.org>
Reviewed by Ojan Vafai.
Make new-run-webkit-tests --lint-test-files log slightly better
messages and return -1 if lint fails. Remove the
'suppress_errors' keyword param to the TestExpectationsFile
class, and clean up logging and exception raising for error
handling. Also add more unit tests and clean up the unit test code a bit.
https://bugs.webkit.org/show_bug.cgi?id=50205
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
* Scripts/webkitpy/layout_tests/layout_package/test_expectations.py
* Scripts/webkitpy/layout_tests/layout_package/test_expectations_unittest.py
* Scripts/webkitpy/layout_tests/port/chromium.py:
* Scripts/webkitpy/style/checkers/test_expectations.py:
* Scripts/webkitpy/style/checkers/test_expectations_unittest.py:
2010-12-09 Tony Chang <tony@chromium.org>
Unreviewed, fix for windows code. We were catching the wrong
exception in Windows.
* Scripts/webkitpy/common/system/filesystem.py:
2010-12-09 Antonio Gomes <agomes@rim.com>
Reviewed by Martin Robinson.
[Gtk] Populate DumpRenderTreeSupportGtk (part IV)
https://bugs.webkit.org/show_bug.cgi?id=48429
Adjust call sites of webkit_thread_count to use the DumpRenderTreeSupportGtk version now.
* DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
(LayoutTestController::workerThreadCount):
2010-12-06 Philippe Normand <pnormand@igalia.com>
Reviewed by Martin Robinson.
[GTK] Support the Mozilla-style Fullscreen Javascript API
https://bugs.webkit.org/show_bug.cgi?id=50572
Enable the Javascript Fullscreen support in the GtkLauncher. Also
make DRT update the Chrome when the WebView settings have been
reset to consistent values.
* DumpRenderTree/gtk/DumpRenderTree.cpp:
(resetDefaultsToConsistentValues):
* GtkLauncher/main.c:
(create_window):
* Scripts/build-webkit:
2010-12-09 Koan-Sin Tan <koansin.tan@gmail.com>
Reviewed by Tor Arne Vestbø
[Qt] run-launcher doesn't start the browser on Mac OS X 10.6
https://bugs.webkit.org/show_bug.cgi?id=50742
check not only QtWebKit.framework/QtWebKit but also libQtWebKit.dylib
* Scripts/webkitdirs.pm:
2010-12-09 Joone Hur <joone@kldp.org>
Reviewed by Martin Robinson.
[GTK] Support for viewport meta tag
https://bugs.webkit.org/show_bug.cgi?id=45443
Add support for testing the viewport properties for WebKitGtk.
* DumpRenderTree/LayoutTestController.cpp:
(dumpConfigurationForViewportCallback): Added this JS callback to set the size of the visible viewport.
(LayoutTestController::staticFunctions):
* DumpRenderTree/LayoutTestController.h:
* DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
(LayoutTestController::dumpConfigurationForViewport): Added this function to test the viewport properties.
* DumpRenderTree/mac/LayoutTestControllerMac.mm:
(LayoutTestController::dumpConfigurationForViewport): Ditto.
* DumpRenderTree/qt/LayoutTestControllerQt.cpp:
(LayoutTestController::dumpConfigurationForViewport): Ditto.
* DumpRenderTree/win/LayoutTestControllerWin.cpp:
(LayoutTestController::dumpConfigurationForViewport): Ditto.
* DumpRenderTree/wx/LayoutTestControllerWx.cpp:
(LayoutTestController::dumpConfigurationForViewport): Ditto.
2010-12-08 William Siegrist <wsiegrist@apple.com>
Reviewed by Eric Seidel.
Migrate buildbot.py from xmlrpc to json.
https://bugs.webkit.org/show_bug.cgi?id=50647
* Scripts/webkitpy/common/net/buildbot.py:
2010-12-08 Tony Chang <tony@chromium.org>
Reviewed by Eric Seidel.
make starting the websocket server more reliable on windows
https://bugs.webkit.org/show_bug.cgi?id=50712
* Scripts/webkitpy/common/system/filesystem.py:
(remove): Substitute method for os.remove to retry on error on Windows
* Scripts/webkitpy/common/system/filesystem_unittest.py:
* Scripts/webkitpy/layout_tests/port/http_server_base.py: Use FileSystem.remove
2010-12-08 Jessie Berlin <jberlin@apple.com>
Reviewed by Adam Roben.
WebKit2: Implement WebChromeClient::exceededDatabaseQuota
https://bugs.webkit.org/show_bug.cgi?id=50656
* MiniBrowser/mac/BrowserWindowController.m:
(-[BrowserWindowController awakeFromNib]):
Indicate that exceededDatabaseQuota is not implemented.
* MiniBrowser/win/BrowserView.cpp:
(BrowserView::create):
Ditto.
* WebKitTestRunner/TestController.cpp:
(WTR::createOtherPage):
Ditto.
(WTR::TestController::initialize):
Ditto.
2010-12-08 Antonio Gomes <agomes@rim.com>
Reviewed by Martin Robinson.
[Gtk] Populate DumpRenderTreeSupportGtk (part III)
https://bugs.webkit.org/show_bug.cgi?id=48429
Moved the following methods from webkitprivate.h to DumpRenderTreeSupportGtk.cpp|h:
* webkit_gc_collect_javascript_objects;
* webkit_gc_collect_javascript_objects_on_alternate_thread;
* webkit_gc_count_javascript_objects;
* webkit_web_frame_layout.
These were all only used by GTK+ DRT.
* DumpRenderTree/gtk/EventSender.cpp:
(dispatchEvent):
* DumpRenderTree/gtk/GCControllerGtk.cpp:
(GCController::collect):
(GCController::collectOnAlternateThread):
(GCController::getJSObjectCount):
2010-12-07 Tony Chang <tony@chromium.org>
Reviewed by Ojan Vafai.
[chromium] force NRWT to use DRT
https://bugs.webkit.org/show_bug.cgi?id=50359
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2010-12-06 Sam Weinig <sam@webkit.org>
Reviewed by Anders Carlsson.
Generalize didChangeLocationWithinPageForFrame for all same document navigations
https://bugs.webkit.org/show_bug.cgi?id=50584
* MiniBrowser/mac/BrowserWindowController.m:
(didSameDocumentNavigationForFrame):
(-[BrowserWindowController awakeFromNib]):
(-[BrowserWindowController didSameDocumentNavigationForFrame:]):
* TestWebKitAPI/Tests/WebKit2/PageLoadDidChangeLocationWithinPageForFrame.cpp:
(TestWebKitAPI::didSameDocumentNavigationForFrame):
(TestWebKitAPI::TEST):
* WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
(WTR::InjectedBundlePage::InjectedBundlePage):
(WTR::InjectedBundlePage::didSameDocumentNavigationForFrame):
* WebKitTestRunner/InjectedBundle/InjectedBundlePage.h:
* WebKitTestRunner/TestController.cpp:
(WTR::TestController::initialize):
2010-12-06 Kinuko Yasuda <kinuko@chromium.org>
Reviewed by Ojan Vafai.
Make TestResultsServer return 404 for non-existing results files.
https://bugs.webkit.org/show_bug.cgi?id=50581
* TestResultServer/handlers/testfilehandler.py:
2010-12-06 Tony Chang <tony@chromium.org>
Reviewed by Ojan Vafai.
[chromium] fix an assert hit in DRT
https://bugs.webkit.org/show_bug.cgi?id=50575
This class uses WTF's RefPtr, but expects chrome's base::RefPtr
semantics.
Otherwise, we hit ASSERTION FAILED: !m_adoptionIsRequired
(third_party/WebKit/JavaScriptCore/wtf/RefCounted.h:37 void
WTF::RefCountedBase::ref()).
* DumpRenderTree/chromium/TestWebWorker.h:
(TestWebWorker::TestWebWorker):
2010-12-02 Ojan Vafai <ojan@chromium.org>
Reviewed by Eric Seidel.
make webkit-patch command work when the git branch is not synced to the remote svn branch
https://bugs.webkit.org/show_bug.cgi?id=50424
* Scripts/webkitpy/common/checkout/scm.py:
* Scripts/webkitpy/common/checkout/scm_unittest.py:
2010-12-06 Patrick Gansterer <paroga@webkit.org>
Reviewed by Andreas Kling.
[WINCE] Add build system
https://bugs.webkit.org/show_bug.cgi?id=50522
* CMakeListsWinCE.txt: Added.
2010-12-03 Mihai Parparita <mihaip@chromium.org>
Reviewed by Tony Chang.
Allow the Chromium port to have Leopard-specific baselines
https://bugs.webkit.org/show_bug.cgi?id=50506
* Scripts/webkitpy/layout_tests/port/chromium_mac.py:
2010-12-03 Tony Chang <tony@chromium.org>
Reviewed by Kent Tamura.
[chromium] fix 2 bugs with inspector tests in DRT
https://bugs.webkit.org/show_bug.cgi?id=50492
Tasks can outlive the tasklist (even when canceled) so this was
causing a crash when ~WebTask() ran. Avoid this by unregistering
when a task is canceled.
Also fix an assert when closing devtool windows by copying some
logic from test_shell.
* DumpRenderTree/chromium/Task.cpp:
(WebTask::~WebTask):
(TaskList::revokeAll):
* DumpRenderTree/chromium/Task.h: Canceling a task now removes it from
the tasklist (since the task can outlive the tasklist).
* DumpRenderTree/chromium/TestShell.h:
(TestShell::devToolsWebView):
* DumpRenderTree/chromium/WebViewHost.cpp:
(WebViewHost::~WebViewHost): Don't load about:blank when closing
a window if the window has devtools loaded. This avoids an
ASSERT and matches test_shell.
2010-12-06 Alejandro G. Castro <alex@igalia.com>
Reviewed by Martin Robinson.
[GTK] Fix GtkLauncher.c style, use webkit style
https://bugs.webkit.org/show_bug.cgi?id=50542
* GtkLauncher/main.c:
(activateUriEntryCb):
(updateTitle):
(linkHoverCb):
(notifyTitleCb):
(notifyLoadStatusCb):
(notifyProgressCb):
(destroyCb):
(goBackCb):
(goForwardCb):
(create_webViewCb):
(webViewReadyCb):
(closeWebViewCb):
(createBrowser):
(createStatusbar):
(createToolbar):
(createWindow):
(main):
2010-12-04 Antonio Gomes <agomes@rim.com>
Reviewed by Martin Robinson.
[Gtk] Populate DumpRenderTreeSupportGtk (Part II)
https://bugs.webkit.org/show_bug.cgi?id=48429
Moved more private method declared with WEBKIT_API to
DumpRenderTreeSupportGtk, since they were only being used by DRT.
* DumpRenderTree/gtk/DumpRenderTree.cpp:
(resetDefaultsToConsistentValues):
* DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
(LayoutTestController::addOriginAccessWhitelistEntry):
(LayoutTestController::execCommand):
(LayoutTestController::isCommandEnabled):
2010-12-04 Antonio Gomes <agomes@rim.com>
Reviewed by Andreas Kling.
Removed extern webkit_web_view_set_group_name from gtk/DumpRenderTree.cpp
Method was removed by http://trac.webkit.org/changeset/71604
* DumpRenderTree/gtk/DumpRenderTree.cpp:
2010-12-05 Alejandro G. Castro <alex@igalia.com>
Reviewed by Xan Lopez.
[GTK] Fix compilation warnings reported by clang
https://bugs.webkit.org/show_bug.cgi?id=50252
* GtkLauncher/main.c:
(create_window): The create_statusbar function does not have
parameters.
2010-12-05 Robert Hogan <robert@webkit.org>
Reviewed by Andreas Kling.
[Qt] Zero-sized font does not yet work
Fix fast/text/font-size-zero.html
Copy other ports by setting minimumFontSize to 0 for DRT runs.
https://bugs.webkit.org/show_bug.cgi?id=49759
* DumpRenderTree/qt/DumpRenderTreeQt.cpp:
(WebCore::WebPage::WebPage):
2010-12-04 Sam Weinig <sam@webkit.org>
Reviewed by Jon Honeycutt.
WebKit2: Need client functions to notify app when the url changes due to a fragment navigation
https://bugs.webkit.org/show_bug.cgi?id=50511
Update tools for new didChangeLocationWithinPageForFrame client function (and changed bundle
signature) and added a test for the functionality (PageLoadDidChangeLocationWithinPageForFrame.cpp)
* MiniBrowser/mac/BrowserWindowController.m:
(didChangeLocationWithinPageForFrame):
(-[BrowserWindowController awakeFromNib]):
(-[BrowserWindowController didChangeLocationWithinPageForFrame:]):
* TestWebKitAPI/PlatformUtilities.cpp:
(TestWebKitAPI::Util::toWK):
* TestWebKitAPI/PlatformUtilities.h:
(TestWebKitAPI::Util::adoptWK):
* TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
* TestWebKitAPI/Tests/WebKit2/PageLoadDidChangeLocationWithinPageForFrame.cpp: Added.
(TestWebKitAPI::nullJavaScriptCallback):
(TestWebKitAPI::didFinishLoadForFrame):
(TestWebKitAPI::didChangeLocationWithinPageForFrame):
(TestWebKitAPI::TEST):
* TestWebKitAPI/Tests/WebKit2/file-with-anchor.html: Added.
* TestWebKitAPI/win/TestWebKitAPI.vcproj:
* TestWebKitAPI/win/copy-resources.cmd:
* WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
(WTR::InjectedBundlePage::InjectedBundlePage):
(WTR::InjectedBundlePage::didChangeLocationWithinPageForFrame):
* WebKitTestRunner/InjectedBundle/InjectedBundlePage.h:
* WebKitTestRunner/TestController.cpp:
(WTR::TestController::initialize):
2010-11-08 Antonio Gomes <agomes@rim.com>
Reviewed by Martin Robinson.
[Gtk] Populate DumpRenderTreeSupportGtk
https://bugs.webkit.org/show_bug.cgi?id=48429
Made the previous calls to webkit_web_frame* functions defined as private
APIs in webkitprivate.h go through DRTSupportGtk.
* DumpRenderTree/gtk/AccessibilityControllerGtk.cpp:
(AccessibilityController::focusedElement):
* DumpRenderTree/gtk/DumpRenderTree.cpp:
(dumpFramesAsText):
(resetDefaultsToConsistentValues):
(dump):
(webViewDocumentLoadFinished):
* DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
(LayoutTestController::counterValueForElementById):
(LayoutTestController::pageNumberForElementById):
(LayoutTestController::numberOfPages):
(LayoutTestController::pauseAnimationAtTimeOnElementWithId):
(LayoutTestController::pauseTransitionAtTimeOnElementWithId):
(LayoutTestController::sampleSVGAnimationForElementAtTime):
(LayoutTestController::numberOfActiveAnimations):
(LayoutTestController::suspendAnimations):
(LayoutTestController::resumeAnimations):
(LayoutTestController::markerTextForListItem):
2010-12-03 Mihai Parparita <mihaip@chromium.org>
Reviewed by Tony Chang.
Rebaseline server: use pretty text diff output
https://bugs.webkit.org/show_bug.cgi?id=50484
Serve *-pretty-diff.html for text diff output.
* Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html:
* Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css:
* Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
* Scripts/webkitpy/tool/commands/rebaselineserver.py:
2010-12-03 Sheriff Bot <webkit.review.bot@gmail.com>
Unreviewed, rolling out r73302.
http://trac.webkit.org/changeset/73302
https://bugs.webkit.org/show_bug.cgi?id=50499
Causes crashes in debug LayoutTests (Requested by xan_ on
#webkit).
* DumpRenderTree/chromium/AccessibilityUIElement.cpp:
(AccessibilityUIElement::AccessibilityUIElement):
(AccessibilityUIElement::parentElementCallback):
* DumpRenderTree/chromium/AccessibilityUIElement.h:
* DumpRenderTree/chromium/CppBoundClass.cpp:
* DumpRenderTree/chromium/CppBoundClass.h:
2010-12-03 Sam Weinig <sam@webkit.org>
Reviewed by Maciej Stachowiak.
Enable <a ping> for Mac/Windows/WebKit2 builds
<rdar://problem/8504473>
https://bugs.webkit.org/show_bug.cgi?id=50488
* TestWebKitAPI/Tests/WebKit2/WKPreferences.cpp:
(TestWebKitAPI::TEST): Test that the default value of
WKPreferencesGetHyperlinkAuditingEnabled is true.
2010-12-03 Chris Guillory <chris.guillory@google.com>
Reviewed by Chris Fleizach.
Include the FrameView widget of a RenderWidget in the accessibility tree.
https://bugs.webkit.org/show_bug.cgi?id=49106
* DumpRenderTree/chromium/AccessibilityUIElement.cpp:
(AccessibilityUIElement::AccessibilityUIElement):
(AccessibilityUIElement::parentElementCallback):
(AccessibilityUIElement::isEqualCallback):
* DumpRenderTree/chromium/AccessibilityUIElement.h:
* DumpRenderTree/chromium/CppBoundClass.cpp:
(CppBoundClass::getFromCppVariant):
* DumpRenderTree/chromium/CppBoundClass.h:
2010-12-03 Mihai Parparita <mihaip@chromium.org>
Reviewed by Tony Chang.
Rebaseline server: organize tests by state, add dry run mode
https://bugs.webkit.org/show_bug.cgi?id=50473
Group tests in the menu by state, so that it's easier to see after
processing the queue which failed.
Add support for the --dry-run flag so that it's easier to test changes
such as this (stubs out filesystem and SCM operations).
* Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
* Scripts/webkitpy/tool/commands/data/rebaselineserver/queue.js:
* Scripts/webkitpy/tool/commands/rebaselineserver.py:
2010-12-03 Brady Eidson <beidson@apple.com>
Reviewed by Anders Carlsson.
Groundwork for <rdar://problem/7660733> and https://bugs.webkit.org/show_bug.cgi?id=50191
WebKit2 Authentication Support
Keep these builds working:
* MiniBrowser/mac/BrowserWindowController.m:
(-[BrowserWindowController awakeFromNib]):
* WebKitTestRunner/TestController.cpp:
(WTR::TestController::initialize):
2010-12-02 Mihai Parparita <mihaip@chromium.org>
Reviewed by Tony Chang.
Rebaseline server: move existing baselines
https://bugs.webkit.org/show_bug.cgi?id=50421
Implement moving of existing baselines: move all files that are about
to be overwriten by update baselines.
Also fix a typo in the UI JS.
* Scripts/webkitpy/tool/commands/data/rebaselineserver/queue.js:
* Scripts/webkitpy/tool/commands/rebaselineserver.py:
* Scripts/webkitpy/tool/commands/rebaselineserver_unittest.py:
2010-12-03 Sheriff Bot <webkit.review.bot@gmail.com>
Unreviewed, rolling out r73211.
http://trac.webkit.org/changeset/73211
https://bugs.webkit.org/show_bug.cgi?id=50450
Broke Chromium's "Webkit Win (dbg)(2)" bot. (Requested by
yutak on #webkit).
* Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
2010-12-03 Sheriff Bot <webkit.review.bot@gmail.com>
Unreviewed, rolling out r73222.
http://trac.webkit.org/changeset/73222
https://bugs.webkit.org/show_bug.cgi?id=50449
r73211 seemed to break Chromium's "Webkit Win (dbg)(2)" bot.
(Requested by yutak on #webkit).
* Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
* Scripts/webkitpy/layout_tests/layout_package/message_broker.py:
* Scripts/webkitpy/layout_tests/layout_package/message_broker_unittest.py:
* Scripts/webkitpy/layout_tests/port/base.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2010-12-03 Sheriff Bot <webkit.review.bot@gmail.com>
Unreviewed, rolling out r73228.
http://trac.webkit.org/changeset/73228
https://bugs.webkit.org/show_bug.cgi?id=50448
r73211 seemed to break Chromium's "Webkit Win (dbg)(2)" bot.
(Requested by yutak on #webkit).
* Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
* Scripts/webkitpy/layout_tests/test_types/image_diff.py:
* Scripts/webkitpy/layout_tests/test_types/test_type_base.py:
* Scripts/webkitpy/layout_tests/test_types/test_type_base_unittest.py:
* Scripts/webkitpy/layout_tests/test_types/text_diff.py:
2010-12-03 Sheriff Bot <webkit.review.bot@gmail.com>
Unreviewed, rolling out r73231.
http://trac.webkit.org/changeset/73231
https://bugs.webkit.org/show_bug.cgi?id=50443
r73211 seemed to broke Chromium's "Webkit Win (dbg)(2)" bot.
(Requested by yutak on #webkit).
* Scripts/webkitpy/layout_tests/layout_package/message_broker.py:
* Scripts/webkitpy/layout_tests/layout_package/message_broker_unittest.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2010-12-03 David Levin <levin@chromium.org>
Reviewed by Shinichiro Hamaji.
check-webkit-style: false positive reported for #if macro
https://bugs.webkit.org/show_bug.cgi?id=48242
* Scripts/webkitpy/style/checkers/cpp.py: Adjusted check to avoid
all preprocessor commands.
* Scripts/webkitpy/style/checkers/cpp_unittest.py: Added test.
2010-12-02 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
Update the EWSTools scripts to use modern-style git repos
https://bugs.webkit.org/show_bug.cgi?id=50402
* EWSTools/create-webkit-git:
* EWSTools/start-commit-queue.sh:
* EWSTools/start-queue.sh:
2010-12-02 Simon Fraser <simon.fraser@apple.com>
Fix Chromium Linux plugin tests by not logging
unconditionally from the event handling code.
* DumpRenderTree/TestNetscapePlugIn/main.cpp:
(handleEventX11):
2010-12-02 Dirk Pranke <dpranke@chromium.org>
Reviewed by Tony Chang.
nrwt multiprocessing - move logic back into run_webkit_tests
This change moves a bunch of logic that I had put into
message_broker back into run_webkit_tests, in a slightly
different format. WorkerMessageBroker needed to become less aware of
the logic the TestRunner class uses, and more generic.
Eventually the MessageBroker will only do generic messaging and
thread/process-pooling, and (almost) all of the
run-webkit-tests-specific logic will be moved to
run_webkit_tests.py and dump_render_tree_thread.py.
The biggest changes are that the Broker can now start a single
worker, but the responsibility for starting all of them is pushed
back to the TestRunner (Manager), and the logic for checking if
the threads are done or wedged is moved back to TestRunner. We
also remove WorkerMessageBroker.cleanup (not needed) and
cancel_workers (they have to be cancelled individually).
The message_broker is now encapsulated inside
TestRunner._run_tests(); it only needs to exist while actually
running the tests.
Also, delete a bunch of tests in message_broker_unittest that no
longer make much sense.
This patch depends on bug 50372.
https://bugs.webkit.org/show_bug.cgi?id=50374
* Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
* Scripts/webkitpy/layout_tests/layout_package/message_broker.py:
* Scripts/webkitpy/layout_tests/layout_package/message_broker_unittest.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2010-12-02 Hayato Ito <hayato@chromium.org>
Reviewed by Eric Seidel.
Refactor test_types's compare_output() method so that it takes a
TestInput object.
compare_output() will need additional information included in a TestInput
object to support reftests. This change is a pre-requirement of that.
https://bugs.webkit.org/show_bug.cgi?id=50362
* Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
* Scripts/webkitpy/layout_tests/test_types/image_diff.py:
* Scripts/webkitpy/layout_tests/test_types/test_type_base.py:
* Scripts/webkitpy/layout_tests/test_types/test_type_base_unittest.py:
* Scripts/webkitpy/layout_tests/test_types/text_diff.py:
2010-12-02 Simon Fraser <simon.fraser@apple.com>
Reviewed by Anders Carlsson.
Add the ability to run script on 'new' and 'paint' in the test plugin
https://bugs.webkit.org/show_bug.cgi?id=50425
Add "onnew" and "onpaintevent" attributes to the test plugin,
so that tests can run JS at these times.
* DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
(pluginAllocate):
* DumpRenderTree/TestNetscapePlugIn/PluginObject.h:
* DumpRenderTree/TestNetscapePlugIn/main.cpp:
(NPP_New):
(NPP_Destroy):
(handleEventCarbon):
(handleEventCocoa):
(NPP_HandleEvent):
2010-12-02 Sam Weinig <sam@webkit.org>
Fix Qt build.
* WebKitTestRunner/qt/PlatformWebViewQt.cpp:
(WTR::PlatformWebView::PlatformWebView):
2010-12-02 Sam Weinig <sam@webkit.org>
Reviewed by Anders Carlsson.
WebKit2 Preferences should be held off the PageGroup, not the Context
https://bugs.webkit.org/show_bug.cgi?id=50414
Update for API changes.
* WebKitTestRunner/PlatformWebView.h:
* WebKitTestRunner/TestController.cpp:
(WTR::createOtherPage):
(WTR::TestController::initialize):
(WTR::TestController::resetStateToConsistentValues):
* WebKitTestRunner/TestController.h:
* WebKitTestRunner/mac/PlatformWebViewMac.mm:
(WTR::PlatformWebView::PlatformWebView):
* WebKitTestRunner/win/PlatformWebViewWin.cpp:
(WTR::PlatformWebView::PlatformWebView):
2010-12-02 Dirk Pranke <dpranke@chromium.org>
Reviewed by Tony Chang.
new-run-webkit-tests: minor cleanup for multiprocessing work
This change:
* moves worker naming into TestShellThread, eliminating a
parameter to the constructor and putting the responsibility in
the correct place.
* eliminates the _WorkerState() class in message_broker, because
it turns out that state really needs to be in run_webkit_tests
* renames the Broker classes to be module-private.
* fixes a bunch of minor commenting and whitespace issues to
make subsequent patches a bit clearer.
* Adds a Port hook for default_worker_model() so that we can
accomodate the potential for different worker models on
different ports.
* merge in the fix from 50420 for the brokenness introduced in
bug 50367.
This patch depends on bug 50367.
https://bugs.webkit.org/show_bug.cgi?id=50372
* Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
* Scripts/webkitpy/layout_tests/layout_package/message_broker.py:
* Scripts/webkitpy/layout_tests/layout_package/message_broker_unittest.py:
* Scripts/webkitpy/layout_tests/port/base.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2010-12-02 Dirk Pranke <dpranke@chromium.org>
Reviewed by Tony Chang.
This change is a bunch of cleanup / refactoring of the file
below. It moves a bunch of free functions to methods on the
TestShellThread class, and restructures other methods to be
closer to the structure we'll need for the multiprocessing
rewrite.
It also makes the logic of --run-singly a *lot* easier to follow
by nestling all of the separate-thread logic into a single
routine.
There should be no semantic changes in this patch, just cleanup.
https://bugs.webkit.org/show_bug.cgi?id=50367
* Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
2010-12-02 Tony Chang <tony@chromium.org>
Unreviewed, rolling out r73164.
http://trac.webkit.org/changeset/73164
https://bugs.webkit.org/show_bug.cgi?id=50359
broke debugger tests on win
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2010-11-30 Mihai Parparita <mihaip@chromium.org>
Reviewed by Tony Chang.
Rebaseline server: add updating of baselines
https://bugs.webkit.org/show_bug.cgi?id=50305
Implements updating of baselines, where we copy -actual.* files over
the current -expected.* files. To do this, we need a
_get_actual_result_files method to get test results files and a
_rebaseline_test method to actually do the file copy and SCM operation.
_rebaseline_test logs output into a buffer, this is useful for both
showing result in the UI and for unit tests.
To make passing around of the various test environment properties
(results directory, filesystem, SCM, etc) easier, add a TestConfig
class for them.
Moving of existing baselines is not implemented yet, this patch is big
enough as it is.
* Scripts/webkitpy/common/system/filesystem.py:
* Scripts/webkitpy/common/system/filesystem_mock.py:
* Scripts/webkitpy/tool/commands/data/rebaselineserver/queue.js:
* Scripts/webkitpy/tool/commands/rebaselineserver.py:
* Scripts/webkitpy/tool/commands/rebaselineserver_unittest.py:
* Scripts/webkitpy/tool/mocktool.py:
2010-12-02 Brent Fulgham <bfulgham@webkit.org>
Unreviewed build fix after vcproj updates.
* DumpRenderTree/win/DumpRenderTreeCairo.vsprops: Make sure
that WIN_CAIRO is defined for this build.
* DumpRenderTree/win/DumpRenderTreePreBuild.cmd: Update file
to reflect new 'Debug_Cairo_CFLite' and 'Release_Cairo_CFLite'
(these were previously 'Debug_Cairo' and 'Release_Cairo'
2010-12-02 Sam Weinig <sam@webkit.org>
Reviewed by Anders Carlsson.
WebKit2 needs API to turn on popup blocking
https://bugs.webkit.org/show_bug.cgi?id=50407
* TestWebKitAPI/Tests/WebKit2/WKPreferences.cpp:
(TestWebKitAPI::TEST):
* WebKitTestRunner/TestController.cpp:
(WTR::TestController::resetStateToConsistentValues):
2010-12-01 Tony Chang <tony@chromium.org>
Reviewed by Ojan Vafai.
[chromium] force NRWT to use DRT
https://bugs.webkit.org/show_bug.cgi?id=50359
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2010-12-02 Martin Robinson <mrobinson@igalia.com>
Reviewed by Xan Lopez.
[GTK] DumpRenderTree has two copies of JavaScriptCore
https://bugs.webkit.org/show_bug.cgi?id=49877
* GNUmakefile.am: Remove the second copy of JavaScriptCore from the
LDADD list for DumpRenderTree.
2010-12-02 Sergio Villar Senin <svillar@igalia.com>
Unreviewed: Added myself to list of Committers.
* Scripts/webkitpy/common/config/committers.py:
2010-12-01 Mihai Parparita <mihaip@chromium.org>
Reviewed by Adam Barth.
build-webkit --chromium always updates DEPS and re-generates project files
https://bugs.webkit.org/show_bug.cgi?id=50340
Make build-webkit --chromium invoke update-webkit-chromium only when
--update-chromium is passed in, so that we don't add 20 seconds to all
builds.
Making all invocations update was added with r61883, with the goal of
making EWS bots always update. The bots will switch to using this flag
(as will all other webkit-patch build steps) so nothing should change
for them.
* Scripts/webkitdirs.pm:
* Scripts/webkitpy/common/config/ports.py:
* Scripts/webkitpy/common/config/ports_unittest.py:
2010-12-01 Dirk Pranke <dpranke@chromium.org>
Reviewed by Tony Chang.
new-run-webkit-tests: add a --dry-run / -n flag
https://bugs.webkit.org/show_bug.cgi?id=50045
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2010-12-01 Steve Falkenburg <sfalken@apple.com>
Reviewed by Adam Roben.
WinCairo build should not use link-time code generation (LTCG)
https://bugs.webkit.org/show_bug.cgi?id=50353
* DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
* DumpRenderTree/win/DumpRenderTree.vcproj:
* DumpRenderTree/win/ImageDiff.vcproj:
* FindSafari/FindSafari.vcproj:
* MiniBrowser/MiniBrowser.vcproj:
* TestWebKitAPI/win/TestWebKitAPI.vcproj:
* WebKitAPITest/WebKitAPITest.vcproj:
* WebKitLauncherWin/WebKitLauncherWin.vcproj:
* WebKitTestRunner/win/InjectedBundle.vcproj:
* WebKitTestRunner/win/WebKitTestRunner.vcproj:
* WinLauncher/WinLauncher.vcproj:
* record-memory-win/record-memory-win.vcproj:
2010-12-01 Martin Robinson <mrobinson@igalia.com>
Touch GetUserAgentWithNullNPPFromNPPNew.cpp in an effort to force
a rebuild of TestNetscapePlugin.la on GTK+. It seems that simply
adding the file to the sources list was not enough to force the rebuild.
* DumpRenderTree/TestNetscapePlugIn/Tests/GetUserAgentWithNullNPPFromNPPNew.cpp:
Remove an empty line at the end of this file to force a rebuild.
2010-12-01 Sam Weinig <sam@webkit.org>
Fix windows build.
* MiniBrowser/win/BrowserView.cpp:
(BrowserView::create):
2010-12-01 Martin Robinson <mrobinson@igalia.com>
Add missing file to the TestNetscapePlugin sources list after r73057.
* GNUmakefile.am: Add missing file.
2010-12-01 Sam Weinig <sam@webkit.org>
Reviewed by Anders Carlsson.
Add PageGrouping to WebKit2 API
https://bugs.webkit.org/show_bug.cgi?id=50332
- Update testing harnesses to deal with new PageGroup API.
* MiniBrowser/mac/WebBundle/WebBundleMain.m:
(WKBundleInitialize):
* TestWebKitAPI/InjectedBundleController.cpp:
(TestWebKitAPI::InjectedBundleController::initialize):
(TestWebKitAPI::InjectedBundleController::didInitializePageGroup):
(TestWebKitAPI::InjectedBundleController::initializeTestNamed):
* TestWebKitAPI/InjectedBundleController.h:
* TestWebKitAPI/InjectedBundleTest.h:
(TestWebKitAPI::InjectedBundleTest::initialize):
(TestWebKitAPI::InjectedBundleTest::didInitializePageGroup):
* TestWebKitAPI/PlatformUtilities.cpp:
(TestWebKitAPI::Util::createContextForInjectedBundleTest):
* TestWebKitAPI/PlatformUtilities.h:
* TestWebKitAPI/PlatformWebView.h:
* TestWebKitAPI/Tests/WebKit2/DocumentStartUserScriptAlertCrash.cpp:
(TestWebKitAPI::TEST):
* TestWebKitAPI/Tests/WebKit2/DocumentStartUserScriptAlertCrash_Bundle.cpp:
(TestWebKitAPI::DocumentStartUserScriptAlertCrashTest::initialize):
* TestWebKitAPI/mac/PlatformWebViewMac.mm:
(TestWebKitAPI::PlatformWebView::PlatformWebView):
* TestWebKitAPI/win/PlatformWebViewWin.cpp:
(TestWebKitAPI::PlatformWebView::PlatformWebView):
* WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
(WTR::InjectedBundle::didInitializePageGroup):
(WTR::InjectedBundle::initialize):
(WTR::InjectedBundle::beginTesting):
* WebKitTestRunner/InjectedBundle/InjectedBundle.h:
(WTR::InjectedBundle::pageGroup):
* WebKitTestRunner/InjectedBundle/LayoutTestController.cpp:
(WTR::LayoutTestController::addUserScript):
(WTR::LayoutTestController::addUserStyleSheet):
(WTR::LayoutTestController::setXSSAuditorEnabled):
* WebKitTestRunner/TestController.cpp:
* WebKitTestRunner/win/PlatformWebViewWin.cpp:
(WTR::PlatformWebView::PlatformWebView):
2010-12-01 Steve Falkenburg <sfalken@apple.com>
Reviewed by Adam Roben.
vcproj changes can't be applied cleanly by the Windows EWS bot
https://bugs.webkit.org/show_bug.cgi?id=50328
* CLWrapper/CLWrapper.sln: Modified property svn:eol-style.
* CLWrapper/CLWrapper.vcproj: Modified property svn:eol-style.
* DumpRenderTree/DumpRenderTree.sln: Modified property svn:eol-style.
* DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePluginCommon.vsprops: Added property svn:eol-style.
* DumpRenderTree/win/DumpRenderTree.vcproj: Modified property svn:eol-style.
* DumpRenderTree/win/DumpRenderTreeApple.vsprops: Added property svn:eol-style.
* DumpRenderTree/win/DumpRenderTreeCFLite.vsprops: Added property svn:eol-style.
* DumpRenderTree/win/DumpRenderTreeCairo.vsprops: Added property svn:eol-style.
* DumpRenderTree/win/DumpRenderTreeCommon.vsprops: Added property svn:eol-style.
* DumpRenderTree/win/ImageDiff.vcproj: Modified property svn:eol-style.
* DumpRenderTree/win/ImageDiffCommon.vsprops: Added property svn:eol-style.
* FindSafari/FindSafari.vcproj: Modified property svn:eol-style.
* FindSafari/FindSafariCommon.vsprops: Added property svn:eol-style.
* MIDLWrapper/MIDLWrapper.sln: Modified property svn:eol-style.
* MIDLWrapper/MIDLWrapper.vcproj: Modified property svn:eol-style.
* MiniBrowser/Configurations/MiniBrowserCFLite.vsprops: Added property svn:eol-style.
* MiniBrowser/Configurations/MiniBrowserCommon.vsprops: Added property svn:eol-style.
* MiniBrowser/Configurations/MiniBrowserCoreFoundation.vsprops: Added property svn:eol-style.
* MiniBrowser/MiniBrowser.vcproj: Added property svn:eol-style.
* TestWebKitAPI/Configurations/TestWebKitAPICFLite.vsprops: Added property svn:eol-style.
* TestWebKitAPI/Configurations/TestWebKitAPICommon.vsprops: Added property svn:eol-style.
* TestWebKitAPI/Configurations/TestWebKitAPICoreFoundation.vsprops: Added property svn:eol-style.
* TestWebKitAPI/win/TestWebKitAPI.sln: Added property svn:eol-style.
* TestWebKitAPI/win/TestWebKitAPI.vcproj: Added property svn:eol-style.
* TestWebKitAPI/win/TestWebKitAPIGenerated.vcproj: Added property svn:eol-style.
* TestWebKitAPI/win/TestWebKitAPIGeneratedCommon.vsprops: Added property svn:eol-style.
* WebKitAPITest/WebKitAPITest.vcproj: Added property svn:eol-style.
* WebKitAPITest/WebKitAPITestCommon.vsprops: Added property svn:eol-style.
* WebKitLauncherWin/WebKitLauncherWin.vcproj: Modified property svn:eol-style.
* WebKitLauncherWin/WebKitLauncherWinCommon.vsprops: Added property svn:eol-style.
* WebKitTestRunner/Configurations/InjectedBundleCFLite.vsprops: Added property svn:eol-style.
* WebKitTestRunner/Configurations/InjectedBundleCommon.vsprops: Added property svn:eol-style.
* WebKitTestRunner/Configurations/InjectedBundleCoreFoundation.vsprops: Added property svn:eol-style.
* WebKitTestRunner/Configurations/WebKitTestRunnerCFLite.vsprops: Added property svn:eol-style.
* WebKitTestRunner/Configurations/WebKitTestRunnerCommon.vsprops: Added property svn:eol-style.
* WebKitTestRunner/Configurations/WebKitTestRunnerCoreFoundation.vsprops: Added property svn:eol-style.
* WebKitTestRunner/WebKitTestRunner.sln: Added property svn:eol-style.
* WebKitTestRunner/win/InjectedBundle.vcproj: Added property svn:eol-style.
* WebKitTestRunner/win/InjectedBundleGenerated.vcproj: Added property svn:eol-style.
* WebKitTestRunner/win/InjectedBundleGeneratedCommon.vsprops: Added property svn:eol-style.
* WebKitTestRunner/win/WebKitTestRunner.vcproj: Added property svn:eol-style.
* WinLauncher/WinLauncher.vcproj: Modified property svn:eol-style.
* WinLauncher/WinLauncherCommon.vsprops: Added property svn:eol-style.
* record-memory-win/record-memory-win-common.vsprops: Added property svn:eol-style.
* record-memory-win/record-memory-win.vcproj: Modified property svn:eol-style.
2010-12-01 Anders Carlsson <andersca@apple.com>
Reviewed by Darin Adler.
Return a correct user agent if NPN_UserAgent is called with a null NPP from NPP_New.
https://bugs.webkit.org/show_bug.cgi?id=50336
* DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
Add new test.
* DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp:
(PluginTest::NPP_New):
Add default implementation/
* DumpRenderTree/TestNetscapePlugIn/PluginTest.h:
* DumpRenderTree/TestNetscapePlugIn/Tests/GetUserAgentWithNullNPPFromNPPNew.cpp: Added.
(GetUserAgentWithNullNPPFromNPPNew::GetUserAgentWithNullNPPFromNPPNew):
(GetUserAgentWithNullNPPFromNPPNew::NPP_New):
Get the user agent.
* DumpRenderTree/TestNetscapePlugIn/main.cpp:
(NPP_New):
Call PluginTest::NPP_New.
* DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
* DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro:
Add new files.
* DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
(webkit_test_plugin_new_instance):
Call PluginTest::NPP_New.
2010-12-01 Kevin Ollivier <kevino@theolliviers.com>
[wx] Build fixes after recent trunk changes.
* DumpRenderTree/wx/LayoutTestControllerWx.cpp:
(LayoutTestController::findString):
* wx/build/settings.py:
2010-12-01 Dan Bernstein <mitz@apple.com>
Reviewed by Darin Adler.
WebKitTestRunner needs layoutTestController.findString
https://bugs.webkit.org/show_bug.cgi?id=50238
* WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl: Added findString().
* WebKitTestRunner/InjectedBundle/LayoutTestController.cpp:
(WTR::LayoutTestController::findString): Updated signature for autogenerated bindings.
* WebKitTestRunner/InjectedBundle/LayoutTestController.h: Ditto.
2010-12-01 Steve Falkenburg <sfalken@apple.com>
Try using svn:eol-style native on a vcproj file.
* DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj: Modified property svn:eol-style.
2010-12-01 Patrick Gansterer <paroga@webkit.org>
Reviewed by Adam Roben.
[WINCE] Add WinCELauncher
https://bugs.webkit.org/show_bug.cgi?id=50217
* WinCELauncher/main.cpp: Added.
2010-11-30 Benjamin Poulain <benjamin.poulain@nokia.com>
Reviewed by Kenneth Rohde Christiansen.
[Qt] Implement layoutTestController.findString
https://bugs.webkit.org/show_bug.cgi?id=50236
Add the missing function to the LayoutTestController.
* DumpRenderTree/qt/LayoutTestControllerQt.cpp:
(LayoutTestController::findString):
* DumpRenderTree/qt/LayoutTestControllerQt.h:
2010-11-30 Tony Chang <tony@chromium.org>
Reviewed by Adam Barth.
[chromium] fix get-int-identifier-special-values.html using TestNetscapePlugIn
https://bugs.webkit.org/show_bug.cgi?id=49036
* DumpRenderTree/chromium/LayoutTestController.cpp: Use IsNumber instead of IsInt32
for CppVariants.
(LayoutTestController::cppVariantToBool):
(LayoutTestController::cppVariantToInt32):
(LayoutTestController::setDatabaseQuota):
(LayoutTestController::evaluateInWebInspector):
(LayoutTestController::setMockGeolocationError):
2010-11-30 Eric Seidel <eric@webkit.org>
Reviewed by Tony Chang.
update-webkit should call git fetch before git svn rebase
https://bugs.webkit.org/show_bug.cgi?id=50273
After discussion on webkit-dev, we've decided to move the "default"
git setup to pulling updates from git.webkit.org in preference
to rebuilding the local svn index using git svn fetch every time.
This change should have no effect on people using the "old" git setup
and should dramatically increase the speed of updates for those using
the "new" git setup along with update-webkit. I'm about to move the
EWS and other queues over to this setup once this lands.
* Scripts/update-webkit:
2010-11-30 Mario Sanchez Prada <msanchez@igalia.com>
Reviewed by Xan Lopez.
GTK: AX: implement press in DRT
https://bugs.webkit.org/show_bug.cgi?id=36146
Implement AccessibilityUIElement::press() for GTK.
* DumpRenderTree/gtk/AccessibilityUIElementGtk.cpp:
(AccessibilityUIElement::press): Implemented.
2010-11-29 Mihai Parparita <mihaip@chromium.org>
Reviewed by Tony Chang.
config.webkit_base_dir returns a path with a trailing slash
https://bugs.webkit.org/show_bug.cgi?id=50197
config.webkit_base_dir() should return a path without a trailing slash,
like all the other methods that return paths.
Undoes MockFileSytem changes by r72640, since they're no longer
necessary for GetBaselinesTest to pass.
* Scripts/webkitpy/common/system/filesystem_mock.py:
* Scripts/webkitpy/layout_tests/port/base.py:
* Scripts/webkitpy/layout_tests/port/config.py:
* Scripts/webkitpy/layout_tests/port/config_unittest.py:
2010-11-30 Mihai Parparita <mihaip@chromium.org>
Reviewed by Tony Chang.
Rebaseline server: baseline display tweaks
https://bugs.webkit.org/show_bug.cgi?id=50207
Determine which baselines were used when running the tests and highlight
them in the UI. Sort platform names alphabetically.
Also makes the server URL be launched in the user's browser
automatically.
* Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css:
* Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
* Scripts/webkitpy/tool/commands/data/rebaselineserver/util.js:
* Scripts/webkitpy/tool/commands/rebaselineserver.py:
* Scripts/webkitpy/tool/commands/rebaselineserver_unittest.py:
2010-11-30 Dan Bernstein <mitz@apple.com>
Reviewed by Anders Carlsson.
<rdar://problem/8710645> WebKitTestRunner needs layoutTestController.findString
https://bugs.webkit.org/show_bug.cgi?id=50238
* WebKitTestRunner/InjectedBundle/LayoutTestController.cpp:
(WTR::LayoutTestController::findString): Added.
* WebKitTestRunner/InjectedBundle/LayoutTestController.h:
2010-11-30 Steve Falkenburg <sfalken@apple.com>
Reviewed by Adam Roben.
All projects on Windows should use cmd files for build events
https://bugs.webkit.org/show_bug.cgi?id=50213
* DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePluginCommon.vsprops:
* DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePluginPostBuild.cmd: Added.
* DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePluginPreBuild.cmd: Added.
* DumpRenderTree/win/DumpRenderTreeCommon.vsprops:
* DumpRenderTree/win/DumpRenderTreePostBuild.cmd: Added.
* DumpRenderTree/win/DumpRenderTreePreBuild.cmd: Added.
* DumpRenderTree/win/ImageDiffCommon.vsprops:
* DumpRenderTree/win/ImageDiffPostBuild.cmd: Added.
* DumpRenderTree/win/ImageDiffPreBuild.cmd: Added.
* MiniBrowser/Configurations/MiniBrowserCFLite.vsprops:
* MiniBrowser/Configurations/MiniBrowserCommon.vsprops:
* MiniBrowser/Configurations/MiniBrowserCoreFoundation.vsprops:
* MiniBrowser/MiniBrowserPostBuild.cmd: Added.
* MiniBrowser/MiniBrowserPreBuild.cmd: Added.
* TestWebKitAPI/Configurations/TestWebKitAPICommon.vsprops:
* TestWebKitAPI/win/TestWebKitAPIPostBuild.cmd: Added.
* TestWebKitAPI/win/TestWebKitAPIPreBuild.cmd: Added.
* WebKitAPITest/WebKitAPITestCommon.vsprops:
* WebKitAPITest/WebKitAPITestPostBuild.cmd: Added.
* WebKitAPITest/WebKitAPITestPreBuild.cmd: Added.
* WebKitTestRunner/Configurations/InjectedBundleCommon.vsprops:
* WebKitTestRunner/Configurations/WebKitTestRunnerCommon.vsprops:
* WebKitTestRunner/win/InjectedBundlePostBuild.cmd: Added.
* WebKitTestRunner/win/InjectedBundlePreBuild.cmd: Added.
* WebKitTestRunner/win/WebKitTestRunnerPostBuild.cmd: Added.
* WebKitTestRunner/win/WebKitTestRunnerPreBuild.cmd: Added.
* WinLauncher/WinLauncherCommon.vsprops:
* WinLauncher/WinLauncherPostBuild.cmd: Added.
* WinLauncher/WinLauncherPreBuild.cmd: Added.
2010-11-29 Dan Bernstein <mitz@apple.com>
GTK DumpRenderTree build fix after r72887.
* DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
(LayoutTestController::findString):
2010-11-29 Dan Bernstein <mitz@apple.com>
Windows DumpRenderTree build fix after r72887.
* DumpRenderTree/win/LayoutTestControllerWin.cpp:
(LayoutTestController::findString): Added empty implementation.
2010-11-29 Dan Bernstein <mitz@apple.com>
Reviewed by Darin Adler.
DumpRenderTree changes for testing the text search API.
https://bugs.webkit.org/show_bug.cgi?id=50038
* DumpRenderTree/LayoutTestController.cpp:
(findStringCallback):
(LayoutTestController::staticFunctions):
* DumpRenderTree/LayoutTestController.h:
* DumpRenderTree/mac/LayoutTestControllerMac.mm:
(LayoutTestController::findString):
* MiniBrowser/mac/BrowserWindowController.m:
(-[BrowserWindowController find:]):
2010-11-29 Johnny Ding <jnd@chromium.org>
Unreviewed: Add myself to the list of Committers.
* Scripts/webkitpy/common/config/committers.py:
2010-11-29 Mark Rowe <mrowe@apple.com>
Reviewed by Adele Peterson.
<rdar://problem/8694997> DumpRenderTree fails to build.
* DumpRenderTree/mac/PerlSupport/Makefile: Skip generating wrappers.
2010-11-29 Mark Rowe <mrowe@apple.com>
Reviewed by Adele Peterson.
<rdar://problem/8694997> DumpRenderTree fails to build.
* DumpRenderTree/mac/PerlSupport/DumpRenderTreeSupportPregenerated.pm: Renamed from WebKitTools/DumpRenderTree/mac/PerlSupport/DumpRenderTreeSupportTiger.pm.
* DumpRenderTree/mac/PerlSupport/DumpRenderTreeSupport_wrapPregenerated.c: Renamed from WebKitTools/DumpRenderTree/mac/PerlSupport/DumpRenderTreeSupport_wrapTiger.c.
* DumpRenderTree/mac/PerlSupport/Makefile:
2010-11-29 Mario Sanchez Prada <msanchez@igalia.com>
Reviewed by Martin Robinson.
[GTK] Null-check needed in DRT's AccessibilityUIElement::allAtributes()
https://bugs.webkit.org/show_bug.cgi?id=50154
Added missing checks to allAttributes().
* DumpRenderTree/gtk/AccessibilityUIElementGtk.cpp:
(AccessibilityUIElement::allAttributes): Added missing checks.
2010-11-29 Laszlo Gombos <laszlo.1.gombos@nokia.com>
Reviewed by Kenneth Rohde Christiansen.
[Qt] Allow pre-generation for package builds for WebKit2
https://bugs.webkit.org/show_bug.cgi?id=50139
Use a build variable for the generated directory path and set the
value based on CONFIG just like for WebKit1.
For non-package builds use a relative base path for the
genrated directory just like for WebKit1.
* WebKitTestRunner/InjectedBundle/qt/InjectedBundle.pro:
* WebKitTestRunner/qt/WebKitTestRunner.pro:
2010-11-28 Laszlo Gombos <laszlo.1.gombos@nokia.com>
Reviewed by Kenneth Rohde Christiansen.
[Qt] Move some build logic from Qt to platform independent code
https://bugs.webkit.org/show_bug.cgi?id=50134
Create target directories inside generate-forwarding-headers.pl.
* Scripts/generate-forwarding-headers.pl:
2010-11-24 Nicolas Dufresne <nicolas.dufresne@collabora.co.uk>
Reviewed by Martin Robinson.
[GTK] Windowless plugins gets broken key input
https://bugs.webkit.org/show_bug.cgi?id=49927
Convert the XEvent::keycode to ASCII before printing it. Not
doing this was causing false positive in
plugins/keyboard-events.html.
* DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
(keycode_to_char):
(webkit_test_plugin_handle_event):
2010-11-24 Jan Erik Hanssen <jhanssen@sencha.com>
Reviewed by Antonio Gomes.
[Qt] Html autofocus not working with QGraphicsWebView
https://bugs.webkit.org/show_bug.cgi?id=43169
Tell the scene to set the webview item as the currently
focused item. This makes the input field with autofocus get
focus as it should.
* QtTestBrowser/webview.cpp:
(WebViewGraphicsBased::WebViewGraphicsBased):
2010-11-24 Dirk Pranke <dpranke@chromium.org>
Reviewed by Tony Chang.
nrwt - clean up create_driver interface and print out the
command line used for DumpRenderTree/TestShell.
This patch adds a driver_cmd_line() call to the Port class in
order to retrive the command line to print it out (as part of
--print config).
In addition, this patch cleans up the Port.create_driver()
signature and the Driver constructor interface to take a worker
number and no longer require the image_path argument (Chromium's
drivers now synthesize the image_path from the worker number
internally).
https://bugs.webkit.org/show_bug.cgi?id=49934
* Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
* Scripts/webkitpy/layout_tests/layout_package/printing.py:
* Scripts/webkitpy/layout_tests/layout_package/printing_unittest.py:
* Scripts/webkitpy/layout_tests/port/base.py:
* Scripts/webkitpy/layout_tests/port/base_unittest.py:
* Scripts/webkitpy/layout_tests/port/chromium.py:
* Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
* Scripts/webkitpy/layout_tests/port/dryrun.py:
* Scripts/webkitpy/layout_tests/port/port_testcase.py:
* Scripts/webkitpy/layout_tests/port/test.py:
* Scripts/webkitpy/layout_tests/port/webkit.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2010-11-24 Ojan Vafai <ojan@chromium.org>
Reviewed by Tony Chang.
[chromium] improve memory usage for test results server and store fewer runs
https://bugs.webkit.org/show_bug.cgi?id=50035
Store fewer runs and don't pass full files around when we only need bools.
Greatly improves error rates on test-results.appspot.com.
* Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
* TestResultServer/handlers/testfilehandler.py:
* TestResultServer/model/dashboardfile.py:
* TestResultServer/model/jsonresults.py:
2010-11-24 Dirk Pranke <dpranke@chromium.org>
Reviewed by Tony Chang.
This patch implements the first part of the manager side of the
Broker objects - it handles creating threads, waiting for them
to complete, and running a single-threaded loop as well.
https://bugs.webkit.org/show_bug.cgi?id=49779
* Scripts/webkitpy/layout_tests/layout_package/message_broker.py:
* Scripts/webkitpy/layout_tests/layout_package/message_broker_unittest.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2010-11-24 Mihai Parparita <mihaip@chromium.org>
Reviewed by David Levin.
DRT should not generate pixel output for text/plain tests
https://bugs.webkit.org/show_bug.cgi?id=50002
Modify Mac, Windows and GTK DRT implementations to disable pixel output
when encountering a text/plain response (the Chromium one already does
this, and none of the other ports special-case text/plain output in
the first place).
* DumpRenderTree/gtk/DumpRenderTree.cpp:
(dump):
* DumpRenderTree/mac/DumpRenderTree.mm:
(dump):
* DumpRenderTree/win/DumpRenderTree.cpp:
(dump):
2010-11-24 Andras Becsi <abecsi@inf.u-szeged.hu>
Unreviewed.
Add my other email address to committers.py.
* Scripts/webkitpy/common/config/committers.py:
2010-11-24 Andras Becsi <abecsi@inf.u-szeged.hu>
Reviewed by Csaba Osztrogonác.
Make build-webkit --minimal build with ENABLE_INSPECTOR=0
https://bugs.webkit.org/show_bug.cgi?id=49975
* Scripts/build-webkit: Add ENABLE_INSPECTOR define to features,
to enable the script to define it 0 in case of a minimal build.
2010-11-24 Martin Robinson <mrobinson@igalia.com>
Reviewed by Xan Lopez.
[GTK] The WebKitWebView should expose a set of signals encapsulating the behavior of the EditorClient
https://bugs.webkit.org/show_bug.cgi?id=49143
* DumpRenderTree/gtk/DumpRenderTree.cpp:
(createWebView): Call the new connectEditingCallbacks function.
* DumpRenderTree/gtk/EditingCallbacks.cpp: Added.
(dumpNodePath):
(dumpRange):
(insertActionString):
(selectionAffinityString):
(shouldBeginEditing):
(shouldEndEditing):
(shouldInsertNode):
(shouldInsertText):
(shouldDeleteRange):
(shouldShowDeleteInterfaceForElement):
(shouldChangeSelectedRange):
(shouldApplyStyle):
(editingBegan):
(userChangedContents):
(editingEnded):
(selectionChanged):
(connectEditingCallbacks):
* DumpRenderTree/gtk/EditingCallbacks.h: Added.
* GNUmakefile.am: Added EditingCallbacks.{h,cpp} to the source list.
2010-11-24 Yi Shen <yi.4.shen@nokia.com>
Reviewed by Andreas Kling.
[Qt] Make QtTestBrowser remember the selected user agent
https://bugs.webkit.org/show_bug.cgi?id=50021
Set user agent for a new page by using the old page's
* QtTestBrowser/mainwindow.cpp:
(MainWindow::setPage):
2010-11-24 Patrick Gansterer <paroga@webkit.org>
Reviewed by Csaba Osztrogonác.
Remove Bakefile build system files
https://bugs.webkit.org/show_bug.cgi?id=49983
r53757 only removed the content, but not the files.
This patch removes that empty files.
* DumpRenderTree/wx/DumpRenderTree.bkl: Removed.
* wx/browser/browser.bkl: Removed.
* wx/build-wxwebkit: Removed.
2010-11-23 Dirk Pranke <dpranke@chromium.org>
Reviewed by Tony Chang.
This patch cleans up the logic used to shard tests into groups a
bit and adds the --worker-model flag to NRWT. The flag is only
used at the moment to control whether to run single-threaded or
not, but eventually will also allow toggling between threads and
processes.
Also add a minor cleanup with _test_is_slow(), which just
eliminates some repetition and gives slightly better encapsulation.
https://bugs.webkit.org/show_bug.cgi?id=49773
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2010-11-23 Mihai Parparita <mihaip@chromium.org>
Reviewed by Tony Chang.
Rebaseline server: list current baselines and platforms
https://bugs.webkit.org/show_bug.cgi?id=49991
List existing baselines (with Trac links) for tests.
Add dropdowns for choosing with platform baselines to target (and what
to do with existing ones).
Also fix MockFileSystem.join to behave more like os.path.join (unit
test was ending up with two consecutive slashes in a layout test
path).
* Scripts/webkitpy/common/system/filesystem_mock.py:
* Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html:
* Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css:
* Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
* Scripts/webkitpy/tool/commands/data/rebaselineserver/queue.js:
* Scripts/webkitpy/tool/commands/data/rebaselineserver/util.js:
* Scripts/webkitpy/tool/commands/rebaselineserver.py:
* Scripts/webkitpy/tool/commands/rebaselineserver_unittest.py: Added.
2010-11-23 Ojan Vafai <ojan@chromium.org>
Reviewed by Tony Chang.
remove timeout argument to urlopen
https://bugs.webkit.org/show_bug.cgi?id=49995
Apparently the version of python we have on the bots doesn't accept a timeout argument.
* Scripts/webkitpy/layout_tests/layout_package/test_results_uploader.py:
2010-11-23 Sheriff Bot <webkit.review.bot@gmail.com>
Unreviewed, rolling out r72628.
http://trac.webkit.org/changeset/72628
https://bugs.webkit.org/show_bug.cgi?id=49994
This patch is causing layout-test failtures on GTK Linux
64-bit Debug (Requested by ctguil on #webkit).
* DumpRenderTree/chromium/AccessibilityUIElement.cpp:
(AccessibilityUIElement::AccessibilityUIElement):
(AccessibilityUIElement::parentElementCallback):
* DumpRenderTree/chromium/AccessibilityUIElement.h:
* DumpRenderTree/chromium/CppBoundClass.cpp:
* DumpRenderTree/chromium/CppBoundClass.h:
2010-11-22 Ojan Vafai <ojan@chromium.org>
Reviewed by Tony Chang.
speculative fix for upload errors: stop using mechanize to upload to test-results.appspot.com
https://bugs.webkit.org/show_bug.cgi?id=49944
* Scripts/webkitpy/layout_tests/layout_package/test_results_uploader.py:
2010-11-23 Chris Guillory <chris.guillory@google.com>
Reviewed by Dimitri Glazkov.
Include the FrameView widget of a RenderWidget in the accessibility tree.
https://bugs.webkit.org/show_bug.cgi?id=49106
* DumpRenderTree/chromium/AccessibilityUIElement.cpp:
(AccessibilityUIElement::AccessibilityUIElement):
(AccessibilityUIElement::parentElementCallback):
(AccessibilityUIElement::isEqualCallback):
* DumpRenderTree/chromium/AccessibilityUIElement.h:
* DumpRenderTree/chromium/CppBoundClass.cpp:
(CppBoundClass::getFromCppVariant):
* DumpRenderTree/chromium/CppBoundClass.h:
2010-11-23 Alex Grilo <abgrilo@profusion.mobi>
Reviewed by Kenneth Rohde Christiansen.
[EFL] Add tiled backing store to EWebLauncher
https://bugs.webkit.org/show_bug.cgi?id=45777
Allow EWebLauncher to choose between single and tiled backing store
* EWebLauncher/main.c:
(on_key_down):
(browserCreate):
(main):
2010-11-23 Simon Hausmann <simon.hausmann@nokia.com>
Reviewed by Tor Arne Vestbø.
[Qt] Review the setUserPermission & friends API
https://bugs.webkit.org/show_bug.cgi?id=46810
Renamed requestPermissionFromUser to featurePermissionRequested
and cancelRequestsForPermission to featurePermissionRequestCanceled.
* DumpRenderTree/qt/DumpRenderTreeQt.cpp:
(WebCore::WebPage::WebPage):
* QtTestBrowser/webpage.cpp:
(WebPage::WebPage):
(WebPage::featurePermissionRequestCanceled):
* QtTestBrowser/webpage.h:
2010-11-23 Simon Hausmann <simon.hausmann@nokia.com>
Reviewed by Tor Arne Vestbø.
[Qt] Review the setUserPermission & friends API
https://bugs.webkit.org/show_bug.cgi?id=46810
Rename setUserPermission to setFeaturePermission
* DumpRenderTree/qt/DumpRenderTreeQt.cpp:
(WebCore::WebPage::requestPermission):
(WebCore::WebPage::permissionSet):
* DumpRenderTree/qt/LayoutTestControllerQt.cpp:
(LayoutTestController::grantDesktopNotificationPermission):
* QtTestBrowser/webpage.cpp:
(WebPage::requestPermission):
2010-11-23 Simon Hausmann <simon.hausmann@nokia.com>
Reviewed by Tor Arne Vestbø.
[Qt] Review the setUserPermission & friends API
https://bugs.webkit.org/show_bug.cgi?id=46810
Rename PermissionDomain to Feature, NotificationsPermissionDomain to Notifications
and GeolocationPermissionDomain to Geolocation.
* DumpRenderTree/qt/DumpRenderTreeQt.cpp:
(WebCore::WebPage::WebPage):
(WebCore::WebPage::requestPermission):
(WebCore::WebPage::cancelPermission):
(WebCore::WebPage::permissionSet):
(WebCore::DumpRenderTree::geolocationPermissionSet):
* DumpRenderTree/qt/DumpRenderTreeQt.h:
* DumpRenderTree/qt/LayoutTestControllerQt.cpp:
(LayoutTestController::grantDesktopNotificationPermission):
* QtTestBrowser/webpage.cpp:
(WebPage::WebPage):
(WebPage::requestPermission):
(WebPage::cancelRequestsForPermission):
* QtTestBrowser/webpage.h:
2010-11-23 Simon Hausmann <simon.hausmann@nokia.com>
Reviewed by Tor Arne Vestbø.
[Qt] Review the setUserPermission & friends API
https://bugs.webkit.org/show_bug.cgi?id=46810
Add a ByUser suffix to PermissionGranted/Denied. In the future
we can add PermissionGrantedByDefault.
* DumpRenderTree/qt/DumpRenderTreeQt.cpp:
(WebCore::WebPage::requestPermission):
(WebCore::WebPage::permissionSet):
* DumpRenderTree/qt/LayoutTestControllerQt.cpp:
(LayoutTestController::grantDesktopNotificationPermission):
* QtTestBrowser/webpage.cpp:
(WebPage::requestPermission):
2010-11-18 Satish Sampath <satish@chromium.org>
Reviewed by Jeremy Orlow.
For speech input event, send an event object containing all the recognition results and metadata.
https://bugs.webkit.org/show_bug.cgi?id=49736
Updated DRT by renaming the mock result method on all platforms to the new signature.
* DumpRenderTree/LayoutTestController.cpp:
(addMockSpeechInputResultCallback):
(LayoutTestController::staticFunctions):
* DumpRenderTree/LayoutTestController.h:
* DumpRenderTree/chromium/LayoutTestController.cpp:
(LayoutTestController::LayoutTestController):
(LayoutTestController::addMockSpeechInputResult):
* DumpRenderTree/chromium/LayoutTestController.h:
* DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
(LayoutTestController::addMockSpeechInputResult):
* DumpRenderTree/mac/LayoutTestControllerMac.mm:
(LayoutTestController::addMockSpeechInputResult):
* DumpRenderTree/qt/LayoutTestControllerQt.cpp:
(LayoutTestController::addMockSpeechInputResult):
* DumpRenderTree/qt/LayoutTestControllerQt.h:
* DumpRenderTree/win/LayoutTestControllerWin.cpp:
(LayoutTestController::addMockSpeechInputResult):
* DumpRenderTree/wx/LayoutTestControllerWx.cpp:
(LayoutTestController::addMockSpeechInputResult):
2010-11-22 Dirk Pranke <dpranke@chromium.org>
Reviewed by Tony Chang.
Fix some spacing issues with log messages with
new-run-webkit-tests --verbose.
https://bugs.webkit.org/show_bug.cgi?id=49936
* Scripts/webkitpy/layout_tests/layout_package/printing.py:
2010-11-22 Ojan Vafai <ojan@chromium.org>
Reviewed by Adam Barth.
update scm to work with different remote branches
https://bugs.webkit.org/show_bug.cgi?id=49949
This works if there are multiple svn tracking branches
and/or if the tracking branch is refs/remots/origin/master
instead of refs/remotes/trunk which is the direction we're trying
to head since that makes fetches faster.
* Scripts/webkitpy/common/checkout/scm.py:
* Scripts/webkitpy/common/checkout/scm_unittest.py:
2010-11-22 Hayato Ito <hayato@chromium.org>
Reviewed by Shinichiro Hamaji.
Ignore reference files which will be used by reftests when collecting
test cases.
https://bugs.webkit.org/show_bug.cgi?id=49835
* Scripts/webkitpy/layout_tests/port/test_files.py:
* Scripts/webkitpy/layout_tests/port/test_files_unittest.py:
2010-11-22 Adam Roben <aroben@apple.com>
Use paths relative to $WebKitVSPropsRedirectionDir to access shared .vsprops files
Apple's Windows build allows placing header files and import libraries for WebKit's
dependencies (CoreGraphics, CFNetwork, SQLite, etc.) outside the source tree via the
$WebKitLibrariesDir environment variable. This is both required for production builds and
convenient for Apple-internal developer builds. Apple's production builds also require that
WebKit's shared .vsprops files be accessed relative to $WebKitLibrariesDir. In production
builds, the files are copied into that directory tree by the
WebKitLibraries/win/tools/WinTools.make file. In Apple-internal developer builds, the
copying is done by
JavaScriptCore/JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreGenerated.make.
This .vsprops copying is problematic in one very important case: when a developer updates
their source tree and then tries to build. Visual Studio only reads .vsprops files when a
project is first loaded. So, when Visual Studio is first opened after the .vsprops files are
updated, it reads in the old files that were already residing in $WebKitLibrariesDir. When a
build is started, JavaScriptCoreGenerated.make copies the new .vsprops files into
$WebKitLibrariesDir, but Visual Studio will not pick up the changes. The rest of the build
will proceed with out-of-date .vsprops files, which will likely result in a build failure.
To fix this, we now use normal relative paths to access the .vsprops files in the source
tree rather than in $WebKitLibrariesDir, but prefix those paths with a new environment
variable, $WebKitVSPropsRedirectionDir. In developer builds, this environment variable is
unset, so the normal relative paths are used to read the .vsprops files out of the source
tree directly. In production builds, this environment variable is set to a fake directory
that will cause the .vsprops files in $WebKitLibrariesDir to be found when the relative path
is resolved.
For example, JavaScriptCore.vcproj uses this path for FeatureDefines.vsprops:
$(WebKitVSPropsRedirectionDir)..\..\..\WebKitLibraries\win\tools\vsprops\FeatureDefines.vsprops
In developer builds, where $WebKitVSPropsRedirectionDir is unset, this will point to the
files in WebKitLibraries\win\tools\vsprops in the source tree. In production builds,
JavaScriptCore.make sets $WebKitVSPropsRedirectionDir to
"$(SRCROOT)\AppleInternal\tools\vsprops\OpenSource\1\2\3\", so the full path for
FeatureDefines.vsprops becomes:
$(SRCROOT)\AppleInternal\tools\vsprops\OpenSource\1\2\3\..\..\..\WebKitLibraries\win\tools\vsprops\FeatureDefines.vsprops
which resolves to:
$(SRCROOT)\AppleInternal\tools\vsprops\OpenSource\WebKitLibraries\win\tools\vsprops\FeatureDefines.vsprops
(We rely on the fact that Windows doesn't care whether the directories "1", "2", and "3"
actually exist since they are matched by an equal number of ".." path components.)
Note that Visual Studio still won't pick up changes made to .vsprops files while Visual
Studio is open, but that problem hasn't seemed to cause developers many headaches so far.
Fixes <http://webkit.org/b/49181> Windows build fails mysteriously when .vsprops files are
updated
Reviewed by Dave Hyatt.
* DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
* DumpRenderTree/win/DumpRenderTree.vcproj:
* DumpRenderTree/win/ImageDiff.vcproj:
* FindSafari/FindSafari.vcproj:
* MiniBrowser/MiniBrowser.vcproj:
* TestWebKitAPI/win/TestWebKitAPI.vcproj:
* TestWebKitAPI/win/TestWebKitAPIGenerated.vcproj:
* WebKitAPITest/WebKitAPITest.vcproj:
* WebKitLauncherWin/WebKitLauncherWin.vcproj:
* WebKitTestRunner/win/InjectedBundle.vcproj:
* WebKitTestRunner/win/WebKitTestRunner.vcproj:
* WinLauncher/WinLauncher.vcproj:
* record-memory-win/record-memory-win.vcproj:
Changed to use paths relative to $WebKitVSPropsRedirectionDir to access shared .vsprops
files.
2010-11-22 Adam Roben <aroben@apple.com>
Make it possible to run tests on Windows without Visual Studio or VC++
Express installed
webkitdir.pm::setupCygwinEnv dies when Visual Studio and VC++ Express
are not installed. But this function doesn't need to be called when we
already have a build available and are just trying to run the tests.
Fixes <http://webkit.org/b/49932> New Windows 7 bot can't run tests
because Visual Studio/VC++ Express aren't installed
Reviewed by Dave Hyatt.
* Scripts/webkitdirs.pm:
(determineConfigurationForVisualStudio): Don't call setupCygwinEnv, as
it is not needed by this function. Also added a FIXME.
(usingVisualStudioExpress): Call setupCygwinEnv directly rather than
relying on determineConfigurationForVisualStudio doing it.
2010-11-19 Hayato Ito <hayato@chromium.org>
Reviewed by Shinichiro Hamaji.
Simplify Driver.run_test() so that it takes only one parameter, TestInput.
https://bugs.webkit.org/show_bug.cgi?id=49812
* Scripts/webkitpy/layout_tests/driver_test.py:
* Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
* Scripts/webkitpy/layout_tests/port/base.py:
* Scripts/webkitpy/layout_tests/port/base_unittest.py:
* Scripts/webkitpy/layout_tests/port/chromium.py:
* Scripts/webkitpy/layout_tests/port/dryrun.py:
* Scripts/webkitpy/layout_tests/port/test.py:
* Scripts/webkitpy/layout_tests/port/webkit.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2010-11-22 Hayato Ito <hayato@chromium.org>
Reviewed by Tony Chang.
[NRWT] Retry a few times in reading a png image to avoid a race condition.
https://bugs.webkit.org/show_bug.cgi?id=49924
* Scripts/webkitpy/layout_tests/port/chromium.py:
2010-11-22 João Paulo Rechi Vita <jprvita@profusion.mobi>
Reviewed by Kenneth Rohde Christiansen.
Fix EWebLauncher zoom discretization
https://bugs.webkit.org/show_bug.cgi?id=49810
* EWebLauncher/main.c:
(nearest_zoom_level_get):
(zoom_level_set):
(on_load_finished):
(on_key_down):
2010-11-21 Sam Weinig <sam@webkit.org>
Rubber-stamped by Dan Bernstein.
Fix ~300 WebKit2 tests by enabling developer extras preference.
* WebKitTestRunner/TestController.cpp:
(WTR::TestController::resetStateToConsistentValues):
2010-11-21 Sam Weinig <sam@webkit.org>
Reviewed by Dan Bernstein.
Textareas should be resizable by default
https://bugs.webkit.org/show_bug.cgi?id=49892
* TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
* TestWebKitAPI/Tests/WebKit2/WKPreferences.cpp: Added.
* TestWebKitAPI/win/TestWebKitAPI.vcproj:
Add preferences test.
2010-11-19 Mihai Parparita <mihaip@chromium.org>
Reviewed by Tony Chang.
Rebaseline server: checksum-only differences and prefetching
https://bugs.webkit.org/show_bug.cgi?id=49841
Add support for image test results where the only difference is in the
checksum.
Add prefetching of the image results from the next test.
* Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html:
* Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css:
* Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
2010-11-19 Dirk Pranke <dpranke@chromium.org>
Reviewed by Tony Chang.
nrwt multiprocessing - add 'worker number' concept, move stuff to worker thread
Add the 'worker number' and 'worker name' concepts to the
TestShellThread objects, and move test_types and test_args from
the TestRunner to the TestShellThread.
https://bugs.webkit.org/show_bug.cgi?id=49768
* Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2010-11-19 Dirk Pranke <dpranke@chromium.org>
Reviewed by Ojan Vafai.
Do some minor cleanup and bug fixing.
https://bugs.webkit.org/show_bug.cgi?id=49777
* Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2010-11-19 Dirk Pranke <dpranke@chromium.org>
Reviewed by Ojan Vafai.
nrwt - config.build_directory() is busted
Fixes a typo that was causing us to usually return the top level
directory WebKitBuild instead of WebKitBuild/{Debug,Release}. The
bug was hidden by test stubs that were too simplistic :(.
https://bugs.webkit.org/show_bug.cgi?id=49815
* Scripts/webkitpy/common/system/executive_mock.py:
* Scripts/webkitpy/layout_tests/port/config.py:
* Scripts/webkitpy/layout_tests/port/config_unittest.py:
2010-11-19 Steve Falkenburg <sfalken@apple.com>
Reviewed by Adam Roben.
Add Debug_Cairo_CFLite and Release_Cairo_CFLite configurations for all vcproj files
https://bugs.webkit.org/show_bug.cgi?id=49819
* DumpRenderTree/DumpRenderTree.sln:
* DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
* DumpRenderTree/win/DumpRenderTree.vcproj:
* DumpRenderTree/win/ImageDiff.vcproj:
* FindSafari/FindSafari.vcproj:
* MiniBrowser/MiniBrowser.vcproj:
* TestWebKitAPI/win/TestWebKitAPI.sln:
* TestWebKitAPI/win/TestWebKitAPI.vcproj:
* TestWebKitAPI/win/TestWebKitAPIGenerated.vcproj:
* WebKitAPITest/WebKitAPITest.vcproj:
* WebKitLauncherWin/WebKitLauncherWin.vcproj:
* WebKitTestRunner/win/InjectedBundle.vcproj:
* WebKitTestRunner/win/InjectedBundleGenerated.vcproj:
* WebKitTestRunner/win/WebKitTestRunner.vcproj:
* WinLauncher/WinLauncher.vcproj:
* record-memory-win/record-memory-win.vcproj:
2010-11-19 Adam Roben <aroben@apple.com>
Add some more Windows slaves
This is probably more than we will be using in the immediate future,
but having some extra slave names makes experimenting with different
configurations easier.
Reviewed by Jon Honeycutt.
* BuildSlaveSupport/build.webkit.org-config/config.json: Added 5 more
Windows slaves and distributed them among the three different Test
builders.
2010-11-19 Kinuko Yasuda <kinuko@chromium.org>
Reviewed by Ojan Vafai.
[Chromium] REGRESSION: Record actual test results in the JSON result summary for non-layout tests
https://bugs.webkit.org/show_bug.cgi?id=49702
Re-landing r72357 with a test fix.
* Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
* Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
* Scripts/webkitpy/layout_tests/layout_package/json_results_generator_unittest.py:
2010-11-19 Avi Drissman <avi@google.com>
Reviewed by Antonio Gomes.
Complete support for Unix editing mode
https://bugs.webkit.org/show_bug.cgi?id=49757
* DumpRenderTree/LayoutTestController.cpp:
(setEditingBehaviorCallback):
* DumpRenderTree/chromium/LayoutTestController.cpp:
(LayoutTestController::setEditingBehavior):
2010-11-19 Adam Roben <aroben@apple.com>
Fix run-javascriptcore-tests when there's a space in the path to
testapi
Reviewed by Sam Weinig.
* Scripts/run-javascriptcore-tests: Use an indirect object when calling
system() to force it to interpret spaces in the path correctly.
2010-11-19 Adam Roben <aroben@apple.com>
Update scripts for .exe name changes after r72327
As of r72327, the "_debug" suffix is only used in Debug_All builds.
Reviewed by Sam Weinig.
* Scripts/run-javascriptcore-tests:
(testapiPath):
* Scripts/sunspider-compare-results:
(pathToBuiltJSC):
* Scripts/webkitdirs.pm:
(jscPath):
(safariPath):
2010-11-19 Steve Falkenburg <sfalken@apple.com>
Reviewed by Darin Adler.
Normalize Cairo/CFLite project/solution configuration names
https://bugs.webkit.org/show_bug.cgi?id=49818
* DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
* DumpRenderTree/win/DumpRenderTree.vcproj:
* DumpRenderTree/win/ImageDiff.vcproj:
* MiniBrowser/MiniBrowser.vcproj:
* TestWebKitAPI/win/TestWebKitAPI.sln:
* TestWebKitAPI/win/TestWebKitAPI.vcproj:
* WebKitAPITest/WebKitAPITest.vcproj:
* WebKitTestRunner/win/InjectedBundle.vcproj:
* WebKitTestRunner/win/WebKitTestRunner.vcproj:
* WinLauncher/WinLauncher.vcproj:
* Scripts/webkitdirs.pm: Updated cairo build configs.
2010-11-18 Steve Falkenburg <sfalken@apple.com>
Reviewed by Adam Roben.
Windows vcproj configuration names should be normalized across projects
https://bugs.webkit.org/show_bug.cgi?id=49776
* DumpRenderTree/DumpRenderTree.sln:
* FindSafari/FindSafari.vcproj:
* TestWebKitAPI/win/TestWebKitAPI.sln:
* TestWebKitAPI/win/TestWebKitAPIGenerated.vcproj:
* TestWebKitAPI/win/TestWebKitAPIGeneratedCommon.vsprops: Added.
* WebKitAPITest/WebKitAPITest.vcproj:
* WebKitAPITest/WebKitAPITestCommon.vsprops:
* WebKitLauncherWin/WebKitLauncherWin.vcproj:
* WebKitTestRunner/WebKitTestRunner.sln:
* WebKitTestRunner/win/InjectedBundleGenerated.vcproj:
* WebKitTestRunner/win/InjectedBundleGeneratedCommon.vsprops: Added.
* record-memory-win/record-memory-win.vcproj:
2010-11-18 Gavin Barraclough <barraclough@apple.com>
Build fix - as per DRT-mac, always enabled developer extras,
in win/qt/gtk DRT.
This is necessary to enable rich exception messages to be generated
following https://bugs.webkit.org/show_bug.cgi?id=49708.
* DumpRenderTree/gtk/DumpRenderTree.cpp:
(shouldEnableDeveloperExtras):
* DumpRenderTree/qt/DumpRenderTreeQt.cpp:
(WebCore::shouldEnableDeveloperExtras):
* DumpRenderTree/win/DumpRenderTree.cpp:
(shouldEnableDeveloperExtras):
2010-11-18 Sheriff Bot <webkit.review.bot@gmail.com>
Unreviewed, rolling out r72357.
http://trac.webkit.org/changeset/72357
https://bugs.webkit.org/show_bug.cgi?id=49784
It broke test-webkitpy tests since the patch didn't have the
corresponding test update (Requested by kinuko on #webkit).
* Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
* Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
2010-11-18 Mihai Parparita <mihaip@chromium.org>
Reviewed by Tony Chang.
Rebaseline server: add rebaseline queue
https://bugs.webkit.org/show_bug.cgi?id=49763
Add a UI queue (a multi-item <select>) where tests to be baselined
can be enqueued (also supports basic removal). A queue is used instead
of immediately doing rebaselines since SCM operations can take a while.
It's better to go through lots of tests quickly to mark them as
needing rebaselining and then batch the SCM operations.
Also adds a simple log where results can be displayed.
* Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html:
* Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css:
* Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
* Scripts/webkitpy/tool/commands/data/rebaselineserver/queue.js: Added.
* Scripts/webkitpy/tool/commands/data/rebaselineserver/util.js:
* Scripts/webkitpy/tool/commands/rebaselineserver.py:
2010-11-18 Gavin Barraclough <barraclough@apple.com>
Reviewed by Oliver Hunt.
Bug 49708 - Stop recompiling functions to regenerate exception info.
Instead only hold info as necessary – keep divot info is the inspector
is enabled, line number info is debugging or profiling, and handler
info for functions with try/catch.
* DumpRenderTree/mac/DumpRenderTree.mm:
(shouldEnableDeveloperExtras):
- always enable the developer tools from DRT, to ensure we
produce rich error messages on JavaScript exceptions.
2010-11-18 Kinuko Yasuda <kinuko@chromium.org>
Reviewed by Ojan Vafai.
[Chromium] REGRESSION: Record actual test results in the JSON result summary for non-layout tests
https://bugs.webkit.org/show_bug.cgi?id=49702
* Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
* Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
2010-11-18 Dirk Pranke <dpranke@chromium.org>
Reviewed by Tony Chang.
new-run-webkit-tests: create first part of 'message_broker' class for multiprocessing fixes
Create the first version of the 'message_broker' package. This
class will encapsulate all of the threading/multiprocessing and
message-sending details for the communication between the
'manager' object and the 'worker' objects. For the moment, it
just holds some routines and tests for logging thread stacks.
There should be no functional changes in this patch, just moving stuff
around.
https://bugs.webkit.org/show_bug.cgi?id=49707
* Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
* Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread_unittest.py: Removed.
* Scripts/webkitpy/layout_tests/layout_package/message_broker.py: Added.
* Scripts/webkitpy/layout_tests/layout_package/message_broker_unittest.py: Added.
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2010-11-18 Steve Falkenburg <sfalken@apple.com>
Reviewed by Adam Roben.
Remove leftover Windows Debug_Internal configurations
https://bugs.webkit.org/show_bug.cgi?id=49758
* DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
* DumpRenderTree/win/DumpRenderTree.vcproj:
* DumpRenderTree/win/ImageDiff.vcproj:
* MiniBrowser/MiniBrowser.vcproj:
* TestWebKitAPI/win/TestWebKitAPI.vcproj:
* WebKitAPITest/WebKitAPITest.vcproj:
* WebKitTestRunner/win/InjectedBundle.vcproj:
* WebKitTestRunner/win/WebKitTestRunner.vcproj:
* WinLauncher/WinLauncher.vcproj:
2010-11-18 Adam Roben <aroben@apple.com>
Add three new Windows XP build slaves
Reviewed by Steve Falkenburg.
* BuildSlaveSupport/build.webkit.org-config/config.json: Added three
new slaves, and added them to the Windows XP Debug (Tests) builder.
2010-11-18 Daniel Bates <dbates@rim.com>
Reviewed by Adam Roben.
update-webkit-support-libs should fall back to existing
WebKitSupportLibrary version if there is no internet connectivity
https://bugs.webkit.org/show_bug.cgi?id=49503
Fall back to existing support libraries (if present) when there is
no internet connection.
Currently, update-webkit-support-libs dies with an "out-of-date"
error when there is no internet connection because it cannot
retrieve versioning information from developer.apple.com. Because
update-webkit-support-libs fails, build-webkit fails. Instead,
if there is no internet connection and the support libraries are
present then we should warn the user and exit() with success so
that build-webkit can work without an internet connection.
* Scripts/update-webkit-support-libs:
2010-11-18 Steve Falkenburg <sfalken@apple.com>
Reviewed by Adam Roben.
Debug_Internal Windows configuration is unnecessary, should be removed
https://bugs.webkit.org/show_bug.cgi?id=49753
* DumpRenderTree/win/DumpRenderTree.cpp:
* MiniBrowser/win/stdafx.h:
* TestWebKitAPI/win/PlatformUtilitiesWin.cpp:
* WebKitTestRunner/win/TestControllerWin.cpp:
2010-11-18 Steve Falkenburg <sfalken@apple.com>
Reviewed by Adam Roben.
Clean up vcproj errors
https://bugs.webkit.org/show_bug.cgi?id=49741
* MiniBrowser/MiniBrowser.vcproj:
* WebKitAPITest/WebKitAPITest.vcproj:
* WebKitTestRunner/win/InjectedBundle.vcproj:
* WebKitTestRunner/win/WebKitTestRunner.vcproj:
2010-11-18 Hayato Ito <hayato@chromium.org>
Reviewed by Shinichiro Hamaji.
[NRWT] Fix a break of '--new-baseline' feature in pixel_test
which is caused by r72249.
https://bugs.webkit.org/show_bug.cgi?id=49751
* Scripts/webkitpy/layout_tests/test_types/image_diff.py:
2010-11-18 Adam Roben <aroben@apple.com>
Implement layoutTestController.waitForPolicyDelegate on Windows
Fixes <http://webkit.org/b/25038> <rdar://problem/6790213>.
Reviewed by Simon Fraser.
* DumpRenderTree/win/LayoutTestControllerWin.cpp:
(LayoutTestController::waitForPolicyDelegate): Implemented by porting
code from LayoutTestControllerMac.mm.
2010-11-18 Hayato Ito <hayato@chromium.org>
Reviewed by Shinichiro Hamaji.
[NRWT] Fix a break of '--new-baseline' feature which is caused by r72249.
https://bugs.webkit.org/show_bug.cgi?id=49740
* Scripts/webkitpy/layout_tests/test_types/text_diff.py:
2010-11-18 Tony Chang <tony@chromium.org>
Reviewed by Adam Barth.
run platform/chromium/plugins/refcount-leaks.html on all platforms
https://bugs.webkit.org/show_bug.cgi?id=49485
Add PluginObject.testObjectCount which returns the number of allocated
TestObjects. Add PluginObject.testCreateTestObject which allocates
and returns a TestObject.
Add TestObject.refCount which returns the number of refs on the
TestObject.
* DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
(pluginGetProperty):
(pluginInvoke):
* DumpRenderTree/TestNetscapePlugIn/TestObject.cpp:
(getTestObjectCount):
(testAllocate):
(testDeallocate):
(testGetProperty):
(testScriptObjectInvoke): Release a ref to a plugin object that was
previously leaked.
* DumpRenderTree/TestNetscapePlugIn/TestObject.h:
2010-11-17 Steve Falkenburg <sfalken@apple.com>
Reviewed by Adam Roben.
WebKitTools projects (WebKitTestRunner, record-memory-win) should use vsprops for common build settings
https://bugs.webkit.org/show_bug.cgi?id=49711
* WebKitTestRunner/Configurations/WebKitTestRunnerCFLite.vsprops: Added.
* WebKitTestRunner/Configurations/WebKitTestRunnerCommon.vsprops: Added.
* WebKitTestRunner/Configurations/WebKitTestRunnerCoreFoundation.vsprops: Added.
* WebKitTestRunner/win/WebKitTestRunner.vcproj:
* record-memory-win/record-memory-win-common.vsprops: Added.
* record-memory-win/record-memory-win.vcproj:
2010-11-18 Andras Becsi <abecsi@inf.u-szeged.hu>
Reviewed by Csaba Osztrogonác.
[Qt][WK2] Only add user agent strings to the list which aren't listed yet.
* MiniBrowser/qt/BrowserWindow.cpp:
(BrowserWindow::updateUserAgentList):
2010-11-17 Adam Roben <aroben@apple.com>
Don't trigger Windows builds when chromium-win test results change
Reviewed by Mark Rowe.
* Scripts/webkitpy/common/config/build.py:
(_should_file_trigger_build): Changed the regular expression used to
search for directory names to only consider full directory names,
rather than matching directory names where the string we care about is
a suffix of the name.
* Scripts/webkitpy/common/config/build_unittest.py:
(ShouldBuildTest): Added a test.
2010-11-17 Hayato Ito <hayato@chromium.org>
Reviewed by Shinichiro Hamaji.
Remove a uri member from TestInput class.
https://bugs.webkit.org/show_bug.cgi?id=49691
* Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2010-11-17 Adam Roben <aroben@apple.com>
Move FindSafari's settings into a vsprops file
Fixes <http://webkit.org/b/49699> FindSafari should use a vsprops file
Reviewed by Steve Falkenburg.
* FindSafari/FindSafari.vcproj:
* FindSafari/FindSafariCommon.vsprops: Added.
2010-11-17 Adam Roben <aroben@apple.com>
Use a minimum font size of 0 in DRT on Windows
This is the Windows equivalent of r72164.
Rubber-stamped by Dave Hyatt.
* DumpRenderTree/win/DumpRenderTree.cpp:
(resetDefaultsToConsistentValues):
2010-11-17 Dirk Pranke <dpranke@chromium.org>
Reviewed by Andreas Kling.
webkit-patch pretty-diff shouldn't need to be run from a checkout root in git
https://bugs.webkit.org/show_bug.cgi?id=49639
* Scripts/webkitpy/common/checkout/scm.py:
2010-11-17 Mihai Parparita <mihaip@chromium.org>
Reviewed by Tony Chang.
Rebaseline server: add loupe for image diffs
https://bugs.webkit.org/show_bug.cgi?id=49692
Add a loupe (magnifiying glass) for inspecting image diffs. Shows an
enlarged area of the expected, actual and diff images side by side.
* Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html:
* Scripts/webkitpy/tool/commands/data/rebaselineserver/loupe.js: Added.
* Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
* Scripts/webkitpy/tool/commands/rebaselineserver.py:
2010-11-17 Steve Falkenburg <sfalken@apple.com>
Reviewed by Adam Roben.
WebKitLauncherWin should use vsprops for shared build settings
https://bugs.webkit.org/show_bug.cgi?id=49696
* WebKitLauncherWin/WebKitLauncherWin.vcproj:
* WebKitLauncherWin/WebKitLauncherWinCommon.vsprops: Added.
2010-11-17 Steve Falkenburg <sfalken@apple.com>
Reviewed by Adam Roben.
WinLauncher should use vsprops for shared build settings
https://bugs.webkit.org/show_bug.cgi?id=49695
* WinLauncher/WinLauncher.cpp:
(_tWinMain):
(PrintView):
(WndProc):
(MyEditProc):
* WinLauncher/WinLauncher.vcproj:
* WinLauncher/WinLauncherCommon.vsprops: Added.
2010-11-17 Steve Falkenburg <sfalken@apple.com>
Delete unused file unintentionally added in last change.
* DumpRenderTree/win/DumpRenderTreeCURL.vsprops: Removed.
2010-11-17 Steve Falkenburg <sfalken@apple.com>
Reviewed by Adam Roben.
DumpRenderTree on Windows should use a vsprops file for shared project settings
https://bugs.webkit.org/show_bug.cgi?id=49690
* DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
* DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePluginCommon.vsprops: Added.
* DumpRenderTree/win/DumpRenderTree.vcproj:
* DumpRenderTree/win/DumpRenderTreeApple.vsprops: Added.
* DumpRenderTree/win/DumpRenderTreeCFLite.vsprops: Added.
* DumpRenderTree/win/DumpRenderTreeCURL.vsprops: Added.
* DumpRenderTree/win/DumpRenderTreeCairo.vsprops: Added.
* DumpRenderTree/win/DumpRenderTreeCommon.vsprops: Added.
* DumpRenderTree/win/ImageDiff.vcproj:
* DumpRenderTree/win/ImageDiffCommon.vsprops: Added.
2010-11-17 Hayato Ito <hayato@chromium.org>
Refactor TestTypeBase.compare_output().
Introduce a TestOutput class and update compare_output() of each test
types so that they can take both actual and expected TestOutput objects.
https://bugs.webkit.org/show_bug.cgi?id=49431
* Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
* Scripts/webkitpy/layout_tests/layout_package/test_output.py: Added.
* Scripts/webkitpy/layout_tests/port/base.py:
* Scripts/webkitpy/layout_tests/port/chromium.py:
* Scripts/webkitpy/layout_tests/port/dryrun.py:
* Scripts/webkitpy/layout_tests/port/test.py:
* Scripts/webkitpy/layout_tests/port/webkit.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
* Scripts/webkitpy/layout_tests/test_types/image_diff.py:
* Scripts/webkitpy/layout_tests/test_types/test_type_base.py:
* Scripts/webkitpy/layout_tests/test_types/text_diff.py:
2010-11-17 Adam Roben <aroben@apple.com>
Make each Windows Test builder use the same OS for all its slaves
Currently each WebKit1 Windows Test builder has one Windows XP slave
and one Windows 7 slave. Having different OSes in a single builder
meant that the builders would alternate between red and green when
there were OS-specific failures.
Fixes <http://webkit.org/b/49688> Windows Test builders alternate
between red and green because of multiple OS versions
Reviewed by Mark Rowe.
* BuildSlaveSupport/build.webkit.org-config/config.json: Renamed
"Windows Release (Tests)" to "Windows 7 Release (Tests)", "Windows
Debug (Tests)" to "Windows XP Debug (Tests)", and "Windows Release
(WebKit2 Tests)" to "Windows 7 Release (WebKit2 Tests)". Updated the
builders for the two WebKit1 builders so that they would all use the
appropriate OS.
2010-11-17 James Robinson <jamesr@chromium.org>
Reviewed by Dan Bernstein.
[chromium] Set minimum font size to 0 for DRT
https://bugs.webkit.org/show_bug.cgi?id=49677
Followup for http://trac.webkit.org/changeset/72141.
Test: http/tests/misc/acid3.html
* DumpRenderTree/chromium/WebPreferences.cpp:
(WebPreferences::reset):
2010-11-17 Andras Becsi <abecsi@webkit.org>
Unreviewed build fix.
[Qt] Add missing file I forgot in r72220.
* MiniBrowser/MiniBrowser.qrc: Added.
2010-11-17 Andras Becsi <abecsi@webkit.org>
Reviewed by Andreas Kling.
[Qt][WK2] Make it possible to set custom user agent strings in MiniBrowser.
https://bugs.webkit.org/show_bug.cgi?id=49627
Add a user agent dialog and a resource file for useragentlist.txt to MiniBrowser.
* MiniBrowser/MiniBrowser.qrc: Added.
Add it here to prevent qmake from detecting it
since it needs to be copied to the build directory.
* MiniBrowser/qt/BrowserWindow.cpp:
(BrowserWindow::BrowserWindow):
(BrowserWindow::updateUserAgentList):
(BrowserWindow::showUserAgentDialog):
* MiniBrowser/qt/BrowserWindow.h:
* MiniBrowser/qt/MiniBrowser.pro:
2010-11-17 Adam Roben <aroben@apple.com>
Make changes to the Mac WebKit2 Skipped file trigger Windows builds
We use that file on Windows!
Fixes <http://webkit.org/b/49643> Changes to
LayoutTests/platform/mac-wk2 should trigger a Windows build, but don't
Reviewed by Steve Falkenburg.
* Scripts/webkitpy/common/config/build.py:
(_should_file_trigger_build): Added the mac-wk2 directory and make it
trigger builds on SnowLeopard and Windows.
* Scripts/webkitpy/common/config/build_unittest.py:
(ShouldBuildTest.test_should_build): Added a test.
2010-11-17 Satish Sampath <satish@chromium.org>
Reviewed by Jeremy Orlow.
Clear the speech input mock explicitly before each test.
https://bugs.webkit.org/show_bug.cgi?id=49660
* DumpRenderTree/chromium/LayoutTestController.cpp:
(LayoutTestController::setMockSpeechInputResult):
* DumpRenderTree/chromium/LayoutTestController.h:
* DumpRenderTree/chromium/WebViewHost.cpp:
(WebViewHost::speechInputController):
(WebViewHost::reset): Invoke speech input mock's clearResults method.
* DumpRenderTree/chromium/WebViewHost.h:
(WebViewHost::speechInputControllerMock): Moved mock from LayoutTestController.
2010-11-17 Alexander Pavlov <apavlov@chromium.org>
Unreviewed, specify IRC nickname in committers.py.
* Scripts/webkitpy/common/config/committers.py:
2010-11-17 Gabor Rapcsanyi <rgabor@inf.u-szeged.hu>
Reviewed by Tony Chang.
[NRWT] Make http locking similar to perl implementation
https://bugs.webkit.org/show_bug.cgi?id=49187
* Scripts/webkitpy/common/system/file_lock.py: Added.
* Scripts/webkitpy/common/system/file_lock_unittest.py: Added.
* Scripts/webkitpy/layout_tests/port/http_lock.py:
2010-11-17 MORITA Hajime <morrita@google.com>
Reviewed by Kent Tamura.
[Chromium][DRT] EventSender.contextClick() should aware spellchecking
https://bugs.webkit.org/show_bug.cgi?id=49366
- EvenSender: Checked WebContextMenuData.misspelledWord and added extra context menu entries
according to the spellchecker suggestion
- MockSpellCheck: add fillSuggestionList to provide fake suggestions.
test_expectations.txt will be changed after this change is ported to test_shell.
* DumpRenderTree/chromium/EventSender.cpp:
(makeMenuItemStringsFor):
(EventSender::contextClick):
* DumpRenderTree/chromium/MockSpellCheck.cpp:
(MockSpellCheck::fillSuggestionList):
(MockSpellCheck::initializeIfNeeded):
* DumpRenderTree/chromium/MockSpellCheck.h:
* DumpRenderTree/chromium/WebViewHost.cpp:
(WebViewHost::mockSpellCheck):
* DumpRenderTree/chromium/WebViewHost.h:
2010-11-16 Dave Hyatt <hyatt@apple.com>
Reviewed by Sam Weinig.
Make sure the pref for minimum font size is just explicitly set to 0 every time,
since the bots have 1 stuck in their plist otherwise.
* DumpRenderTree/mac/DumpRenderTree.mm:
(resetDefaultsToConsistentValues):
2010-11-16 Mihai Parparita <mihaip@chromium.org>
Reviewed by Tony Chang.
Rebaseline server: compute diffs client-side
https://bugs.webkit.org/show_bug.cgi?id=49640
The image diff output from the DRT is pretty bad on some ports (at
least the Mac one), so it's better to compute diffs on the client by
using <canvas>.
* Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html:
* Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css:
* Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
2010-11-16 Mihai Parparita <mihaip@chromium.org>
Reviewed by Tony Chang.
Rebaseline server: display test results
https://bugs.webkit.org/show_bug.cgi?id=49626
Adds basic result display to the rebaseline server. On the Python side
this involves:
- Parsing the unexpected_results.json into a dictionary.
- Serving it as JSON under /results.json.
(the JSON -> dict -> JSON transform isn't strictly necessary right now,
but I'll need to have access to the parsed results on the Python side
for follow-up changes).
On the web UI side this adds:
- Markup for display image and text results (expected, actual, diff),
and JS for populating it.
- Markup for breaking down test results by failure type and directory,
and JS for populating it.
* Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html:
* Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css:
* Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
* Scripts/webkitpy/tool/commands/data/rebaselineserver/util.js: Added.
* Scripts/webkitpy/tool/commands/rebaselineserver.py:
2010-11-16 Dirk Pranke <dpranke@chromium.org>
Reviewed by Ojan Vafai.
See Chromium issue http://codereview.chromium.org/5133001/ - we
are modifying the build bots to pass in the builder name with
the "GPU" string appended instead of appending it in the code.
https://bugs.webkit.org/show_bug.cgi?id=49636
* Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
2010-11-16 Steve Falkenburg <sfalken@apple.com>
Reviewed by Adam Roben.
Disable LTCG for Windows Release builds. Add new Release_LTCG configuration.
https://bugs.webkit.org/show_bug.cgi?id=49632
* DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
* DumpRenderTree/win/DumpRenderTree.vcproj:
* DumpRenderTree/win/ImageDiff.vcproj:
* MiniBrowser/MiniBrowser.vcproj:
* TestWebKitAPI/win/TestWebKitAPI.vcproj:
* WebKitLauncherWin/WebKitLauncherWin.vcproj:
* WebKitTestRunner/win/InjectedBundle.vcproj:
* WebKitTestRunner/win/WebKitTestRunner.vcproj:
* WinLauncher/WinLauncher.vcproj:
* record-memory-win/record-memory-win.vcproj:
2010-11-16 Dirk Pranke <dpranke@chromium.org>
Reviewed by Ojan Vafai.
new-run-webkit-tests: rename TestInfo to TestInput, move image hash to work thread
Rename the TestInfo class to TestInput to be clearer about its
function, and move the checksum-reading code into dump_render_tree_thread
to avoid cross-thread access.
https://bugs.webkit.org/show_bug.cgi?id=49573
* Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2010-11-16 Dave Hyatt <hyatt@apple.com>
Fix bustage. Remove the minimum font size pref setting in DumpRenderTree.
* DumpRenderTree/mac/DumpRenderTree.mm:
(resetDefaultsToConsistentValues):
2010-11-16 Stephanie Lewis <slewis@apple.com>
Reviewed by Geoff Garen.
<rdar://problem/8624267> Leak creating offscreen webview running fast/dom tests
Use shared DeviceOrientationProviderMock. The old allocation method created a new WebDeviceOrientationProviderMock
for every WebView, and the WebKit API doesn't support that ownership model.
I also fixed a leak in setMockDeviceOrientation by adding a missing release.
* DumpRenderTree/mac/DumpRenderTree.mm:
(createWebViewAndOffscreenWindow):
* DumpRenderTree/mac/LayoutTestControllerMac.mm:
(LayoutTestController::setMockDeviceOrientation):
2010-11-16 Dirk Pranke <dpranke@chromium.org>
Reviewed by Ojan Vafai.
new-run-webkit-tests: log the process id in --verbose mode
As part of the switch to multi-process mode, we should log the
process id when running w/ --verbose.
https://bugs.webkit.org/show_bug.cgi?id=49571
* Scripts/webkitpy/layout_tests/layout_package/printing.py:
2010-11-16 Adam Roben <aroben@apple.com>
Only trigger Windows builds when files that we actually use on Windows
are changed
A new Scheduler subclass, PlatformSpecificScheduler, has been added.
It uses the new webkitpy.common.config.build module to determine
whether a particular change should trigger a build on a particular
platform. The Windows builders have been switched to use a
PlatformSpecificScheduler.
The logic to determine whether or not a particular change should
trigger a build on a given platform has only been implemented/tested
for Windows. I tried to make it easy to add more platforms in the
future, but I don't have enough familiarity with all platforms to be
able to implement it for them.
Fixes <http://webkit.org/b/49407> Windows builders kick off builds for
lots irrelevant changes (e.g., rebaselining Chromium test results)
Reviewed by Eric Seidel.
* BuildSlaveSupport/build.webkit.org-config/config.json: Use a
PlatformSpecificScheduler for the Windows builders.
* BuildSlaveSupport/build.webkit.org-config/master.cfg:
(PlatformSpecificScheduler.__init__): Added. Stores our platform, sets
up a ChangeFilter that filters to our branch and will call through to
our filter method, and calls up to our base class.
(PlatformSpecificScheduler.filter): Calls through to
build.should_build to find out whether we should trigger a build for
this change.
* Scripts/webkitpy/common/net/build.py: Added.
(_should_file_trigger_build): Uses a set of directories and regexp
patterns to determine whether the given file should trigger a build on
the given platform. As mentioned earlier, this has only been
implemented for Windows, though I did try to make some guesses about
other platforms.
(should_build): Returns true if any of the files should trigger a
build on the given platform.
* Scripts/webkitpy/common/net/build_unittest.py: Added.
(ShouldBuildTest.test_should_build): Does some basic testing to make
sure we're triggering builds for the right files. It only tests
Windows for now, though I tried to make some guesses about other
platforms.
2010-11-16 Simon Hausmann <simon.hausmann@nokia.com>
Reviewed by Andreas Kling.
[Qt] Remove synchronous QWebPage::checkPermissions signal
https://bugs.webkit.org/show_bug.cgi?id=46810
* DumpRenderTree/qt/DumpRenderTreeQt.cpp:
(WebCore::WebPage::WebPage):
* DumpRenderTree/qt/LayoutTestControllerQt.cpp:
(LayoutTestController::grantDesktopNotificationPermission): When granting
permission, grant it directly on the QWebPage/Frame, that will remember it.
* QtTestBrowser/webpage.cpp:
(WebPage::WebPage):
* QtTestBrowser/webpage.h:
2010-11-16 Andras Becsi <abecsi@inf.u-szeged.hu>
Reviewed by Csaba Osztrogonác.
[Qt] Add isPrinting and setPrinting methods to DRT's LayoutTestController.
* DumpRenderTree/qt/LayoutTestControllerQt.cpp:
(LayoutTestController::reset):
* DumpRenderTree/qt/LayoutTestControllerQt.h:
(LayoutTestController::isPrinting):
(LayoutTestController::setPrinting):
2010-11-16 Balazs Kelemen <kbalazs@webkit.org>
Reviewed by Andreas Kling.
[Qt][WK2] Add a way to turn on tiled backing store in MiniBrowser
https://bugs.webkit.org/show_bug.cgi?id=49587
* MiniBrowser/qt/BrowserView.cpp:
(BrowserView::BrowserView): Added BackingStoreType
argument to ctor. Propagating it to the QGraphicsWKView.
* MiniBrowser/qt/BrowserView.h:
* MiniBrowser/qt/BrowserWindow.cpp:
(BrowserWindow::BrowserWindow):
(BrowserWindow::newWindow): Create the new BrowserWindow
with the same BackingStoreType that this instance has.
* MiniBrowser/qt/BrowserWindow.h: Added BackingStoreType member.
* MiniBrowser/qt/main.cpp:
(main): Use tiled backing store if got -tiled command line argument.
Avoid redundant copy of command line args.
2010-11-16 John Knottenbelt <jknotten@chromium.org>
Reviewed by Jeremy Orlow.
Move DeviceOrientationClientMock from LayoutTestController to WebViewHost.
https://bugs.webkit.org/show_bug.cgi?id=48506
This ensures that there is one mock per page WebView / WebViewHost. This is a
design constraint of the DeviceOrientationClientMock because the m_controller
field can only store a pointer to a single instance of the corresponding
DeviceOrientationController.
Test: fast/dom/DeviceOrientation/no-page-cache.html
* DumpRenderTree/chromium/LayoutTestController.cpp:
(LayoutTestController::setMockDeviceOrientation):
(LayoutTestController::deviceOrientationClientMock):
* DumpRenderTree/chromium/LayoutTestController.h:
* DumpRenderTree/chromium/WebViewHost.cpp:
(WebViewHost::deviceOrientationClientMock):
(WebViewHost::deviceOrientationClient):
(WebViewHost::reset):
* DumpRenderTree/chromium/WebViewHost.h:
2010-11-15 Andras Becsi <abecsi@webkit.org>
Reviewed by Andreas Kling.
[Qt][WK2] Avoid polling in WebKitTestRunner.
https://bugs.webkit.org/show_bug.cgi?id=49542
Make the WK2 testing session about 10-15% faster and decrease random flakiness
resulting from timing skew by removing the millisecond polling from TestControllerQt.
* WebKitTestRunner/TestController.cpp:
(WTR::TestController::didFinishLoadForFrame):
* WebKitTestRunner/TestController.h:
* WebKitTestRunner/TestInvocation.cpp:
(WTR::TestInvocation::didReceiveMessageFromInjectedBundle):
* WebKitTestRunner/mac/TestControllerMac.mm:
(WTR::TestController::notifyDone): Add method stub.
* WebKitTestRunner/qt/TestControllerQt.cpp:
(WTR::TestControllerRunLoop::instance):
(WTR::TestControllerRunLoop::start):
(WTR::TestControllerRunLoop::stop):
(WTR::TestControllerRunLoop::TestControllerRunLoop):
(WTR::TestControllerRunLoop::timerEvent):
(WTR::TestController::notifyDone):
(WTR::TestController::platformRunUntil):
* WebKitTestRunner/win/TestControllerWin.cpp:
(WTR::TestController::notifyDone): Add method stub.
2010-11-15 Darin Adler <darin@apple.com>
* Scripts/webkitpy/common/net/bugzilla: Added property svn:ignore.
2010-11-15 Benjamin Poulain <benjamin.poulain@nokia.com>
Reviewed by Andreas Kling.
[Qt] [WK2] Add Ctrl+L shortcut to Qt Minibrowser
https://bugs.webkit.org/show_bug.cgi?id=49544
Add the shortcut to BrowserWindow.
* MiniBrowser/qt/BrowserWindow.cpp:
(BrowserWindow::BrowserWindow):
(BrowserWindow::openLocation):
* MiniBrowser/qt/BrowserWindow.h:
2010-11-14 David Levin <levin@chromium.org>
Reviewed by Daniel Bates.
check-webkit-style should detect PassRefPtr usage in functions.
https://bugs.webkit.org/show_bug.cgi?id=49513
* Scripts/webkitpy/style/checkers/cpp.py:
(check_for_function_lengths): Revert a comment change that I
accidentally made in r71986.
(check_pass_ptr_usage): Added the code to do the check.
(process_line): Added the call to check_pass_ptr_usage.
(CppChecker): Added the new error category.
* Scripts/webkitpy/style/checkers/cpp_unittest.py:
(CppStyleTestBase::perform_pass_ptr_check): Runs the new check for
testing purposes.
(PassPtrTest::*): The class/functions to unit test the new
functionality.
2010-11-14 David Levin <levin@chromium.org>
Reviewed by Shinichiro Hamaji.
check-webkit-style function detection and the line count style checks should be separate.
https://bugs.webkit.org/show_bug.cgi?id=49512
* Scripts/webkitpy/style/checkers/cpp.py: Do the separation.
* Scripts/webkitpy/style/checkers/cpp_unittest.py: Adjust the test to
call the detection function and fix line counts in two places now that
the code really only counts the lines in the body of the function.
2010-11-14 Andreas Kling <kling@webkit.org>
Reviewed by Antonio Gomes.
http/tests/plugins tests print "Unhandled variable" to stderr under Qt
https://bugs.webkit.org/show_bug.cgi?id=33438
Stifle "Unhandled variable" warning to match what the other
TestNetscapePlugin does (NPP_GetValue in TestNetscapePlugIn/main.cpp
simply returns NPERR_GENERIC_ERROR for unhandled variables.)
* DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
(webkit_test_plugin_get_value):
2010-11-13 David Levin <levin@chromium.org>
Reviewed by Eric Seidel.
check-webkit-style function detection crashes on functions in templates.
https://bugs.webkit.org/show_bug.cgi?id=49504
For example "template <bool shouldClose(const Element*)>".
* Scripts/webkitpy/style/checkers/cpp.py: If not function is detected, bail out.
* Scripts/webkitpy/style/checkers/cpp_unittest.py: Modified the complex
function detection test to expose this issue.
2010-11-12 David Levin <levin@chromium.org>
Reviewed by Shinichiro Hamaji.
check-webkit-style function detection doesn't detect indented functions declaractions.
https://bugs.webkit.org/show_bug.cgi?id=49446
Indented function declarations occur inside class definitions, so
they are a pretty common (and worth detecting).
* Scripts/webkitpy/style/checkers/cpp.py:
Changed regex to allow indentation.
Changed the function start detection to only happen when not in a
function.
Changed function end detection to work based on matching braces
instead of finding a close brace at the beginning of the line.
Fixed close_expression to do what it says when it doesn't find
the close.
* Scripts/webkitpy/style/checkers/cpp_unittest.py: Indented function test.
2010-11-12 Daniel Bates <dbates@rim.com>
Rubber-stamped by Nikolas Zimmermann.
Make do-file-rename rename files in the directory WebKit2.
This also makes the list of searched directories in do-file-rename
consistent with the list of searched directories in the script do-webcore-rename.
* Scripts/do-file-rename:
2010-11-12 Dirk Pranke <dpranke@chromium.org>
Reviewed by James Robinson.
Attempt yet again to land the fix for bug 49360 (respecting
set-webkit-configuration). We need to handle the cases where
trying to run webkit-build-directory to find out where the
default configuration might be fails (that shows up on some
Chromium bots that apparently don't have perl installed).
https://bugs.webkit.org/show_bug.cgi?id=49360
* Scripts/webkitpy/layout_tests/port/config.py:
* Scripts/webkitpy/layout_tests/port/config_standalone.py: Added.
* Scripts/webkitpy/layout_tests/port/config_unittest.py:
2010-11-12 Mihai Parparita <mihaip@chromium.org>
Reviewed by Adam Barth.
webkit-patch land logs commit message twice
https://bugs.webkit.org/show_bug.cgi?id=49482
Don't log the comment that's used by bugzilla.close_bug_as_fixed, since
that's just the commit message, which we now output following the
actual commit.
* Scripts/webkitpy/common/net/bugzilla/bugzilla.py:
2010-11-12 Tony Chang <tony@chromium.org>
Reviewed by Adam Barth.
run platform/chromium/plugins/script-object-invoke.html on all platforms
https://bugs.webkit.org/show_bug.cgi?id=49280
* DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
(testScriptObjectInvoke):
(pluginInvoke):
2010-11-12 Mihai Parparita <mihaip@chromium.org>
Reviewed by Adam Barth.
webkit-patch land-cowboy (and other commands) should display the committed SVN revision after landing a patch
https://bugs.webkit.org/show_bug.cgi?id=49471
Right now there's no indication of a successful commit.
* Scripts/webkitpy/tool/commands/download_unittest.py:
* Scripts/webkitpy/tool/steps/commit.py:
2010-11-12 Kinuko Yasuda <kinuko@chromium.org>
Reviewed by Ojan Vafai.
Include detailed test modifiers in results.json and enable incremental uploading for non-layout tests
https://bugs.webkit.org/show_bug.cgi?id=49354
Also moved/integrated the upload method from run_webkit_tests.py to json_results_generator.py.
* Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
* Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
* Scripts/webkitpy/layout_tests/layout_package/json_results_generator_unittest.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2010-11-12 Tony Chang <tony@chromium.org>
Reviewed by Kent Tamura.
[chromium] Load a blank page when closing DRT WebViewHosts
https://bugs.webkit.org/show_bug.cgi?id=49417
In test_shell, when we close a window, we first load about:blank
and run GC to fire the destruction logic of the page. In test_shell,
this happens in ~TestShell. In DRT, we manage virtual windows as
WebViewHosts, so we need to replicate this logic in ~WebViewHost.
Managing focus is a bit trickier, since normally you would get window
events to reset the focus to the right window. We do it manually
instead.
This fixes 2 plugin tests on DRT.
* DumpRenderTree/chromium/TestShell.cpp:
(TestShell::~TestShell): Move window close logic to ~WebViewHost.
(TestShell::closeWindow): Move widget close() to WVH destructor and
manage focus manually.
(TestShell::closeRemainingWindows): Comment typo.
* DumpRenderTree/chromium/WebViewHost.cpp:
(WebViewHost::closeWidget):
(invokeCloseWidget):
(WebViewHost::closeWidgetSoon): Close the window asynchronously (matches
test_shell).
(WebViewHost::~WebViewHost):
* DumpRenderTree/chromium/WebViewHost.h:
2010-11-12 David Levin <levin@chromium.org>
Reviewed by Adam Barth.
cpp.py needs some tidying.
https://bugs.webkit.org/show_bug.cgi?id=49443
* Scripts/webkitpy/style/checkers/cpp.py:
Fix parameter ordering for iteratively_replace_matches_with_char
to mirror other similar functions (and improved comments, etc.)
Also, use a python-ism (for/else) to get rid of a variable that
was there to determine if something had been found in the loop.
2010-11-12 Robert Kroeger <rjkroege@chromium.org>
Reviewed by James Robinson.
[Chromium/DRT] Make EventSendingController send complete touch events.
https://bugs.webkit.org/show_bug.cgi?id=49285
* DumpRenderTree/chromium/EventSender.cpp:
(EventSender::addTouchPoint):
(EventSender::updateTouchPoint):
(EventSender::sendCurrentTouchEvent):
2010-11-12 Mihai Parparita <mihaip@chromium.org>
Unreviewed rollout of r71858.
Rollout out r71858 since it breaks new-run-webkit-httpd as used by the
NaCl tests.
* Scripts/webkitpy/layout_tests/port/config.py:
* Scripts/webkitpy/layout_tests/port/config_standalone.py: Removed.
* Scripts/webkitpy/layout_tests/port/config_unittest.py:
2010-11-12 Benjamin Poulain <benjamin.poulain@nokia.com>
Reviewed by Kenneth Rohde Christiansen.
[Qt] TestController::runUntil doesn't honor the timeout parameter
https://bugs.webkit.org/show_bug.cgi?id=48941
Implement the missing timeout of WebKitTestRunner for Qt.
A QElapsedTimer has been added to RunUntilConditionLoop in order
to measure how long the loop has been running. When the timer
is bigger or equal than the timeout, the loop ends even if the
condition is not met.
* WebKitTestRunner/qt/TestControllerQt.cpp:
(WTR::RunUntilConditionLoop::start):
(WTR::RunUntilConditionLoop::run):
(WTR::RunUntilConditionLoop::timerEvent):
(WTR::TestController::platformRunUntil):
2010-11-11 David Levin <levin@chromium.org>
Reviewed by Shinichiro Hamaji.
check-webkit-style function detection doesn't handle templates with spaces.
https://bugs.webkit.org/show_bug.cgi?id=49427
* Scripts/webkitpy/style/checkers/cpp.py: Stripped out templates when
finding the function name.
* Scripts/webkitpy/style/checkers/cpp_unittest.py: Made the test have a
template with a space.
2010-11-11 Ademar de Souza Reis Jr <ademar.reis@openbossa.org>
Reviewed by Shinichiro Hamaji.
Enable check-webkit-style on Qt files
https://bugs.webkit.org/show_bug.cgi?id=48322
Adding exception for _q_... private slots and
..._data() methods in tests
* Scripts/webkitpy/style/checker.py: enable Qt dirs, add exceptions
* Scripts/webkitpy/style/checker_unittest.py: improve unittests
2010-11-11 David Levin <levin@chromium.org>
Reviewed by Shinichiro Hamaji.
check-webkit-style function detection doesn't work for templates and destructors.
https://bugs.webkit.org/show_bug.cgi?id=49425
* Scripts/webkitpy/style/checkers/cpp.py: Changed regex for templates and destructors.
* Scripts/webkitpy/style/checkers/cpp_unittest.py: Made the test hit this issue.
2010-11-11 Joone Hur <joone@kldp.org>
Reviewed by Antonio Gomes.
[GTK][DRT] Implement LayoutTestController::nodesFromRect
https://bugs.webkit.org/show_bug.cgi?id=46598
Support nodesFromRect in DRT
* DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
(LayoutTestController::nodesFromRect):
2010-11-11 MORITA Hajime <morrita@google.com>
Reviewed by Kent Tamura.
editing/selection/context-menu-on-text.html fails on chromium
https://bugs.webkit.org/show_bug.cgi?id=45898
Enhance Chromium's EventSender::contextClick() to return
array of menu strings. The implementation incomplete because
Although actual context menu is implemented by the browser,
it reflects editability of underlying context thus makes
context-menu-on-text.html passed.
* DumpRenderTree/chromium/EventSender.cpp:
(makeMenuItemStringsFor): Added.
(EventSender::contextClick): Now returns array of menu strings.
* DumpRenderTree/chromium/WebViewHost.cpp:
(WebViewHost::showContextMenu): Now storing given WebContextMenuData object.
(WebViewHost::clearContextMenuData): Added.
(WebViewHost::lastContextMenuData): Added.
* DumpRenderTree/chromium/WebViewHost.h:
2010-11-11 Dirk Pranke <dpranke@chromium.org>
Unreviewed, build fix.
fix build breakage caused by fix for bug 49380 (r71858).
config_standalone needs to ensure that the packages it needs are
in sys.path.
https://bugs.webkit.org/show_bug.cgi?id=49419
* Scripts/webkitpy/layout_tests/port/config_standalone.py:
2010-11-11 Dirk Pranke <dpranke@chromium.org>
Reviewed by Adam Roben.
Fix NRWT to respect set-webkit-configuration again :(
This change fixes a typo in config.py that was causing the wrong
value to be read initially and us never actually looking into
the filesystem to get the default configuration.
* Scripts/webkitpy/layout_tests/port/config.py:
* Scripts/webkitpy/layout_tests/port/config_standalone.py:
* Scripts/webkitpy/layout_tests/port/config_unittest.py:
2010-11-11 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
Split out Bug, Attachment and CommitterValidator from bugzilla.py
https://bugs.webkit.org/show_bug.cgi?id=49403
CommitterValidator really had nothing to do with bugzilla.py.
I've put it next to committers.py for now since its rather tied to that file.
* Scripts/webkitpy/common/config/committervalidator.py: Added.
* Scripts/webkitpy/common/config/committervalidator_unittest.py: Added.
* Scripts/webkitpy/common/net/bugzilla/__init__.py:
* Scripts/webkitpy/common/net/bugzilla/attachment.py: Added.
* Scripts/webkitpy/common/net/bugzilla/bug.py: Added.
* Scripts/webkitpy/common/net/bugzilla/bug_unittest.py: Added.
* Scripts/webkitpy/common/net/bugzilla/bugzilla.py:
* Scripts/webkitpy/common/net/bugzilla/bugzilla_unittest.py:
* Scripts/webkitpy/tool/bot/feeders.py:
* Scripts/webkitpy/tool/commands/queues.py:
2010-11-11 Dirk Pranke <dpranke@chromium.org>
Reviewed by Eric Seidel.
fix missing 'pid' argument for http lockfile checking on win32
https://bugs.webkit.org/show_bug.cgi?id=49363
* Scripts/webkitpy/common/system/executive.py:
2010-11-11 Eric Seidel <eric@webkit.org>
Rubber-stamped by Adam Barth.
Move bugzilla.py into its own module in preparation for splitting one-file-per-class
https://bugs.webkit.org/show_bug.cgi?id=49402
* Scripts/webkitpy/common/net/bugzilla/__init__.py: Added.
* Scripts/webkitpy/common/net/bugzilla/bugzilla.py: Renamed from WebKitTools/Scripts/webkitpy/common/net/bugzilla.py.
* Scripts/webkitpy/common/net/bugzilla/bugzilla_unittest.py: Renamed from WebKitTools/Scripts/webkitpy/common/net/bugzilla_unittest.py.
2010-11-10 Chris Guillory <chris.guillory@google.com>
Reviewed by Chris Fleizach.
Use IAccessibleComparable to compare accessibility objects.
https://bugs.webkit.org/show_bug.cgi?id=49118
* DumpRenderTree/AccessibilityUIElement.cpp:
(AccessibilityUIElement::isEqual):
* DumpRenderTree/AccessibilityUIElement.h:
* DumpRenderTree/win/AccessibilityUIElementWin.cpp:
(comparableObject):
(AccessibilityUIElement::isEqual):
2010-11-10 Csaba Osztrogonác <ossy@webkit.org>
Reviewed by David Hyatt.
HTML5 Ruby support should be mandatory feature
https://bugs.webkit.org/show_bug.cgi?id=49272
Remove Ruby as optional feature.
* Scripts/build-webkit:
2010-11-10 Adam Roben <aroben@apple.com>
Print VC++ Express build logs in the same order that the projects are
built
Fixes <http://webkit.org/b/49326> print-vse-failure-logs prints logs
in an unhelpful order
Reviewed by Eric Seidel.
* Scripts/print-vse-failure-logs:
(PrintVisualStudioExpressLogs._build_order): Added. Uses
print-msvc-project-dependencies to figure out the order in which
projects are built.
(PrintVisualStudioExpressLogs._sort_buildlogs): Added. Sorts the logs
based on their build order and project name.
(PrintVisualStudioExpressLogs._obj_directory): Moved code to find the
scripts directory from here...
(PrintVisualStudioExpressLogs._scripts_directory): ...to here.
(PrintVisualStudioExpressLogs.main): Sort the logs before printing
them.
2010-11-10 Adam Roben <aroben@apple.com>
Only print the interesting text from VC++ Express build logs
Fixes <http://webkit.org/b/49325> It's hard to find the interesting
output from print-vse-failure-logs
Reviewed by Eric Seidel.
* Scripts/print-vse-failure-logs: Removed unnecessary os.path import,
added newly-required imports. Removed a comment that was essentially a
FIXME, and is now fixed.
(PrintVisualStudioExpressLogs._relevant_text): Added. Uses
BeautifulSoup to extract the relevant text from the build log.
(PrintVisualStudioExpressLogs.main): Only print the relevant text.
2010-11-09 Mihai Parparita <mihaip@chromium.org>
Reviewed by Tony Chang.
[Chromium] http/tests/security/XFrameOptions fail with Chromium DRT
https://bugs.webkit.org/show_bug.cgi?id=49286
The tests in http/tests/security/XFrameOptions started to fail after
r71297, since it changed the WebViewHost::assignIdentifierToRequest
logic to always assign identifiers to requests, even if we're not going
to be dumping resource load callbacks at that point. These tests
only call LayoutTestController.dumpResourceLoadCallbacks after the
request is started, thus their expectations have "<unknown>" as the
identifier. Change WebViewHost to the previous behavior, which also
match the other ports'.
* DumpRenderTree/chromium/WebViewHost.cpp:
(WebViewHost::assignIdentifierToRequest):
(WebViewHost::removeIdentifierForRequest):
2010-11-08 Tony Chang <tony@chromium.org>
Reviewed by Adam Barth.
run platform/chromium/plugins/return-npobject.html on all platforms
https://bugs.webkit.org/show_bug.cgi?id=49103
This tests that we can get an NPObject returned through a method on
an NPAPI Object.
* DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
(testCallbackReturn):
(pluginInvoke):
2010-11-09 Sheriff Bot <webkit.review.bot@gmail.com>
Unreviewed, rolling out r71672.
http://trac.webkit.org/changeset/71672
https://bugs.webkit.org/show_bug.cgi?id=49276
It doesn't work on Chromium Windows bot (Requested by Ossy on
#webkit).
* Scripts/webkitpy/common/system/file_lock.py: Removed.
* Scripts/webkitpy/common/system/file_lock_unittest.py: Removed.
* Scripts/webkitpy/layout_tests/port/http_lock.py:
2010-11-09 Gabor Rapcsanyi <rgabor@inf.u-szeged.hu>
Reviewed by Tony Chang.
[NRWT] Make http locking similar to perl implementation
https://bugs.webkit.org/show_bug.cgi?id=49187
* Scripts/webkitpy/common/system/file_lock.py: Added.
* Scripts/webkitpy/common/system/file_lock_unittest.py: Added.
* Scripts/webkitpy/layout_tests/port/http_lock.py:
2010-11-09 James Kozianski <koz@chromium.org>
Reviewed by Eric Seidel.
Clean up imports for webkit-patch.
https://bugs.webkit.org/show_bug.cgi?id=49083
Move some imports from commands into main.py where they are used.
Move command imports into the __init__ of the commands module.
* Scripts/webkitpy/tool/commands/__init__.py:
* Scripts/webkitpy/tool/commands/download.py:
* Scripts/webkitpy/tool/commands/queues.py:
* Scripts/webkitpy/tool/main.py:
2010-11-09 Yi Shen <yi.4.shen@nokia.com>
Reviewed by Andreas Kling.
[Qt][QtTestBrowser] Don't erase incorrect url in the Url Bar
https://bugs.webkit.org/show_bug.cgi?id=49047
* QtTestBrowser/mainwindow.cpp:
(MainWindow::setAddressUrl):
2010-11-09 Andras Becsi <abecsi@webkit.org>
Reviewed by Csaba Osztrogonác.
Improve old-run-webkit-tests --verbose to show DumpTool opening/closing
https://bugs.webkit.org/show_bug.cgi?id=49252
* Scripts/old-run-webkit-tests:
2010-11-09 Csaba Osztrogonác <ossy@webkit.org>
Reviewed by Tor Arne Vestbø.
Output of old-run-webkit-tests is buggy in verbose mode
https://bugs.webkit.org/show_bug.cgi?id=49249
* Scripts/old-run-webkit-tests: Redundant outputs removed.
2010-11-08 Nicolas Dufresne <nicolas.dufresne@collabora.co.uk>
Reviewed by Martin Robinson.
[GTK] Link with target name set does not work
https://bugs.webkit.org/show_bug.cgi?id=48865
When a new page is created with a name (target=myFrame), the new
mainFrame could not be found because they where not stored in the
same PageGroup. As PageGroup are not exposed externally so the
simpliest solution is to use a global page group name. This also fixes
issue with visited link coloration across pages. After this change the
private function webkit_web_view_set_group_name() was no longer used
so it was removed completly.
* DumpRenderTree/gtk/DumpRenderTree.cpp:
(createWebView):
2010-11-08 John Knottenbelt <jknotten@chromium.org>
Reviewed by Kent Tamura.
WebViewHost::reset() uses placement new.
https://bugs.webkit.org/show_bug.cgi?id=49069
WebViewHost is using placement destruction / new to simulate a fresh
WebViewHost object at the same address. This is because the WebView remains
open across tests and maintains a pointer to the WebViewHost.
This change resets member variables explictly instead of the placement new dance.
* DumpRenderTree/chromium/WebViewHost.cpp:
(WebViewHost::WebViewHost):
(WebViewHost::reset):
2010-11-08 Dirk Pranke <dpranke@chromium.org>
Reviewed by Eric Seidel.
Fix the webkit_base_dir logic in webkitpy/layout_tests/port/config.py.
It turns out that NRWT can't use the code in scm.find_checkout_root()
because the Chromium bots don't do full checkouts of the WebKit
tree; they only check out subdirectories like WebKitTools/Scripts.
Until we can figure out a better approach for this, I've
restored the base_dir-detecting code from NRWT, which works in
any directory tree, scm or no.
This also restores the files modified in r71475 and r71474.
https://bugs.webkit.org/show_bug.cgi?id=49151
* WebKitTools/Scripts/webkitpy/layout_tests/port/config.py
* WebKitTools/Scripts/webkitpy/layout_tests/port/config_unittest.py
2010-11-08 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
webkit-patch failure-reason explodes if a build is missing
https://bugs.webkit.org/show_bug.cgi?id=49195
This is likely a recent regression. Adding a unit test for this case.
* Scripts/webkitpy/common/net/buildbot.py:
* Scripts/webkitpy/common/net/buildbot_unittest.py:
2010-11-08 Dan Bernstein <mitz@apple.com>
Reviewed by Darin Adler.
Ignore the system scrollbar setting.
* DumpRenderTree/mac/DumpRenderTree.mm:
(resetDefaultsToConsistentValues): Set the scrollbar default.
2010-11-05 Adam Roben <aroben@apple.com>
Make webkitpy.common.system.executive_unittest pass when running under
Win32 Python
Fixes <http://webkit.org/b/49033>.
Reviewed by Dave Levin and Eric Seidel.
* Scripts/webkitpy/common/system/executive.py:
(Executive._run_command_with_teed_output): Pass the arguments through
encode_argument_if_needed rather than using Cygwin-specific code here.
(Executive.run_and_throw_if_fail): Use child_process_encoding to decode
the output.
(Executive.run_command): Use encode_argument_if_needed to encode the
arguments and child_process_encoding to decode the output.
(Executive._child_process_encoding): Returns the encoding that should be
used when communicating with child processes. On Windows we use mbcs,
which maps to the current code page. On all other platforms we use
UTF-8.
(Executive._should_encode_child_process_arguments): Returns True if
arguments to child processes need to be encoded. This is currently
only needed on Cygwin and Win32 Python 2.x.
(Executive._encode_argument_if_needed): Encode the argument using
child_process_encoding if we need to encode arguments to child
processes on this platform.
* Scripts/webkitpy/common/system/executive_unittest.py:
(never_ending_command): Added. Returns arguments to run a command that
will not quit until we kill it. On Windows we use wmic, on other
platforms we use yes.
(ExecutiveTest.test_run_command_with_unicode): Changed to expect the
mbcs encoding to be used and for output from the child processes to
have been roundtripped through encode/decode on Win32 Python. When
UTF-8 is the encoding the roundtripping is undetectable, but with mbcs
it's possible that some characters will not be able to be converted
and will be replaced by question marks; the round-tripping allows us
to expect this result.
(ExecutiveTest.test_kill_process):
(ExecutiveTest.test_kill_all):
Use never_ending_command instead of invoking "yes" directly. Expect an
exit code of 1 when using Win32 Python, as that's what seems to happen.
2010-11-08 Adam Roben <aroben@apple.com>
Roll out r71532
It broke the build for Cygwin 1.7 installs. Cygwin 1.7's default
.bashrc unsets %TEMP%, which broke copy-tools.cmd.
* Scripts/webkitdirs.pm:
2010-11-08 Tony Chang <tony@chromium.org>
Reviewed by Adam Barth.
run platform/chromium/plugins/nested-plugin-objects.html on all platforms
https://bugs.webkit.org/show_bug.cgi?id=49094
This tests that objects created by plugins are proplery cleaned up.
* DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
(testPassTestObject):
(pluginInvoke):
* DumpRenderTree/TestNetscapePlugIn/TestObject.cpp:
(testAllocate):
(testDeallocate):
(testGetProperty):
(testConstruct):
2010-11-08 Adam Roben <aroben@apple.com>
Mark Windows builds triggered from Perl as being non-interactive
This affects whether some of our scripts will show alerts vs. printing
to the build log.
Fixes <http://webkit.org/b/49181> Windows build fail mysteriously when
.vsprops files are updated
Reviewed by Steve Falkenburg.
* Scripts/webkitdirs.pm:
(buildVisualStudioProject): Set WEBKIT_NONINTERACTIVE_BUILD to 1.
2010-11-08 Gabor Rapcsanyi <rgabor@inf.u-szeged.hu>
Reviewed by Ojan Vafai.
Make http locking default in NRWT.
https://bugs.webkit.org/show_bug.cgi?id=48053
* Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2010-11-08 Gabor Rapcsanyi <rgabor@inf.u-szeged.hu>
Reviewed by Csaba Osztrogonác.
[NRWT] If the http lock fails we shouldn't do any locking
https://bugs.webkit.org/show_bug.cgi?id=49164
If something goes wrong with the locking, the test should keep going.
* Scripts/webkitpy/layout_tests/port/http_lock.py:
2010-11-08 Adam Roben <aroben@apple.com>
Switch back to using kCGImageAlphaPremultipliedFirst when generating
pixel dumps on Windows
I changed this behavior in r71418 thinking that it was required for
getting plugins to show up in pixel dumps. But it doesn't seem to be
necessary, and was making it impossible to compare new Windows pixel
dumps with existing Windows or Mac pixel dumps (because ImageDiff won't
compare an image with alpha to an image without alpha).
Fixes <http://webkit.org/b/49172> REGRESION (r71418): Can't compare
new Windows pixel results to existing Windows or Mac results
Reviewed by Antti Koivisto.
* DumpRenderTree/win/PixelDumpSupportWin.cpp:
(createBitmapContextFromWebView): Replaced kCGImageAlphaNoneSkipFirst
with kCGImageAlphaPremultipliedFirst.
2010-11-08 Csaba Osztrogonac <ossy@webkit.org>
Unreviewed, rolling out r71466.
http://trac.webkit.org/changeset/71466
https://bugs.webkit.org/show_bug.cgi?id=48865
It broke layout tests on GTK bots.
* DumpRenderTree/gtk/DumpRenderTree.cpp:
(createWebView):
2010-11-08 Gabor Rapcsanyi <rgabor@inf.u-szeged.hu>
Reviewed by Csaba Osztrogonác.
Enable running of Qt API tests on BuildBot
https://bugs.webkit.org/show_bug.cgi?id=49004
* BuildSlaveSupport/build.webkit.org-config/master.cfg:
2010-11-08 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
Add clean-review-queue command to remove closed bugs from the webkit.org/pending-review
https://bugs.webkit.org/show_bug.cgi?id=49160
Bugzilla doesn't automatically remove r? when a bug gets closed.
This script takes care of that for webkit.org.
* Scripts/webkitpy/common/net/bugzilla.py:
* Scripts/webkitpy/tool/commands/upload.py:
2010-11-07 Fumitoshi Ukai <ukai@chromium.org>
Unreviewed, rolling out r71474.
http://trac.webkit.org/changeset/71474
https://bugs.webkit.org/show_bug.cgi?id=48280
breaks chromium webkit tests
https://bugs.webkit.org/show_bug.cgi?id=49151
* Scripts/webkitpy/layout_tests/port/base.py:
* Scripts/webkitpy/layout_tests/port/base_unittest.py:
* Scripts/webkitpy/layout_tests/port/config.py:
* Scripts/webkitpy/layout_tests/port/config_unittest.py:
* Scripts/webkitpy/layout_tests/port/google_chrome_unittest.py:
* Scripts/webkitpy/layout_tests/port/webkit.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
* Scripts/webkitpy/tool/mocktool.py:
2010-11-07 Fumitoshi Ukai <ukai@chromium.org>
Unreviewed, rolling out r71475.
http://trac.webkit.org/changeset/71475
breaks chromium webkit tests
https://bugs.webkit.org/show_bug.cgi?id=49151
* Scripts/webkitpy/common/newstringio.py: Removed.
* Scripts/webkitpy/common/newstringio_unittest.py: Removed.
* Scripts/webkitpy/common/system/executive_mock.py: Removed.
* Scripts/webkitpy/common/system/filesystem_mock.py: Removed.
* Scripts/webkitpy/layout_tests/port/config_mock.py: Removed.
2010-11-06 Dirk Pranke <dpranke@chromium.org>
Unreviewed, build fix.
Add files inexplicably not committed in r71474 as part of the
fix for bug 48280.
* Scripts/webkitpy/common/newstringio.py: Added.
* Scripts/webkitpy/common/newstringio_unittest.py: Added.
* Scripts/webkitpy/common/system/executive_mock.py: Added.
* Scripts/webkitpy/common/system/filesystem_mock.py: Added.
* Scripts/webkitpy/layout_tests/port/config_mock.py: Added.
2010-11-06 Dirk Pranke <dpranke@chromium.org>
Reviewed by Eric Siedel.
new-run-webkit-tests: update port/base and port/webkit to use the
new FileSystem and Config abstractions, pulling more logic out of
the base Port classes into separate, mockable objects.
Also create a MockFileSystem object, a MockConfig object, move
MockExecutive into common/system to be next to executive, and
update the config object to use a FileSystem.
https://bugs.webkit.org/show_bug.cgi?id=48280
* Scripts/webkitpy/common/newstringio.py: Added.
* Scripts/webkitpy/common/newstringio_unittest.py: Added.
* Scripts/webkitpy/common/system/executive_mock.py: Added.
* Scripts/webkitpy/common/system/filesystem_mock.py: Added.
* Scripts/webkitpy/layout_tests/port/base.py:
* Scripts/webkitpy/layout_tests/port/base_unittest.py:
* Scripts/webkitpy/layout_tests/port/config.py:
* Scripts/webkitpy/layout_tests/port/config_mock.py:
* Scripts/webkitpy/layout_tests/port/config_unittest.py:
* Scripts/webkitpy/layout_tests/port/google_chrome_unittest.py:
* Scripts/webkitpy/layout_tests/port/webkit.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
* Scripts/webkitpy/tool/mocktool.py:
2010-11-06 Dirk Pranke <dpranke@chromium.org>
Unreviewed, build breakage.
Apparently I uploaded the wrong version of the file to fix 49122
and neither Eric or I noticed - it was missing a dirname() call.
Fixing ...
https://bugs.webkit.org/show_bug.cgi?id=49122
* Scripts/webkitpy/common/checkout/scm.py:
2010-11-06 Dirk Pranke <dpranke@chromium.org>
Reviewed by Eric Seidel.
webkitpy/tool/* unittests change cwd and don't clean up properly
https://bugs.webkit.org/show_bug.cgi?id=49122
* Scripts/webkitpy/common/checkout/scm.py:
2010-11-05 Chris Marrin <cmarrin@apple.com>
Reviewed by Simon Fraser.
suspendAnimations/resumeAnimations not present in WebKit2
https://bugs.webkit.org/show_bug.cgi?id=49109
* WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl:
* WebKitTestRunner/InjectedBundle/LayoutTestController.cpp:
(WTR::LayoutTestController::suspendAnimations):
(WTR::LayoutTestController::resumeAnimations):
* WebKitTestRunner/InjectedBundle/LayoutTestController.h:
2010-11-05 Tony Chang <tony@chromium.org>
Reviewed by David Levin.
cleanup style in TestNetscapePlugIn/PluginObject.cpp
https://bugs.webkit.org/show_bug.cgi?id=49044
* DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
(getPluginClass):
(pluginGetProperty):
(pluginSetProperty):
(variantToIdentifier):
(testCallback):
(getURL):
(getURLNotify):
(testInvokeDefault):
(testGetProperty):
(testDocumentOpen):
(testWindowOpen):
(pluginInvoke):
(handleCallback):
(notifyStream):
* DumpRenderTree/TestNetscapePlugIn/PluginObject.h:
2010-11-05 Eric Seidel <eric@webkit.org>
Reviewed by David Levin.
Add QueueStatusServer/__init__.py so others can run the QueueStatusServer tests
https://bugs.webkit.org/show_bug.cgi?id=49032
I wrote this file as part of bug 47847, but I forgot to commit it.
No one else noticed it missing because test-webkitpy knows how
to recover in the case where it can't import QueueStatusServer
(which generally occurs due to not having installed the AppEngine SDK).
* QueueStatusServer/__init__.py: Added.
* QueueStatusServer/model/workitems_unittest.py:
- Remove a test which fails. This was probably landed (by me)
from my other machine, which since this __init__.py was missing
I never noticed the failure and landed this invalid test.
Sadly we can't really test remove_work_item as it depends
on .key() working. .key() will throw unless the object
has already been saved it seems.
This may be a mis-design in our remove_work_item implementation,
but for now, just removing the test.
2010-11-04 Adam Roben <aroben@apple.com>
Teach check-webkit-style about TestNetscapePlugIn
Fixes <http://webkit.org/b/49030> check-webkit-style is ignorant of
TestNetscapePlugIn's coding conventions
Reviewed by John Sullivan.
* Scripts/webkitpy/style/checker.py: Lump TestNetscapePlugIn in with
WebKitAPITest and TestWebKitAPI in having different include order and
naming conventions than WebCore. Added some comments about why the
exceptions exist.
2010-11-04 Adam Roben <aroben@apple.com>
Add a test that shows that windowed plugins are able to paint
Somehow we never had a test for this before.
Fixes <http://webkit.org/b/49029> Should add a test that shows
windowed plugins can paint
Reviewed by Jon Honeycutt.
* DumpRenderTree/TestNetscapePlugIn/Tests/win/DrawsGradient.cpp: Added.
(DrawsGradient::DrawsGradient):
(DrawsGradient::wndProc): We handle the WM_PAINT and WM_PRINTCLIENT messages.
(DrawsGradient::onPaint):
(DrawsGradient::onPrintClient):
These both just call through to paint.
(DrawsGradient::paint): Fills our client area with some gradients.
* DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj: Link
against Msimg32.lib for ::GradientFill and added DrawsGradient.
* DumpRenderTree/win/PixelDumpSupportWin.cpp:
(createBitmapContextFromWebView): Use WM_PRINT instead of
WM_PRINTCLIENT so that ::DefWindowProc will send
WM_PRINT/WM_PRINTCLIENT messages to the WebView's child windows.
Replaced kCGImageAlphaPremultipledFirst with kCGImageAlphaNoneSkipFirst
because GDI doesn't support alpha and kCGImageBitmapOrder32Little with
kCGImageBitmapOrder32Host because, who knows, maybe someday Windows
will run on a big-endian platform.
2010-11-04 Adam Roben <aroben@apple.com>
Extract much of NPNInvalidateRectInvalidatesWindow's code into a
WindowedPluginTest base class
The base class takes care of subclassing the plugin's window so that a
custom WNDPROC is called. This will make it easier to write tests that
need to handle window messages.
Fixes <http://webkit.org/b/49028> It's hard to write a PluginTest with
a custom WNDPROC
Reviewed by Jon Honeycutt.
* DumpRenderTree/TestNetscapePlugIn/Tests/win/NPNInvalidateRectInvalidatesWindow.cpp:
(NPNInvalidateRectInvalidatesWindow::NPNInvalidateRectInvalidatesWindow):
(NPNInvalidateRectInvalidatesWindow::NPP_SetWindow):
(NPNInvalidateRectInvalidatesWindow::wndProc):
(NPNInvalidateRectInvalidatesWindow::testInvalidateRect):
Moved code from here to WindowedPluginTest. Changed to use window()
instead of m_window.
* DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
Added WindowedPluginTest and added TestNetscapePlugIn/win to the
include path. Also reordered the include path to make a little more
sense and simplified the entry that added TestNetscapePlugIn itself to
the include path.
* DumpRenderTree/TestNetscapePlugIn/win/WindowedPluginTest.cpp: Added.
(WindowedPluginTest::WindowedPluginTest):
(WindowedPluginTest::NPP_SetWindow):
(WindowedPluginTest::staticWndProc):
* DumpRenderTree/TestNetscapePlugIn/win/WindowedPluginTest.h: Added.
(WindowedPluginTest::window):
Code came from NPNInvalidateRectInvalidatesWindow.
2010-11-04 Adam Roben <aroben@apple.com>
Add a plugin test to show that windowed plugins are clipped correctly
Fixes <http://webkit.org/b/49024> <rdar://problem/8487847> Windowed
plugins aren't clipped in WebKit2 on Windows
Reviewed by Jon Honeycutt.
* DumpRenderTree/TestNetscapePlugIn/Tests/win/WindowRegionIsSetToClipRect.cpp: Added.
(WindowRegionIsSetToClipRect::WindowRegionIsSetToClipRect): Initialize members.
(WindowRegionIsSetToClipRect::NPP_SetWindow): Check that our window
region matches the clip rect we know we should have based on
window-region-is-set-to-clip-rect.html, and check that our window class
doesn't have the CS_PARENTDC style.
* DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
Added WindowRegionIsSetToClipRect.
2010-11-05 Alejandro G. Castro <alex@igalia.com>
Reviewed by Martin Robinson.
[GTK] Avoid font hinting in the DRT
https://bugs.webkit.org/show_bug.cgi?id=48548
Change the settings to avoid font hinting, it was causing
different results depending on the distribution.
* DumpRenderTree/gtk/DumpRenderTree.cpp:
(initializeGtkFontSettings):
* DumpRenderTree/gtk/fonts/fonts.conf:
2010-11-05 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
build-webkit spams system.log by repeatedly checking for PBXProductDirectory in com.apple.Xcode
https://bugs.webkit.org/show_bug.cgi?id=49051
This is a speculative fix. The unit tests cover these methods, however
I don't know if this will fully stop the system.log spam.
* Scripts/webkitpy/layout_tests/port/base.py:
* Scripts/webkitpy/layout_tests/port/webkit.py:
2010-11-05 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
Add basic support for showing bot id on /queue-status/ pages
https://bugs.webkit.org/show_bug.cgi?id=49037
This support is really simple. Eventually we'll want to
show the bot id in the lock table too, but we don't have
that information stored in the server yet.
* QueueStatusServer/handlers/queuestatus.py:
* QueueStatusServer/handlers/queuestatus_unittest.py: Added.
* QueueStatusServer/templates/includes/singlequeuestatus.html:
2010-11-04 Dirk Pranke <dpranke@chromium.org>
Reviewed by Adam Barth.
Create a filesystem wrapper that we can use to enforce
particular conventions and use for mocking and dependency
injection down the line.
https://bugs.webkit.org/show_bug.cgi?id=48144
* Scripts/webkitpy/common/system/filesystem.py: Added.
* Scripts/webkitpy/common/system/filesystem_unittest.py: Added.
2010-11-04 Mihai Parparita <mihaip@chromium.org>
Reviewed by Tony Chang.
run_webkit_tests_unittest fails under Python 2.5
https://bugs.webkit.org/show_bug.cgi?id=49043
Switch from itertools.chain.from_iterable (which was added in 2.6)
to using itertools.chain directly.
* Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2010-11-04 Mihai Parparita <mihaip@chromium.org>
Unreviewed fix to webkit-patch rebaseline-server so that it runs under
Python 2.5 (it needs an import to use the with statement).
* Scripts/webkitpy/tool/commands/rebaselineserver.py:
2010-11-04 Dirk Pranke <dpranke@chromium.org>
Reviewed by Adam Barth.
new-run-webkit-tests: split out webkit-specific configuration stuff into a new module
The current NRWT code has webkit-specific configuration code (like
_script_path, default configuration, etc.) mixed in with
layout-test-specific stuff in port/base. The configuration code
should be split out into a separate module for easier mocking,
testing, and isolation.
https://bugs.webkit.org/show_bug.cgi?id=48264
* Scripts/webkitpy/layout_tests/port/config.py: Added.
* Scripts/webkitpy/layout_tests/port/config_unittest.py: Added.
2010-11-04 Mihai Parparita <mihaip@chromium.org>
Reviewed by Tony Chang.
Rebaseline server: initial framework
https://bugs.webkit.org/show_bug.cgi?id=48892
Adds the basic framework for the rebaseline server (details at
http://webkit.org/b/47761). Includes the rebaseline-server webkit-patch
command, which starts an HTTP server that can serve static files or
invoke handler methods on a class.
* Scripts/webkitpy/tool/commands/__init__.py:
* Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html: Added.
* Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css: Added.
* Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js: Added.
* Scripts/webkitpy/tool/commands/rebaselineserver.py: Added.
2010-11-04 Dirk Pranke <dpranke@chromium.org>
Reviewed by James Robinson.
new-run-webkit-tests wasn't using DRT by default for
--platform chromium-gpu
The default value was set to False instead of None, which meant
that the platform specific logic wasn't firing to change the
value to True (b/c we were afraid we'd be overriding the user
preference).
https://bugs.webkit.org/show_bug.cgi?id=49038
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2010-11-04 Tony Chang <tony@chromium.org>
Reviewed by Adam Barth.
make platform/chromium/plugins/multiple-plugins.html pass on all platforms
https://bugs.webkit.org/show_bug.cgi?id=49026
Have the testObject.bar property return the string "bar". This lets
us run and pass platform/chromium/plugins/multiple-plugins.html on all
platforms.
* DumpRenderTree/TestNetscapePlugIn/TestObject.cpp:
(testGetProperty):
2010-11-04 Mihai Parparita <mihaip@chromium.org>
Reviewed by Tony Chang.
NRWT doesn't have good test coverage for --run-chunk, --batch-size, --run-part, etc.
https://bugs.webkit.org/show_bug.cgi?id=48878
Add get_tests_run so that it's easy to see which tests get run (and with
what batching) for a given flag combination. Flesh out the various
test cases that have FIXMEs.
Also fixes an off-by-one error (batch sizes were one larger than
expected) and makes --run-part also have wraparound behavior, like
--run-chunk.
* Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2010-11-04 Erik Arvidsson <arv@chromium.org>
Reviewed by Dimitri Glazkov.
Support box-sizing without the vendor prefix
https://bugs.webkit.org/show_bug.cgi?id=36713
Based on patch by Peter Beverloo <peter@lvp-media.com>
* iExploder/htdocs/cssproperties.in: Change -webkit-box-sizing to box-sizing.
2010-11-04 Csaba Osztrogonác <ossy@webkit.org>
Unreviewed rollout r71340, because it broke Chromium Windows bot.
* Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2010-11-04 Gabor Rapcsanyi <rgabor@inf.u-szeged.hu>
Reviewed by Ojan Vafai.
Make http locking default in NRWT.
https://bugs.webkit.org/show_bug.cgi?id=48053
* Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2010-11-04 Gabor Rapcsanyi <rgabor@inf.u-szeged.hu>
Reviewed by Eric Seidel.
[NRWT] Clear invalid http locks on Windows platform as well
https://bugs.webkit.org/show_bug.cgi?id=48515
* Scripts/webkitpy/common/system/executive.py:
* Scripts/webkitpy/common/system/executive_unittest.py:
* Scripts/webkitpy/layout_tests/port/http_lock.py:
2010-11-02 Adam Roben <aroben@apple.com>
Reduce our dependence on coreutils when running Python tests
This patch introduces versions of the standard echo and cat utilities
implemented in Python. They are probably missing features of their
coreutils equivalents, but they can do what's necessary for our Python
tests. This is useful on Windows, where these utilities typically
aren't available.
Fixes <http://webkit.org/b/48883> executive_unittest relies on echo
and cat utilities from coreutils, which are not present on Windows
Reviewed by Eric Seidel.
* Scripts/webkitpy/common/system/executive_unittest.py: Changed to use
our Python-based echo and cat.
* Scripts/webkitpy/common/system/fileutils.py: Added.
(make_stdout_binary): On Windows, puts sys.stdout into binary mode so
that \n won't be translated into \r\n. I couldn't think of a good way
to test this directly without touching the filesystem, but it is tested
indirectly by echo_unittest.
* Scripts/webkitpy/test/cat.py: Added.
(command_arguments): Returns a list for invoking cat with the given arguments.
(main): Acts like a simplified version of the coreutils cat utility.
* Scripts/webkitpy/test/cat_unittest.py: Added.
(CatTest.assert_cat): Runs cat with the given input and ensures the
output matches the input.
(CatTest.test_basic): Performs a simple test of cat.
(CatTest.test_no_newline): Tests what happens when the input string
doesn't have a trailing newline.
(CatTest.test_unicode): Tests passing a unicode string to cat.
(CatTest.test_as_command): Tests running cat as a separate command.
* Scripts/webkitpy/test/echo.py: Added.
(command_arguments): Returns a list for invoking echo with the given arguments.
(main): Acts like a simplified version of the coreutils echo utility.
* Scripts/webkitpy/test/echo_unittest.py: Added.
(EchoTest.test_basic): Performs a simple test of echo.
(EchoTest.test_no_newline): Tests passing -n to echo to suppress the
trailing newline.
(EchoTest.test_unicode): Tests passing unicode and non-unicode strings
to echo.
(EchoTest.test_argument_order): Tests what happens when -n is not the
first argument.
(EchoTest.test_empty_arguments): Tests what happens when you pass [] to
echo.main.
(EchoTest.test_no_arguments): Tests what happens when you call
echo.main with no arguments.
(EchoTest.test_as_command): Tests running echo as a separate command.
2010-11-04 Renata Hodovan <reni@inf.u-szeged.hu>
Unreviewed: Add myself to the list of Committers.
* Scripts/webkitpy/common/config/committers.py:
2010-11-04 Andreas Kling <kling@webkit.org>
Reviewed by Kenneth Rohde Christiansen.
[Qt] Hook into QWebPage::scrollRequested for resizes-to-contents mode
In RTC mode the QGraphicsWebView item is the size of the contents,
scrolling works a bit differently (we need to react to scrollRequested.)
Normally QGraphicsView will replay the last mouse event when scrolling,
so to prevent WebKit from getting confused by this we temporarily make
the QGraphicsView non-interactive.
* QtTestBrowser/webview.cpp:
(WebViewGraphicsBased::setPage):
(WebViewGraphicsBased::scrollRequested):
* QtTestBrowser/webview.h:
2010-11-04 Tor Arne Vestbø <tor.arne.vestbo@nokia.com>
Reviewed by Andreas Kling.
Use OS(MAC_OS_X) rather than PLATFORM(MAC) in TestControllerQt
PLATFORM(MAC) is not defined for the Qt port, as it refers to
the Mac-port, not the Mac OS X operating system.
* WebKitTestRunner/qt/TestControllerQt.cpp:
(WTR::isExistingLibrary):
2010-11-04 Leonid Ebril <leonid.ebril@nokia.com>
Reviewed by Andreas Kling.
[Qt] Adding iphone user agent string the user agent list for QtTestBrowser
https://bugs.webkit.org/show_bug.cgi?id=48869
* QtTestBrowser/useragentlist.txt:
2010-11-03 Adam Roben <aroben@apple.com>
Always use uppercase drive names in strings returned by abspath_to_uri
Some versions of cygpath use lowercase drive letters while others use
uppercase, which makes it hard to test the output of code that uses
cygpath.
Fixes <http://webkit.org/b/48914> webkitpy.common.system.path_unittest
fails with Cygwin 1.5
Reviewed by Eric Seidel.
* Scripts/webkitpy/common/system/path.py:
(cygpath): Updated the docstring to indicate that only absolute paths
should be passed for now (though relative paths will work fine).
(_Cygpath.convert): Upper-case the first letter of the converted Windows path.
2010-11-03 George Guo <George.Guo@Nokia.com>
Reviewed by Andreas Kling.
[Qt] QtTestBrowser : set mmp rule pageddata in Symbian
https://bugs.webkit.org/show_bug.cgi?id=48767
Paging is needd on Symbian devices to support benchmarks tests like
dromaeo.com and Celtic Kane that need a lot of memory to run
* QtTestBrowser/QtTestBrowser.pro:
2010-11-03 Jenn Braithwaite <jennb@chromium.org>
Reviewed by Dmitry Titov.
Chromium: Update resource tracking when moving a frame between documents
https://bugs.webkit.org/show_bug.cgi?id=48363
* DumpRenderTree/chromium/WebViewHost.cpp:
(WebViewHost::assignIdentifierToRequest):
Always put resource id in map so we can make assumptions about its
presence.
(WebViewHost::removeIdentifierForRequest):
Added.
(WebViewHost::didFinishResourceLoad):
(WebViewHost::didFailResourceLoad):
Check existence of resource id before removing from map.
* DumpRenderTree/chromium/WebViewHost.h:
2010-11-03 Victor Wang <victorw@chromium.org>
Reviewed by Adam Barth.
[Chromium] update buildbot names in chromium rebaseline tool.
https://bugs.webkit.org/show_bug.cgi?id=48881
* Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
2010-11-03 Darin Adler <darin@apple.com>
Updated Xcode projects by opening them with Xcode 3.2.4.
Updated svn:ignore for Xcode projects.
* MiniBrowser/MiniBrowser.xcodeproj: Added property svn:ignore.
* TestWebKitAPI/TestWebKitAPI.xcodeproj: Added property svn:ignore.
* WebKitLauncher/WebKitLauncher.xcodeproj: Modified property svn:ignore.
* WebKitLauncher/WebKitLauncher.xcodeproj/project.pbxproj: Updated with Xcode 3.2.4.
2010-11-03 Adam Roben <aroben@apple.com>
Stop waiting for messages from the web process after a timeout elapses
TestController::runUntil now takes a timeout parameter that specifies
how long to wait for the condition to become true. WebKitTestRunner
takes a --timeout flag that specifies how long the timeout should be.
run-webkit-tests passes this flag to WebKitTestRunner so its timeout
will be similar to run-webkit-tests.
Fixes <http://webkit.org/b/43047> <rdar://problem/8365833>
WebKitTestRunner waits forever if the web process crashes
Reviewed by Darin Adler and Anders Carlsson.
* Scripts/old-run-webkit-tests:
(top level): Moved the GuardMalloc timeout adjustment here from
readFromDumpToolWithTimer.
(openDumpTool): Make WTR use a timeout similar to but slightly shorter
than the one that was specified on the command line.
* WebKitTestRunner/TestController.cpp:
(WTR::TestController::TestController): Initialize our timeout values.
(WTR::TestController::initialize): Parse the --timeout flag and use it
to modify our timeout values.
(WTR::TestController::resetStateToConsistentValues): Changed to use a
short timeout while waiting for the web process to reset and to return
a boolean indicating whether we were able to reset the web process.
Uses a 5-second timeout while waiting for the process to be reset.
(WTR::TestController::runTest): Changed to return a boolean indicating
whether we were able to reset the web process (and thus run the test).
(WTR::TestController::runUntil): Call through to platformRunUntil.
(WTR::TestController::runTestingServerLoop):
(WTR::TestController::run):
Changed to bail if any test can't be run. This will cause the process
to exit. (Unfortunately this will make run-webkit-tests think we
crashed; see <http://webkit.org/b/48943>.)
* WebKitTestRunner/TestController.h: Added platformRunUntil,
m_longTimeout, and m_shortTimeout.
* WebKitTestRunner/TestInvocation.cpp:
(WTR::TestInvocation::invoke): Use a short timeout when waiting for the
initial response and a long timeout when waiting for the test to
complete. Check whether runUntil timed out and print an error message
if so.
* WebKitTestRunner/mac/TestControllerMac.mm:
(WTR::TestController::platformRunUntil): Renamed from runUntil. Pass
[NSDate distantPast] to -[NSRunLoop runMode:beforeDate:] so that we
won't block waiting for the run loop. Only loop until the timeout
elapses.
* WebKitTestRunner/qt/TestControllerQt.cpp:
(WTR::TestController::platformRunUntil): Renamed from runUntil. Added a
FIXME about honoring the timeout.
* WebKitTestRunner/win/TestControllerWin.cpp:
(WTR::TestController::platformRunUntil): Renamed from runUntil. Use
::MsgWaitForMultipleObjectsEx to implement the timeout. Changed to use
::PeekMessageW so that we don't block waiting for messages to become
available.
2010-11-03 Adam Roben <aroben@apple.com>
Add a plugin test that evaluates JS after removing the plugin element
from the document
This test replaces platform/win/plugins/plugin-delayed-destroy.html.
That test was made to prevent a crash very similar to this one, but
unfortunately tested only the mechanism that prevented the crash and
not whether the crash itself was prevented. Since WebKit2 uses a
different mechanism to prevent the crash, the test was failing even
though WebKit2 was not vulnerable to the crash. This new test crashes
if there is no mechanism in place to prevent it and passes in both
WebKit1 and WebKit2.
Fixes <http://webkit.org/b/46711> <rdar://problem/8485903>
platform/win/plugins/plugin-delayed-destroy.html fails in WebKit2
Reviewed by Anders Carlsson.
* DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
* DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
* DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro:
* GNUmakefile.am:
Added new file.
* DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
(pluginDeallocate): Make sure we delete the PluginTest object. This
prevents a leak and also allows us to test the crash.
* DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp:
(PluginTest::executeScript): Made this into a non-static member
function.
(PluginTest::waitUntilDone):
(PluginTest::notifyDone):
Updated for changes to executeScript.
* DumpRenderTree/TestNetscapePlugIn/PluginTest.h: Added executeScript.
* DumpRenderTree/TestNetscapePlugIn/Tests/EvaluateJSAfterRemovingPluginElement.cpp: Added.
(EvaluateJSAfterRemovingPluginElement::EvaluateJSAfterRemovingPluginElement):
Initialize ourselves and tell the test harness to wait.
(EvaluateJSAfterRemovingPluginElement::NPP_DestroyStream): Remove our
plugin element from the document, then execute some JavaScript. If
WebKit does not have appropriate mechanisms in place, we'll be
destroyed inside the first call to executeScript and crash on the
second call.
2010-11-02 Stephen White <senorblanco@chromium.org>
Reviewed by Tony Chang.
[chromium] Fix LayoutTestController UMRs.
https://bugs.webkit.org/show_bug.cgi?id=48872
* DumpRenderTree/chromium/LayoutTestController.cpp:
(LayoutTestController::LayoutTestController):
2010-11-03 Kent Tamura <tkent@chromium.org>
Reviewed by Dimitri Glazkov.
REGRESSION: rebaseline-chromium-webkit-tests uses non-zero tolerance for
image dup detection
https://bugs.webkit.org/show_bug.cgi?id=48744
* Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
- Make a function for option parsing for ease of test
- Set 0 to options.tolerance
* Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests_unittest.py:
- Add a test for this change
2010-11-02 Gyuyoung Kim <gyuyoung.kim@samsung.com>
Unreviewed: Add myself to the list of Committers.
* Scripts/webkitpy/common/config/committers.py:
2010-11-02 Anders Carlsson <andersca@apple.com>
Fix build.
* TestWebKitAPI/PlatformUtilities.cpp:
(TestWebKitAPI::Util::createContextForInjectedBundleTest):
2010-11-02 Sam Weinig <sam@webkit.org>
Reviewed by Anders Carlsson.
Add a way to send startup messages on the context which can be posted when a process launches
<rdar://problem/8617928>
https://bugs.webkit.org/show_bug.cgi?id=48838
* MiniBrowser/mac/WebBundle/WebBundleMain.m:
(WKBundleInitialize):
* TestWebKitAPI/InjectedBundleController.cpp:
(TestWebKitAPI::InjectedBundleController::initialize):
(TestWebKitAPI::InjectedBundleController::didReceiveMessage):
* TestWebKitAPI/InjectedBundleController.h:
* TestWebKitAPI/InjectedBundleMain.cpp:
(WKBundleInitialize):
* TestWebKitAPI/PlatformUtilities.cpp:
(TestWebKitAPI::Util::createContextForInjectedBundleTest):
* WebKitTestRunner/InjectedBundle/InjectedBundleMain.cpp:
(WKBundleInitialize):
Update implementations of WKBundleInitialize to take an initial userData
argument. Change TestWebKitAPI to use the new initial userData to initialize
each test's bundle.
2010-11-02 Benjamin Kalman <kalman@google.com>
Reviewed by Ojan Vafai.
new-run-webkit-tests doesn't strip "LayoutTests/" from prefix, unlike old-run-webkit-tests
https://bugs.webkit.org/show_bug.cgi?id=48794
* Scripts/webkitpy/layout_tests/run_webkit_tests.py: Strip the "LayoutTests/" prefix from test argument paths.
2010-11-02 Adam Roben <aroben@apple.com>
Skip webkitpy.layout_tests.run_webkit_tests_unittest.MainTest on Cygwin
Python 2.5.x
It is known to hang on that version of Python. See
<http://webkit.org/b/48614>.
Reviewed by Adam Barth.
* Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py: Use
skip_if to skip MainTest on Cygwin Python 2.5.x.
* Scripts/webkitpy/test/skip.py: Added.
(skip_if): If the passed-in condition is false, find all the test_*
methods of the given class and replace them with a function that just
logs that we're skipping these tests. This is loosely based on the
unittest.skip_if decorator added in Python 3.1.
(_skipped_method): Returns a function that just logs that the tests are
being skipped. This is used in place of the actual test_* functions.
* Scripts/webkitpy/test/skip_unittest.py: Added.
(SkipTest.setUp): Create a logger for ourselves and give it a handler
that logs to a private stream.
(SkipTest.tearDown): Reset the logger.
(SkipTest.create_fixture_class): Returns a class that we use to test
skip_if. It has a single test_* method, test_foo, that just calls
through to a callback.
(SkipTest.foo_callback): Record that test_foo was called.
(SkipTest.test_skip_if_false): Pass skip_if a False condition and test
that test_foo does get called.
(SkipTest.test_skip_if_true): Pass skip_if a True condition and test
that test_foo does not get called and the appropriate message gets
logged.
2010-11-02 Adam Barth <abarth@webkit.org>
Reviewed by Eric Seidel.
webkit-patch should tell check-webkit-style which files were changed so
check-webkit-style doesn't have to stat the whole working copy again
https://bugs.webkit.org/show_bug.cgi?id=48792
* Scripts/webkitpy/tool/mocktool.py:
* Scripts/webkitpy/tool/steps/checkstyle.py:
2010-11-02 Robert Kroeger <rjkroege@chromium.org>
Reviewed by James Robinson.
[Chromium/DRT] Make EventSendingController honour leapForward for touch events.
https://bugs.webkit.org/show_bug.cgi?id=48777
* DumpRenderTree/chromium/EventSender.cpp:
(EventSender::sendCurrentTouchEvent):
2010-11-02 Adam Roben <aroben@apple.com>
Only track resource identifiers in DRT when dumpResourceLoadCallbacks
is on
This reverts Windows to our pre-r71097 behavior. That patch made us
track all resource identifiers, including the main resource, so the
main resource's URL started appearing in test output instead of
"<unknown>". Arguably having the main resource's URL is better, but all
other platforms print "<unknown>" and we want to match.
Fixes <http://webkit.org/b/48837> <rdar://problem/8620351> REGRESSION
(r71097): Many http tests failing on Windows
Reviewed by Anders Carlsson.
* DumpRenderTree/win/ResourceLoadDelegate.cpp:
(ResourceLoadDelegate::identifierForInitialRequest): Don't add the
identifier to the URL map if we're not supposed to dump resource load
callbacks.
(ResourceLoadDelegate::removeIdentifierForRequest): Always remove the
identifier from the URL map even if we're already "done". There's no
point in keeping out-of-date identifiers around.
2010-11-01 Jenn Braithwaite <jennb@chromium.org>
Reviewed by Adam Roben.
Windows: Update resource tracking when moving a frame between documents
https://bugs.webkit.org/show_bug.cgi?id=48364
* DumpRenderTree/win/DumpRenderTree.cpp:
(createWebViewAndOffscreenWindow):
(main):
Give each WebView its own ResourceLoadDelegate instance in order to
make assertions about resource ids on a particular WebView.
* DumpRenderTree/win/ResourceLoadDelegate.cpp:
(ResourceLoadDelegate::identifierForInitialRequest):
Always add id to the map.
(ResourceLoadDelegate::removeIdentifierForRequest):
Added.
(ResourceLoadDelegate::willSendRequest):
(ResourceLoadDelegate::didReceiveAuthenticationChallenge):
(ResourceLoadDelegate::didReceiveResponse):
(ResourceLoadDelegate::didFinishLoadingFromDataSource):
(ResourceLoadDelegate::didFailLoadingWithError):
(ResourceLoadDelegate::descriptionSuitableForTestResult):
Replace static descriptionSuitableForTestResult with static member function to access identifier map.
* DumpRenderTree/win/ResourceLoadDelegate.h:
(ResourceLoadDelegate::urlMap):
Moved within class so that each WebView has its own id map.
2010-11-01 Ojan Vafai <ojan@chromium.org>
Reviewed by Dimitri Glazkov.
remove debug code from run_webkit_tests.py
https://bugs.webkit.org/show_bug.cgi?id=48800
Remove temporary debug code and make --master-name required
if --test-results-server is set now that all clients set
--master-name.
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2010-11-01 Tony Chang <tony@chromium.org>
Reviewed by Kent Tamura.
add plugin event logging for linux
https://bugs.webkit.org/show_bug.cgi?id=48779
This is taken from Chromium's fork of the layout test plugin:
http://git.chromium.org/gitweb/?p=chromium.git/.git;a=blob;f=webkit/tools/npapi_layout_test_plugin/main.cpp;h=3ebdada2f049b3624756438cff852364f86a2ede;hb=HEAD#l348
* DumpRenderTree/TestNetscapePlugIn/main.cpp:
(handleEventX11):
(NPP_HandleEvent):
2010-11-01 Dirk Pranke <dpranke@chromium.org>
Reviewed by James Robinson.
new-run-webkit-tests: use DRT, child-processes=1 for GPU tests by default
https://bugs.webkit.org/show_bug.cgi?id=48790
* Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
* Scripts/webkitpy/layout_tests/port/chromium_gpu_unittest.py:
2010-11-01 Adam Barth <abarth@webkit.org>
Reviewed by Eric Seidel.
Teach check-webkit-style how to accept a list of files to diff on the
command line
https://bugs.webkit.org/show_bug.cgi?id=48784
In a future patch, webkit-patch will use this option to improve
performance. I'm landing this in two pieces to avoid causing a version
skew problem for the style-bot.
* Scripts/check-webkit-style:
* Scripts/webkitpy/style/optparser.py:
* Scripts/webkitpy/style/optparser_unittest.py:
* Scripts/webkitpy/style_references.py:
2010-11-01 Anders Carlsson <andersca@apple.com>
Reviewed by John Sullivan.
Tear down the related WebProcessProxy when a WebContext is deallocated
https://bugs.webkit.org/show_bug.cgi?id=48769
* TestWebKitAPI/Tests/WebKit2/FailedLoad.cpp:
(TestWebKitAPI::didFailProvisionalLoadWithErrorForFrame):
We don't support empty URLs anymore, update test to expect a null URL instead.
2010-11-01 Søren Gjesse <sgjesse@chromium.org>
Reviewed by Andreas Kling.
Fix warning when compiling the chromium port of DumpRenderShell
with clang.
https://bugs.webkit.org/show_bug.cgi?id=48414
* DumpRenderTree/chromium/TestShell.h:
(TestShell::javaScriptFlagsForLoad):
2010-11-01 Adam Roben <aroben@apple.com>
Fix typo from r71022
* BuildSlaveSupport/build.webkit.org-config/config.json:
2010-11-01 Adam Roben <aroben@apple.com>
Trigger the Windows Release WebKit2 tests when a Release build
finishes, not when a Debug build finishes
Fixes <http://webkit.org/b/48754> Windows Release WebKit2 tests are
triggered at the wrong time
Reviewed by Sam Weinig.
* BuildSlaveSupport/build.webkit.org-config/config.json: Fixed
triggerable name and trigger.
2010-11-01 Mario Sanchez Prada <msanchez@igalia.com>
Unreviewed. Adding my IRC nickname to the list of committers.
* Scripts/webkitpy/common/config/committers.py:
2010-10-31 Robert Hogan <robert@webkit.org>
Reviewed by Antonio Gomes.
[Qt] [Gtk] Plug-ins having upper case in mime type are failing to load
Qt and Gtk are case-sensitive when storing the declared mime-type
of plugins. Since plugin mime-types are lowercased prior to searching
for them in the plugin database, ensure they are loaded with the
mime-type in lower case too.
Change the test netscape plugin to declare its mimetype in sentence
case so that the correct behaviour is enforced.
https://bugs.webkit.org/show_bug.cgi?id=36815
* DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
(NP_GetMIMEDescription):
2010-10-31 Robert Hogan <robert@webkit.org>
Reviewed by Antonio Gomes.
[Qt] Support nodesFromRect in DRT
https://bugs.webkit.org/show_bug.cgi?id=48716
* DumpRenderTree/qt/LayoutTestControllerQt.cpp:
(LayoutTestController::nodesFromRect):
* DumpRenderTree/qt/LayoutTestControllerQt.h:
2010-10-30 Andreas Kling <kling@webkit.org>
Reviewed by Antonio Gomes.
[Qt] QtTestBrowser: Switching view type moves the embedded inspector
https://bugs.webkit.org/show_bug.cgi?id=48705
Reinsert the embedded inspector into the splitter after changing
between QWebView/QGraphicsWebView.
* QtTestBrowser/launcherwindow.cpp:
(LauncherWindow::init):
(LauncherWindow::initializeView):
* QtTestBrowser/webinspector.h:
(WebInspector::WebInspector):
2010-10-28 Antonio Gomes <agomes@rim.com>
Reviewed by Ojan Vafai.
Needs a "LinuxEditingBehavior", perhaps with a better name
https://bugs.webkit.org/show_bug.cgi?id=36627
Adding support to Mac's, GTK+'s, Windows' and Chromium's LayoutTestController class to test the newly introduced Unix editing behavior.
* DumpRenderTree/chromium/LayoutTestController.cpp:
(LayoutTestController::setEditingBehavior):
* DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
(LayoutTestController::setEditingBehavior):
* DumpRenderTree/mac/LayoutTestControllerMac.mm:
(LayoutTestController::setEditingBehavior):
* DumpRenderTree/win/LayoutTestControllerWin.cpp:
(LayoutTestController::setEditingBehavior):
2010-10-29 Andreas Kling <kling@webkit.org>
Reviewed by Kenneth Rohde Christiansen.
[Qt] QtTestBrowser: Add keyboard shortcut to toggle full screen (F11)
https://bugs.webkit.org/show_bug.cgi?id=48695
* QtTestBrowser/launcherwindow.cpp:
(LauncherWindow::createChrome):
2010-10-29 Andreas Kling <kling@webkit.org>
Reviewed by Kenneth Rohde Christiansen.
[Qt] QtTestBrowser: Fix uninitialized read in FpsTimer
https://bugs.webkit.org/show_bug.cgi?id=48675
FpsTimer::m_timer was never initialized and passed to QObject::killTimer()
on startup with in -graphicsbased mode.
* QtTestBrowser/fpstimer.cpp:
(FpsTimer::FpsTimer):
(FpsTimer::numFrames):
(FpsTimer::stop):
(FpsTimer::timerEvent):
* QtTestBrowser/fpstimer.h:
2010-10-29 Dirk Pranke <dpranke@chromium.org>
Reviewed by Tony Chang.
new-run-webkit-tests: change TestResults to be serializable
In preparation for changing new-run-webkit-tests from
multithreaded to multiprocess, we need to make sure the data
going between the threads is easily serialized over a socket.
This change adds serialization/pickling for the TestResults and
TestFailure objects (using cPickle).
The TestFailure objects included a "has_wdiff" flag for Text
results, but the flag wasn't being used, so I've removed it,
simplifying the state to basically a set of enum objects with
associated methods.
https://bugs.webkit.org/show_bug.cgi?id=48616
* Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
* Scripts/webkitpy/layout_tests/layout_package/printing_unittest.py:
* Scripts/webkitpy/layout_tests/layout_package/test_failures.py:
* Scripts/webkitpy/layout_tests/layout_package/test_failures_unittest.py:
* Scripts/webkitpy/layout_tests/layout_package/test_results.py: Added.
* Scripts/webkitpy/layout_tests/layout_package/test_results_unittest.py:Added.
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
* Scripts/webkitpy/layout_tests/test_types/text_diff.py:
2010-10-29 Dirk Pranke <dpranke@chromium.org>
Reviewed by Ojan Vafai.
new-run-webkit-tests: need to provide separate dashboard results for GPU tests
https://bugs.webkit.org/show_bug.cgi?id=48687
Modify the steps to upload the JSON files to the dashboards so
that the GPU tests don't conflict with the regular tests on a
port. We do this by modifying the --builder-name parameter to
append " - GPU", which should cause the app to treat the results
as a completely new builder. This is a little non-obvious, but
keeps us from having to restructure the app.
* Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
* Scripts/webkitpy/layout_tests/port/chromium_gpu_unittest.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2010-10-29 Ojan Vafai <ojan@chromium.org>
Reviewed by Tony Chang.
[chromium] add debug logging to help diagnose flakiness dashboard issues
https://bugs.webkit.org/show_bug.cgi?id=48657
The appengine app thinks it's getting empty files uploaded. Add some logging
to see if new-run-webkit-tests agrees.
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2010-10-29 Dimitri Glazkov <dglazkov@chromium.org>
Reviewed by Ojan Vafai.
[Chromium/DRT] Add master-name flag to new-run-webkit-tests.
https://bugs.webkit.org/show_bug.cgi?id=48649
* BuildSlaveSupport/build.webkit.org-config/master.cfg: Added flag.
2010-10-29 Nicolas Dufresne <nicolas.dufresne@collabora.co.uk>
Reviewed by Martin Robinson.
Enable popup window in GtkLauncher
https://bugs.webkit.org/show_bug.cgi?id=48335
The GtkLauncher application does not support opening new window when
a link with "target=_blank" is clicked or similar call to
window.open(). Instead, GtkLauncher does nothing which breaks
navigation of some websites.
* GtkLauncher/main.c:
(activate_uri_entry_cb):
(update_title):
(link_hover_cb):
(notify_title_cb):
(notify_load_status_cb):
(notify_progress_cb):
(destroy_cb):
(go_back_cb):
(go_forward_cb):
(create_web_view_cb):
(web_view_ready_cb):
(close_web_view_cb):
(create_browser):
(create_statusbar):
(create_toolbar):
(create_window):
(main):
2010-10-29 Adam Roben <aroben@apple.com>
Teach check-webkit-style about WebKit2's idiosyncracies
Fixes <http://webkit.org/b/48638> Style bot complains about a number
of WebKit2 conventions
Reviewed by Anders Carlsson.
* Scripts/webkitpy/style/checker.py: Excluded some rules for various
WebKit2-related files. Also updated the excluded rules for
WebKitAPITest to match the current code.
2010-10-29 Csaba Osztrogonác <ossy@webkit.org>
Reviewed by Adam Roben and David Kilzer.
Fix and cleanup of build systems
https://bugs.webkit.org/show_bug.cgi?id=48342
* Scripts/build-webkit:
- Remove unnecessary ENABLE_SANDBOX option.
- Add ENABLE_FULLSCREEN_API option.
2010-10-28 Adam Roben <aroben@apple.com>
Switch the Windows WebKit2 bot to the Release configuration
We only have one machine testing WebKit2 on Windows right now, and
Debug is just too slow for it to keep up.
Fixes (hopefully!) <http://webkit.org/b/48615> Windows WebKit2 bot is
always way behind
Reviewed by Jon Honeycutt.
* BuildSlaveSupport/build.webkit.org-config/config.json:
2010-10-29 Leandro Gracia Gil <leandrogracia@google.com>
Reviewed by Jeremy Orlow.
Added a second parameter to setMockSpeechInputResult for
the language used in speech input.
https://bugs.webkit.org/show_bug.cgi?id=47089
* DumpRenderTree/LayoutTestController.cpp:
(setMockSpeechInputResultCallback):
* DumpRenderTree/LayoutTestController.h:
* DumpRenderTree/chromium/LayoutTestController.cpp:
(LayoutTestController::setMockSpeechInputResult):
* DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
(LayoutTestController::setMockSpeechInputResult):
* DumpRenderTree/mac/LayoutTestControllerMac.mm:
(LayoutTestController::setMockSpeechInputResult):
* DumpRenderTree/qt/LayoutTestControllerQt.cpp:
(LayoutTestController::setMockSpeechInputResult):
* DumpRenderTree/qt/LayoutTestControllerQt.h:
* DumpRenderTree/win/LayoutTestControllerWin.cpp:
(LayoutTestController::setMockSpeechInputResult):
* DumpRenderTree/wx/LayoutTestControllerWx.cpp:
(LayoutTestController::setMockSpeechInputResult):
(LayoutTestController::pageSizeAndMarginsInPixels):
2010-10-28 Ojan Vafai <ojan@chromium.org>
Reviewed by Tony Chang.
[chromium] add a result-small.json file for the test dashboard
https://bugs.webkit.org/show_bug.cgi?id=48547
Output both a results.json file and a results-small.json file.
The dashboard will load results-small.json by default so it loads faster.
* TestResultServer/model/jsonresults.py:
* TestResultServer/model/jsonresults_unittest.py:
Added a bunch of sys.path hackery. Unforunately, this uses hardcoded
paths. That obviously needs to be fixed, but at least this way it
clearly documents what paths are necessary.
2010-10-28 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
webkit-patch upload calls changed_files more often than it should
https://bugs.webkit.org/show_bug.cgi?id=48567
Passing changed_files around everywhere isn't a very elegant solution
but it's the one we have for the moment. I think keeping an explicit
cache on Checkout (or making StepState() a real class) is a better
long-term option.
Previously bug_id_for_this_commit was calling changed_files and the
result was never getting cached on the state. Now we're explicitly
caching the result on the state and passing that to the bug_id_for_this_commit call.
I looked into building unit tests for this. Doing so would require
using a real Checkout object with a MockSCM and overriding the appropriate
calls on SCM to count how often we're stating the file system.
That's a useful set of tests to build for a separate change.
* Scripts/webkitpy/common/checkout/api.py:
* Scripts/webkitpy/tool/commands/download.py:
* Scripts/webkitpy/tool/commands/upload.py:
* Scripts/webkitpy/tool/mocktool.py:
2010-10-28 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
Make suggest-reviewers slightly faster
https://bugs.webkit.org/show_bug.cgi?id=48562
Add @memoized to one more common call.
* Scripts/webkitpy/common/checkout/api.py:
2010-10-28 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
webkit-patch suggest-reviewers -g 260550a6e30b7bf34f16bdb4a5396cf26264fc1c is still very slow
https://bugs.webkit.org/show_bug.cgi?id=48536
This patch makes it about 40 seconds faster, but it still
takes 1:40. This will require more refinement.
The suggested reviewers list appears to be the same.
I think the next step may be to have it stop the search after
5 reviewers are found. We never want to suggest 30 people.
* Scripts/webkitpy/common/checkout/scm.py:
- Using --remove-empty to theoretically stop lookups past
when a file is removed. I'm not entirely clear that the option
does what it says it does. Example:
git log --pretty=format:%H -5 --remove-empty -- /Projects/WebKit/WebCore/platform/wx/SearchPopupMenuWx.h
returns only one commit
vs.
git log --pretty=format:%H -5 -- /Projects/WebKit/WebCore/platform/wx/SearchPopupMenuWx.h
which returns 5. I was not aware that wx files were ever removed from the repository?
2010-10-28 Kinuko Yasuda <kinuko@chromium.org>
Reviewed by David Levin.
[Chromium] Support FileSystem in chromium DRT
https://bugs.webkit.org/show_bug.cgi?id=47643
* DumpRenderTree/chromium/WebViewHost.cpp:
(WebViewHost::openFileSystem): Added.
* DumpRenderTree/chromium/WebViewHost.h:
(WebViewHost::openFileSystem): Added.
2010-10-28 Sam Weinig <sam@webkit.org>
Reviewed by Anders Carlsson.
WKURLRefs should be allowed to be null
<rdar://problem/8575621>
https://bugs.webkit.org/show_bug.cgi?id=48535
* TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
* TestWebKitAPI/Tests/WebKit2/PageLoadBasic.cpp:
(TestWebKitAPI::didStartProvisionalLoadForFrame):
(TestWebKitAPI::didCommitLoadForFrame):
(TestWebKitAPI::didFinishLoadForFrame):
Test that URLs are null pointers when unset.
2010-10-28 Chris Fleizach <cfleizach@apple.com>
Reviewed by Adele Peterson.
AX: multi select group option does not handle setting of AXSelectedChildren correctly
https://bugs.webkit.org/show_bug.cgi?id=48464
Add support for querying information about selected children to DRT, including:
selectedChildAtIndex
selectedChildrenCount
setSelectedChild
* DumpRenderTree/AccessibilityUIElement.cpp:
(selectedChildAtIndexCallback):
(setSelectedChildCallback):
(selectedChildrenCountCallback):
(AccessibilityUIElement::setSelectedChild):
(AccessibilityUIElement::selectedChildrenCount):
(AccessibilityUIElement::selectedChildAtIndex):
(AccessibilityUIElement::getJSClass):
* DumpRenderTree/AccessibilityUIElement.h:
* DumpRenderTree/mac/AccessibilityUIElementMac.mm:
(AccessibilityUIElement::selectedChildAtIndex):
(AccessibilityUIElement::selectedChildrenCount):
(AccessibilityUIElement::setSelectedChild):
2010-10-27 Sam Weinig <sam@webkit.org>
Reviewed by Anders Carlsson.
Add WebKit2 API for window feature getter/setters
<rdar://problem/8590373>
https://bugs.webkit.org/show_bug.cgi?id=48496
* MiniBrowser/mac/BrowserWindowController.m:
(-[BrowserWindowController awakeFromNib]):
* MiniBrowser/win/BrowserView.cpp:
(BrowserView::create):
* WebKitTestRunner/TestController.cpp:
(WTR::createOtherPage):
(WTR::TestController::initialize):
2010-10-28 Søren Gjesse <sgjesse@chromium.org>
Reviewed by Tony Chang.
Added support for the DumpRenderTree flags --multiple-loads and --js-flags to the Python test runner.
https://bugs.webkit.org/show_bug.cgi?id=48236
* Scripts/webkitpy/layout_tests/port/chromium.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2010-10-28 Sergio Villar Senin <svillar@igalia.com>
Reviewed by Martin Robinson.
[GTK] WebKitWebFrame's load-status is not properly notified to the tests
https://bugs.webkit.org/show_bug.cgi?id=48048
DRT now listens to WebKitWebFrame load-status signals for each
created frame instead of the load-status signal of the
WebKitWebView that only notifies about changes in the main frame.
* DumpRenderTree/gtk/DumpRenderTree.cpp:
(webViewLoadFinished):
(webFrameLoadStatusNotified):
(frameCreatedCallback):
(createWebView):
(main):
2010-10-28 Tony Chang <tony@chromium.org>
Reviewed by Kent Tamura.
[chromium] fix textInputController.{selectedRange,markedRange}
https://bugs.webkit.org/show_bug.cgi?id=48487
* DumpRenderTree/chromium/TextInputController.cpp:
(TextInputController::markedRange): Return arrays of ints, rather than a string
(TextInputController::selectedRange): Ditto.
2010-10-27 Ojan Vafai <ojan@chromium.org>
Reviewed by Tony Chang.
[chromium] add a master-name flag to new-run-webkit-tests
https://bugs.webkit.org/show_bug.cgi?id=48488
The test results server now allows adding a master name to
the uploaded files. This lets us distinguish bots that have
the same name, but are on different masters.
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2010-10-28 Adam Roben <aroben@apple.com>
Don't append a newline to the test output if the frame has no document
element in WebKitTestRunner
Fixes <http://webkit.org/b/48526> Extra trailing newline when running
plugins/document-open.html in WebKitTestRunner
Reviewed by Anders Carlsson.
* WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
(WTR::toJS): Added. Turns a UTF-8 C string into a JSStringRef.
(WTR::hasDocumentElement): Added. Uses the JSC API to figure out
whether the frame has a document element.
(WTR::dumpFrameText): Match DRT by bailing (rather than appending an
empty string and a newline) if the frame has no document element.
2010-10-28 Adam Roben <aroben@apple.com>
Skip npn-invalidate-rect-invalidates-window.html on headless XP
machines
TestNetscapePlugIn never receives a WM_PAINT message on headless XP
machines, so this test times out. Fixing the test is covered by
<http://webkit.org/b/48333>.
* Scripts/old-run-webkit-tests: Skip
npn-invalidate-rect-invalidates-window.html on Windows if accelerated
compositing support is disabled, which likely means we're on a headless
XP machine.
2010-10-28 Kimmo Kinnunen <kimmok@iki.fi>
Adding myself as a committer.
* Scripts/webkitpy/common/config/committers.py:
2010-10-27 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
webkit-patch suggest-reviewers -g 260550a6e30b7bf34f16bdb4a5396cf26264fc1c is very slow
https://bugs.webkit.org/show_bug.cgi?id=48500
This doesn't fix the problem, but it makes things slightly better.
Each git svn find-rev call takes about .25 seconds on my desktop.
This patch uses a new memoized class to avoid those calls when possible.
The real slowness is still git log on some files, like:
git log --pretty=format:%H -5 -- /Projects/WebKit/WebCore/platform/wx/SearchPopupMenuWx.h
I'm not yet sure how to make the pathological git logs better.
* Scripts/webkitpy/common/checkout/scm.py:
* Scripts/webkitpy/common/memoized.py: Added.
* Scripts/webkitpy/common/memoized_unittest.py: Added.
2010-10-27 Eric Seidel <eric@webkit.org>
Reviewed by Ojan Vafai.
EWS bots should not use --quiet when running build-webkit
https://bugs.webkit.org/show_bug.cgi?id=48482
--quiet is only correct when the sub-process does the error reporting.
In the case of _can_build() the parent process is reporting the error.
We'd like the full build log at queues.webkit.org so someone can look
at the log and understand why the EWS is failing to build trunk.
* Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
* Scripts/webkitpy/tool/commands/earlywarningsystem.py:
* Scripts/webkitpy/tool/commands/earlywarningsystem_unittest.py:
* Scripts/webkitpy/tool/mocktool.py:
2010-10-27 Brian Weinstein <bweinstein@apple.com>
More Windows build fixage. Rename a variable that was named string.
* WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
(WTR::hasPrefix):
2010-10-27 Ojan Vafai <ojan@chromium.org>
Reviewed by Dimitri Glazkov.
[chromium] Make the test results server store which master the bot is on
https://bugs.webkit.org/show_bug.cgi?id=48478
The chromium bots recently changed so that there are multiple slaves with
the same name on different masters. Up till now, the test results server
assumed slave names were unique. Adds a master field to the file in order
to distinguish.
Also, for files that currently lack a master or testtype, set them appropriately.
* TestResultServer/handlers/testfilehandler.py:
* TestResultServer/index.yaml:
* TestResultServer/model/jsonresults.py:
* TestResultServer/model/testfile.py:
* TestResultServer/templates/showfilelist.html:
* TestResultServer/templates/uploadform.html:
2010-10-26 Darin Adler <darin@apple.com>
Reviewed by Sam Weinig.
WebKitTestRunner needs to support layoutTestController.dumpBackForwardList
https://bugs.webkit.org/show_bug.cgi?id=42322
rdar://problem/8193631
WebKitTestRunner needs to support layoutTestController.clearBackForwardList
https://bugs.webkit.org/show_bug.cgi?id=42333
rdar://problem/8193643
* WebKitTestRunner/Configurations/InjectedBundle.xcconfig:
Renamed the product to WebKitTestRunnerInjectedBundle to avoid
name conflicts in the build directory.
* WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl:
Added dumpBackForwardList and clearBackForwardList.
* WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
(WTR::InjectedBundle::InjectedBundle): Removed initialization of m_mainPage.
(WTR::InjectedBundle::didCreatePage): Removed unneeded underscore.
(WTR::InjectedBundle::willDestroyPage): Ditto.
(WTR::InjectedBundle::didReceiveMessage): Ditto.
(WTR::InjectedBundle::initialize): Ditto.
(WTR::InjectedBundle::didCreatePage): Changed code to use m_pages
instead of m_mainPage and m_otherPages.
(WTR::InjectedBundle::willDestroyPage): Ditto.
(WTR::InjectedBundle::page): Ditto.
(WTR::InjectedBundle::beginTesting): Ditto.
(WTR::InjectedBundle::done): Ditto.
(WTR::InjectedBundle::closeOtherPages): Ditto.
(WTR::InjectedBundle::dumpBackForwardListsForAllPages): Added.
* WebKitTestRunner/InjectedBundle/InjectedBundle.h: Changed
page function to longer be inline, and pageCount function
to use m_pages. Added dumpBackForwardListsForAllPages, and
removed some underscores. Replaced m_mainPage and m_otherPageas
with m_pages.
* WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
(WTR::adoptWK): Added.
(WTR::hasPrefix): Added.
(WTR::InjectedBundlePage::reset): Added code to set up
m_previousTestBackForwardListItem.
(WTR::InjectedBundlePage::dump): Added code to call
dumpBackForwardListsForAllPages.
(WTR::compareByTargetName): Added.
(WTR::dumpBackForwardListItem): Added.
(WTR::InjectedBundlePage::dumpBackForwardList): Added.
* WebKitTestRunner/InjectedBundle/InjectedBundlePage.h: Added
dumpBackForwardList and m_previousTestBackForwardListItem.
* WebKitTestRunner/InjectedBundle/LayoutTestController.cpp:
(WTR::LayoutTestController::LayoutTestController): Initialize
m_shouldDumpBackForwardListsForAllWindows to false.
(WTR::LayoutTestController::clearBackForwardList): Added.
* WebKitTestRunner/InjectedBundle/LayoutTestController.h:
Added dumpBackForwardList, clearBackForwardList,
shouldDumpBackForwardListsForAllWindows, and
m_shouldDumpBackForwardListsForAllWindows.
* WebKitTestRunner/TestController.cpp:
(WTR::TestController::initialize): Set up
didReceiveSynchronousMessageFromInjectedBundle.
(WTR::TestController::didReceiveSynchronousMessageFromInjectedBundle):
Added.
* WebKitTestRunner/TestController.h: Ditto.
* WebKitTestRunner/TestInvocation.cpp:
(WTR::TestInvocation::didReceiveSynchronousMessageFromInjectedBundle):
Added.
* WebKitTestRunner/TestInvocation.h: Ditto.
* WebKitTestRunner/WebKitTestRunner.xcodeproj/project.pbxproj:
Renamed bundle to WebKitTestRunnerInjectedBundle (see above).
* WebKitTestRunner/mac/TestControllerMac.mm:
(WTR::TestController::initializeInjectedBundlePath): Ditto.
2010-10-27 Chris Rogers <crogers@google.com>
Reviewed by Chris Marrin.
Add ENABLE_WEB_AUDIO feature enable flag (initially disabled) to build-webkit
https://bugs.webkit.org/show_bug.cgi?id=48279
* Scripts/build-webkit:
2010-10-27 Eric Seidel <eric@webkit.org>
Unreviewed.
build-webkit should collect Visual Studio Express logs and display them
https://bugs.webkit.org/show_bug.cgi?id=39199
It turns out my previous patch wasn't actually working on the win-ews
machine. So I've fixed my mistakes from before.
* Scripts/build-webkit:
- Windows VSE builds change the CWD while building. Why? Who knows.
* Scripts/print-vse-failure-logs:
- windows VSE builds don't use Debug/Release as I expected, so I've
moved off of --configuration to --top-level and added the /obj
optimization while I was there.
2010-10-27 Dimitri Glazkov <dglazkov@chromium.org>
Unreviewed, rolling out r70674.
http://trac.webkit.org/changeset/70674
https://bugs.webkit.org/show_bug.cgi?id=48053
Broke Chromium Windows build.
* Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2010-10-25 Tony Chang <tony@chromium.org>
Reviewed by Anders Carlsson.
compile TestNetscapePlugIn on chromium linux
https://bugs.webkit.org/show_bug.cgi?id=48274
* DumpRenderTree/TestNetscapePlugIn/main.cpp:
(NP_Initialize): On Linux, plugin funcs are set in initialize.
(NPP_New): Mark the plugin as windowless.
(NPP_GetValue): Handle mime type values.
(NP_GetMIMEDescription):
(NP_GetValue):
* DumpRenderTree/chromium/TestNetscapePlugIn/ForwardingHeaders/WebKit/npapi.h: define TRUE and FALSE, which are in webkit's npapi.h.
* DumpRenderTree/chromium/TestNetscapePlugIn/ForwardingHeaders/WebKit/npfunctions.h: Pull in npapi.h to get TRUE/FALSE (matches webkit's npfunctions.h)
2010-10-27 Kenneth Rohde Christiansen <kenneth@webkit.org>
Reviewed by Anders Carlsson.
Remove contentSizeChanged callbacks as it is no longer
part of the public UIClient.
Make WKPageContentsSizeChangedCallback be a private API
https://bugs.webkit.org/show_bug.cgi?id=48409
* MiniBrowser/mac/BrowserWindowController.m:
(-[BrowserWindowController awakeFromNib]):
* MiniBrowser/win/BrowserView.cpp:
(BrowserView::create):
* WebKitTestRunner/TestController.cpp:
(WTR::createOtherPage):
(WTR::TestController::initialize):
2010-10-27 Gabor Rapcsanyi <rgabor@inf.u-szeged.hu>
Reviewed by Ojan Vafai.
Make http locking default in NRWT.
https://bugs.webkit.org/show_bug.cgi?id=48053
* Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2010-10-27 Gabor Rapcsanyi <rgabor@inf.u-szeged.hu>
Reviewed by Tony Chang.
[NRWT] Fix http lock on Windows platform
https://bugs.webkit.org/show_bug.cgi?id=48321
* Scripts/webkitpy/layout_tests/port/http_lock.py:
2010-10-27 Satish Sampath <satish@chromium.org>
Unreviewed, rolling out r70665.
http://trac.webkit.org/changeset/70665
https://bugs.webkit.org/show_bug.cgi?id=47089
Need to address Alexey's review comments.
* DumpRenderTree/LayoutTestController.cpp:
(setMockSpeechInputResultCallback):
* DumpRenderTree/LayoutTestController.h:
* DumpRenderTree/chromium/LayoutTestController.cpp:
(LayoutTestController::setMockSpeechInputResult):
* DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
(LayoutTestController::setMockSpeechInputResult):
* DumpRenderTree/mac/LayoutTestControllerMac.mm:
(LayoutTestController::setMockSpeechInputResult):
* DumpRenderTree/qt/LayoutTestControllerQt.cpp:
(LayoutTestController::setMockSpeechInputResult):
* DumpRenderTree/qt/LayoutTestControllerQt.h:
* DumpRenderTree/win/LayoutTestControllerWin.cpp:
(LayoutTestController::setMockSpeechInputResult):
* DumpRenderTree/wx/LayoutTestControllerWx.cpp:
(LayoutTestController::setMockSpeechInputResult):
2010-10-27 Leandro Gracia Gil <leandrogracia@google.com>
Reviewed by Jeremy Orlow.
Added a second parameter to setMockSpeechInputResult for
the language used in speech input.
https://bugs.webkit.org/show_bug.cgi?id=47089
* DumpRenderTree/LayoutTestController.cpp:
(setMockSpeechInputResultCallback):
* DumpRenderTree/LayoutTestController.h:
* DumpRenderTree/chromium/LayoutTestController.cpp:
(LayoutTestController::setMockSpeechInputResult):
* DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
(LayoutTestController::setMockSpeechInputResult):
* DumpRenderTree/mac/LayoutTestControllerMac.mm:
(LayoutTestController::setMockSpeechInputResult):
* DumpRenderTree/qt/LayoutTestControllerQt.cpp:
(LayoutTestController::setMockSpeechInputResult):
* DumpRenderTree/qt/LayoutTestControllerQt.h:
* DumpRenderTree/win/LayoutTestControllerWin.cpp:
(LayoutTestController::setMockSpeechInputResult):
* DumpRenderTree/wx/LayoutTestControllerWx.cpp:
(LayoutTestController::setMockSpeechInputResult):
(LayoutTestController::pageSizeAndMarginsInPixels):
2010-10-27 Adam Roben <aroben@apple.com>
Reset TestNetscapePlugIn's NPP_GetValue pointer when
NullNPPGetValuePointer finishes running
Fixes <http://webkit.org/b/48435> REGRESSION (r70655): Many plugins
tests are failing on Qt
Reviewed by Anders Carlsson.
* DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp:
(PluginTest::NPP_Destroy):
* DumpRenderTree/TestNetscapePlugIn/PluginTest.h:
Added a do-nothing NPP_Destroy implementation.
* DumpRenderTree/TestNetscapePlugIn/Tests/NullNPPGetValuePointer.cpp:
(NullNPPGetValuePointer::NullNPPGetValuePointer): Save the original
NPP_GetValue pointer in m_originalNPPGetValuePointer so we can restore
it later.
(NullNPPGetValuePointer::NPP_Destroy): Added. Restores the original
NPP_GetValue pointer so it can be used in other tests.
* DumpRenderTree/TestNetscapePlugIn/main.cpp:
(NPP_Destroy):
* DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
(webkit_test_plugin_destroy_instance):
Call through to the PluginTest when NPP_Destroy is called.
2010-10-26 Brian Weinstein <bweinstein@apple.com>
Reviewed by Adam Roben.
WebKit2 shouldn't try to send an empty user agent
https://bugs.webkit.org/show_bug.cgi?id=48397
Add a test that when we set our custom user agent to an empty string, we don't send
an empty user agent.
* TestWebKitAPI/Tests/WebKit2/PreventEmptyUserAgent.cpp: Added.
(TestWebKitAPI::didRunJavaScript): Make sure that the result of navigator.userAgent isn't empty.
(TestWebKitAPI::TEST): Set our custom user agent to the empty string, and run navigator.userAgent.
* TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj: Add the new file.
* TestWebKitAPI/win/TestWebKitAPI.vcproj: Ditto.
2010-10-27 Adam Roben <aroben@apple.com>
Check in file I forgot in r70653
* DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
(NP_Initialize): Save the NPPluginFuncs struct the browser passed to us
so we can be naughty and modify it later.
2010-10-27 Adam Roben <aroben@apple.com>
Test that WebKit doesn't crash if the plugin passes 0 for its
NPP_GetValue pointer
Test for <http://webkit.org/b/48433> Crash in
NetscapePlugin::shouldLoadSrcURL when using Shockwave Director 10.3 in
WebKit2 on Windows
Reviewed by Eric Carlson.
* DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
* DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
* DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro:
* GNUmakefile.am:
Added NullNPPGetValuePointer.cpp.
* DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
* DumpRenderTree/TestNetscapePlugIn/PluginObject.h:
Set up a place to store the NPPluginFuncs struct the browser passed to us.
* DumpRenderTree/TestNetscapePlugIn/Tests/NullNPPGetValuePointer.cpp: Added.
(NullNPPGetValuePointer::NullNPPGetValuePointer): Null out the
NPP_GetValue pointer we passed to the browser to simulate a plugin that
doesn't implement NPP_GetValue.
(NullNPPGetValuePointer::NPP_GetValue): Print an error message. If this
function is called, it means that WebKit has changed in a way that
makes this test invalid.
* DumpRenderTree/TestNetscapePlugIn/main.cpp:
(NP_GetEntryPoints): Save the NPPluginFuncs struct the browser passed
to us so we can be naughty and modify it later.
2010-10-27 Ademar de Souza Reis Jr <ademar.reis@openbossa.org>
Reviewed by Andreas Kling.
Remove references to ancient QGVLauncher and QtLauncher
https://bugs.webkit.org/show_bug.cgi?id=48430
QtTestBrowser substitutes both and has checks enabled.
* Scripts/webkitpy/style/checker.py: remove references
* Scripts/webkitpy/style/checker_unittest.py: ditto
2010-10-27 Gabor Rapcsanyi <rgabor@inf.u-szeged.hu>
Reviewed by Ojan Vafai.
[NRWT] Don't use image hash when it's no need in single test mode.
https://bugs.webkit.org/show_bug.cgi?id=48326
* Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
2010-10-27 David Kilzer <ddkilzer@apple.com>
Fix leak of CFMutableDictionaryRef in createXMLStringFromWebArchiveData()
Reviewed by Adam Roben.
Follow-up fix for: <https://bugs.webkit.org/show_bug.cgi?id=48278>
* DumpRenderTree/cf/WebArchiveDumpSupport.cpp:
(createXMLStringFromWebArchiveData): Use RetainPtr<> to fix a
leak introduced in r70613.
2010-10-27 Gabor Rapcsanyi <rgabor@inf.u-szeged.hu>
Reviewed by Csaba Osztrogonác.
[NRWT] Add platform specific baseline search paths for Qt port
https://bugs.webkit.org/show_bug.cgi?id=48428
* Scripts/webkitpy/layout_tests/port/qt.py:
2010-10-27 Adam Roben <aroben@apple.com>
Catch exceptions when checking if we're inside a git working directory
Fixes <http://webkit.org/b/48420> REGRESSION (r70562): test-webkitpy
fails on systems without git installed
Reviewed by Anders Carlsson.
* Scripts/webkitpy/common/net/credentials.py:
(Credentials._credentials_from_git): Put the call to
Git.in_working_directory inside the try/except since it, too, attempts
to execute git and thus will throw on systems that don't have git
installed.
2010-10-27 Nikolas Zimmermann <nzimmermann@rim.com>
Reviewed by David Kilzer.
Convert DumpRenderTree webarchive code to CoreFoundation
https://bugs.webkit.org/show_bug.cgi?id=48278
CFPropertyListCreateWithData and CFPropertyListCreateData are only available in 10.6+.
Replace CFPropertyListCreateWithData by a combination of CFReadStreamCreateWithBytesNoCopy and CFPropertyListCreateFromStream.
Replace CFPropertyListCreateData by CFPropertyListCreateXMLData.
These changes are wrapped in BUILDING_ON_LEOPARD, as the methods are deprecated on 10.6+.
* DumpRenderTree/cf/WebArchiveDumpSupport.cpp:
(createXMLStringFromWebArchiveData):
2010-10-26 Antonio Gomes <agomes@rim.com>
Reviewed by Martin Robinson.
Remove the absolute path used to include DumpRenderTreeSupportGtk.h from LayoutTestController.h
It turns out this is rather unneeded since WebKitTools/GNUMakefile.am has WebKit/gtk/ in its include
path, and then we can just do #include "WebCoreSupport/DumpRenderTreeSupportGtk.h"
* DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
2010-10-26 David Kilzer <ddkilzer@apple.com>
Rename WebArchiveDumpSupport.mm to WebArchiveDumpSupport.cpp
Reviewed by Adam Roben.
Part 4 of 4: <http://webkit.org/b/48278> Convert DumpRenderTree webarchive code to CoreFoundation
* DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
Updated for file moves and renames.
* DumpRenderTree/cf/WebArchiveDumpSupport.cpp: Renamed from DumpRenderTree/mac/WebArchiveDumpSupport.mm.
* DumpRenderTree/cf/WebArchiveDumpSupport.h: Renamed from DumpRenderTree/mac/WebArchiveDumpSupport.h.
2010-10-26 David Kilzer <ddkilzer@apple.com>
Convert WebArchiveDumpSupport.mm from NS objects to CF types
Reviewed by Adam Roben.
Part 3 of 4: <http://webkit.org/b/48278> Convert DumpRenderTree webarchive code to CoreFoundation
In order to share WebArchive code between the Mac and Windows
ports, the code in WebArchiveDumpSupport.mm was converted from
Cocoa to C++ using CoreFoundation (CF) types.
* DumpRenderTree/mac/DumpRenderTree.mm:
(dump): Renamed serializeWebArchiveToXML() to
createXMLStringFromWebArchiveData() and added HardAutorelease()
to prevent leaks.
* DumpRenderTree/mac/WebArchiveDumpSupport.h:
(createXMLStringFromWebArchiveData): Renamed from
serializeWebArchiveToXML(). Changed to use CF types.
(createCFURLResponseFromResponseData): Renamed from
unarchiveNSURLResponseFromResponseData(). Changed to use CF
types for its parameter and return type.
* DumpRenderTree/mac/WebArchiveDumpSupport.mm: Replaced use of
NS objects with CF types. It will be renamed to *.cpp in a
future commit.
(convertMIMEType): Changed to use case-insensitive string
comparisons.
(convertWebResourceDataToString):
(normalizeHTTPResponseHeaderFields):
(normalizeWebResourceURL):
(convertWebResourceResponseToDictionary):
(compareResourceURLs):
(createXMLStringFromWebArchiveData):
* DumpRenderTree/mac/WebArchiveDumpSupportMac.mm:
(createCFURLResponseFromResponseData): Renamed from
unarchiveNSURLResponseFromResponseData(). Changed to use CF
types for its parameter and return type.
2010-10-26 David Kilzer <ddkilzer@apple.com>
Extract use of NSKeyedUnarchiver from WebArchiveDumpSupport.mm
Reviewed by Adam Roben.
Part 2 of 4: <http://webkit.org/b/48278> Convert DumpRenderTree webarchive code to CoreFoundation
There is no equivalent to NSKeyedUnarchiver in CoreFoundation,
so extract it into a platform-specific source file.
* DumpRenderTree/mac/WebArchiveDumpSupport.h:
(unarchiveNSURLResponseFromResponseData): Added declaration.
* DumpRenderTree/mac/WebArchiveDumpSupport.mm:
(convertWebResourceResponseToDictionary): Extracted code to
unarchiveNSURLResponseFromResponseData() in
WebArchiveDumpSupportMac.mm. Updated logic to return early if
nil is returned from unarchiveNSURLResponseFromResponseData().
* DumpRenderTree/mac/WebArchiveDumpSupportMac.mm:
(unarchiveNSURLResponseFromResponseData): Added. Extracted
code from convertWebResourceResponseToDictionary() in
WebArchiveDumpSupport.mm.
2010-10-26 David Kilzer <ddkilzer@apple.com>
Extract call to -[WebHTMLRepresentation supportedNonImageMIMETypes] from WebArchiveDumpSupport.mm
Reviewed by Adam Roben.
Part 1 of 4: <http://webkit.org/b/48278> Convert DumpRenderTree webarchive code to CoreFoundation
The call to -[WebHTMLRepresentation supportedNonImageMIMETypes]
is not cross-platform between Mac and Windows, so extract it
into a platform-specific source file.
* DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj: Added
WebArchiveDumpSupportMac.mm to the project.
* DumpRenderTree/mac/WebArchiveDumpSupport.h:
(supportedNonImageMIMETypes): Added declaration.
* DumpRenderTree/mac/WebArchiveDumpSupport.mm:
(convertWebResourceDataToString): Replaced call to
-[WebHTMLRepresentation supportedNonImageMIMETypes] with
supportedNonImageMIMETypes().
* DumpRenderTree/mac/WebArchiveDumpSupportMac.mm: Added.
(supportedNonImageMIMETypes): Added. Extracted from
WebArchiveDumpSupport.mm.
2010-10-26 Antonio Gomes <agomes@rim.com>
Reviewed by Martin Robinson.
[GTK] Implement DumpRenderTreeSupportGtk (similarly to DumpRenderTreeSupportQt idea)
https://bugs.webkit.org/show_bug.cgi?id=48199
Implements support to WebKitTabToLinksPreferenceKey through LayoutTestController::overridePreference.
The corresponding DumpRenderTreeSupportGtk method is called in the DRT context only.
* DumpRenderTree/gtk/DumpRenderTree.cpp:
(resetDefaultsToConsistentValues):
(createWebView):
* DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
(LayoutTestController::overridePreference):
2010-10-26 Ariya Hidayat <ariya@sencha.com>
Unreviewed, change the order of my emails for bugzilla autocompletion.
* Scripts/webkitpy/common/config/committers.py:
2010-10-26 Eric Seidel <eric@webkit.org>
Reviewed by David Kilzer.
build-webkit should collect Visual Studio Express logs and display them
https://bugs.webkit.org/show_bug.cgi?id=39199
* Scripts/build-webkit:
* Scripts/print-vse-failure-logs: Added.
* Scripts/webkitdirs.pm:
2010-10-26 David Kilzer <ddkilzer@apple.com>
<http://webkit.org/b/48224> build-webkit: add support for --meter-tag switch
Reviewed by Csaba Osztrogonác.
* Scripts/build-webkit: Added support for --meter-tag switch.
It should have been added with r60820.
2010-10-26 Ademar de Souza Reis Jr. <ademar.reis@openbossa.org>
Reviewed by Dumitru Daniliuc.
check-webkit-style fails on operator+=, operator-=, ... methods
https://bugs.webkit.org/show_bug.cgi?id=48258
* Scripts/webkitpy/style/checkers/cpp.py: Added exceptions
* Scripts/webkitpy/style/checkers/cpp_unittest.py: Added unit-tests
2010-10-26 Ojan Vafai <ojan@chromium.org>
Reviewed by Tony Chang.
remove DEFER support from new-run-webkit-tests
https://bugs.webkit.org/show_bug.cgi?id=48387
DEFER was needed when we were trying to ship Chrome beta.
Now it's just extra complication.
* Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
* Scripts/webkitpy/layout_tests/layout_package/test_expectations.py:
* Scripts/webkitpy/layout_tests/layout_package/test_expectations_unittest.py:
* Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
* Scripts/webkitpy/style/checkers/test_expectations_unittest.py:
2010-10-26 Eric Seidel <eric@webkit.org>
Unreviewed. Adding import with_statement to fix python 2.5.
Teach webkit-patch how to read credentials from the environment
https://bugs.webkit.org/show_bug.cgi?id=48275
* Scripts/webkitpy/common/net/credentials_unittest.py:
- import with_statement to fix python 2.5.
2010-10-26 Alexey Proskuryakov <ap@apple.com>
Reviewed by Sam Weinig.
https://bugs.webkit.org/show_bug.cgi?id=48375
<rdar://problem/8392724> Need delegate calls in PageLoaderClient to indicate if we have
loaded insecure content
Updated for WebKit2 changes.
* MiniBrowser/mac/BrowserWindowController.m:
(didDisplayInsecureContentForFrame):
(didRunInsecureContentForFrame):
(-[BrowserWindowController awakeFromNib]):
* WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
(WTR::InjectedBundlePage::InjectedBundlePage):
(WTR::InjectedBundlePage::didDisplayInsecureContentForFrame):
(WTR::InjectedBundlePage::didRunInsecureContentForFrame):
* WebKitTestRunner/InjectedBundle/InjectedBundlePage.h:
* WebKitTestRunner/TestController.cpp:
(WTR::TestController::initialize):
2010-10-26 Adam Roben <aroben@apple.com>
Clear up confusion between "3D rendering" and "accelerated compositing"
WebKit has no software-based 3D rendering implementation, so 3D
rendering can only be enabled when accelerated compositing is. But DRT
was falsely reporting that 3D rendering was available on machines that
don't support accelerated compositing, leading to much confusion.
Reviewed by Darin Adler.
Fixes <http://webkit.org/b/48370> REGRESSION (r70540): Many
transforms/3d tests are failing on the XP bots
* DumpRenderTree/win/DumpRenderTree.cpp:
(main): When ENABLE(3D_RENDERING) is turned on, only report that 3D
rendering is available when acclerated compositing is also available.
* Scripts/old-run-webkit-tests: Skip tests that use the
-webkit-transform-3d media query when 3D rendering is disabled,
rather than when accelerated compositing is disabled.
2010-10-26 Eric Seidel <eric@webkit.org>
Reviewed by Ojan Vafai.
Remove rietveld code now that it's unused
https://bugs.webkit.org/show_bug.cgi?id=48359
Was cool that we added this support, but now that it's
not used it makes little sense to keep it around.
We can always restore this code from SVN if we need it.
* Scripts/webkitpy/common/config/__init__.py:
* Scripts/webkitpy/common/net/bugzilla.py:
* Scripts/webkitpy/common/net/bugzilla_unittest.py:
* Scripts/webkitpy/common/net/rietveld.py: Removed.
* Scripts/webkitpy/common/net/rietveld_unittest.py: Removed.
* Scripts/webkitpy/thirdparty/__init__.py:
* Scripts/webkitpy/tool/commands/download.py:
* Scripts/webkitpy/tool/commands/download_unittest.py:
* Scripts/webkitpy/tool/commands/queues.py:
* Scripts/webkitpy/tool/commands/queues_unittest.py:
* Scripts/webkitpy/tool/main.py:
* Scripts/webkitpy/tool/mocktool.py:
* Scripts/webkitpy/tool/steps/__init__.py:
* Scripts/webkitpy/tool/steps/postcodereview.py: Removed.
2010-10-26 Anders Carlsson <andersca@apple.com>
Reviewed by Sam Weinig.
pageDidScroll callback should be on the UI process client rather than (or in addition to) the web process client
https://bugs.webkit.org/show_bug.cgi?id=48366
<rdar://problem/8595202>
* MiniBrowser/mac/BrowserWindowController.m:
(-[BrowserWindowController awakeFromNib]):
* WebKitTestRunner/TestController.cpp:
(WTR::createOtherPage):
(WTR::TestController::initialize):
2010-10-26 Eric Seidel <eric@webkit.org>
Reviewed by Tony Chang.
Teach webkit-patch how to read credentials from the environment
https://bugs.webkit.org/show_bug.cgi?id=48275
This makes it possible for svn users to have their bugzilla credentials
stored in their environment instead of typing them every time.
We need this for making it easy to run the win-ews bot (which currently
uses svn instead of git).
* Scripts/webkitpy/common/net/credentials.py:
* Scripts/webkitpy/common/net/credentials_unittest.py:
2010-10-26 Kenichi Ishibashi <bashi@google.com>
Reviewed by Kent Tamura.
Input Method inserts conversion candidates unexpectedly
https://bugs.webkit.org/show_bug.cgi?id=46868
Adds setComposition() to TextInputController to make DRT emulate
an input method behavior.
* DumpRenderTree/chromium/TextInputController.cpp:
(TextInputController::TextInputController):
(TextInputController::setComposition): Added.
* DumpRenderTree/chromium/TextInputController.h:
2010-10-26 Simon Fraser <simon.fraser@apple.com>
Reviewed by Sam Weinig.
Support layoutTestController.layerTreeAsText in WebKitTestRunner
https://bugs.webkit.org/show_bug.cgi?id=42145
Implement layerTreeAsText() in WebKitTestRunner.
* WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl:
* WebKitTestRunner/InjectedBundle/LayoutTestController.cpp:
(WTR::LayoutTestController::layerTreeAsText):
* WebKitTestRunner/InjectedBundle/LayoutTestController.h:
2010-10-26 Adam Barth <abarth@webkit.org>
Reviewed by Eric Seidel.
webkit-patch rollout produces incorrect patch when using svn move
https://bugs.webkit.org/show_bug.cgi?id=48244
We need to flush our caches when we modify the working copy.
* Scripts/webkitpy/tool/steps/revertrevision.py:
2010-10-26 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
commit-queue flaky test message can list the same author more than once
https://bugs.webkit.org/show_bug.cgi?id=48268
tonikitoo reported to me over IRC this morning that he's seen
the commit-queue report flaky tests with author lists like:
"adam, adam and adam", suggesting we're not uniquing authors
before writing the message.
I fixed the uniquing and added a bunch more unit testing.
* Scripts/webkitpy/tool/commands/queues.py:
* Scripts/webkitpy/tool/commands/queues_unittest.py:
* Scripts/webkitpy/tool/mocktool.py:
2010-10-26 Adam Roben <aroben@apple.com>
Pull in the FeatureDefines*.vsprops files when building DRT
This ensures that various ENABLE() macros will get set correctly.
(Prior to r70320 we were picking up the ENABLE(3D_RENDERING) flag
through wtf/Platform.h.)
Fixes <http://webkit.org/b/48343> REGRESSION (r70320): DumpRenderTree
says 3D_RENDERING is disabled even when it is enabled
Reviewed by Ada Chan.
* DumpRenderTree/win/DumpRenderTree.vcproj: Added
FeaturesDefines*.vsprops to all configurations.
* DumpRenderTree/win/DumpRenderTreeWin.h: Touched to force a rebuild.
2010-10-26 Søren Gjesse <sgjesse@chromium.org>
Reviewed by Dimitri Glazkov.
Added options --multiple-loads and --js-flags to chromium DumpRenderTree. The option --multiple-loads=X
is used to have DumpRenderTree load each test it runs X times. To be able to have more fine-grained control
of how the JavaScript engine behaves for each load the flag --js-flags can specify a list of flag-sets like this
--js-flags="--xxx,--noxxx --yyy,--noyyy"
First load will run with --xxx, the second with --yyy and the third without any (the 'no' prefix is handled by
V8 to turn off the flag).
The changes to the Python test runner will be in a separate change.
* DumpRenderTree/chromium/DumpRenderTree.cpp:
(runTest):
(main):
* DumpRenderTree/chromium/TestShell.cpp:
(TestShell::TestShell):
(TestShell::runFileTest):
(TestShell::testFinished):
* DumpRenderTree/chromium/TestShell.h:
(TestShell::loadCount):
(TestShell::setLoadCount):
(TestShell::javaScriptFlagsForLoad):
(TestShell::setJavaScriptFlags):
(TestShell::setDumpWhenFinished):
2010-10-26 Adam Roben <aroben@apple.com>
Skip more tests that depend on accelerated compositing when accelerated
compositing is disabled
Fixes <http://webkit.org/b/48329> Some tests fail when accelerated
compositing is disabled
Reviewed by John Sullivan.
* Scripts/old-run-webkit-tests: Added more tests to skip when
accelerated compositing is disable. The tests all have output that
changes depending on whether accelerated compositing is enabled.
2010-10-25 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
webkit-patch should clean up svn locks when passed --force-clean
https://bugs.webkit.org/show_bug.cgi?id=48269
* Scripts/webkitpy/common/checkout/scm.py:
2010-10-25 Anders Carlsson <andersca@apple.com>
Reviewed by Adam Roben.
Add a pageDidScroll BundleUIClient callback
https://bugs.webkit.org/show_bug.cgi?id=48260
<rdar://problem/8531159>
* WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
(WTR::InjectedBundlePage::InjectedBundlePage):
2010-10-25 Johnny Ding <jnd@chromium.org>
Reviewed by Tony Chang.
Dump the gesture status of frame in frame load callbacks in DumpRenderTree
by adding a new method dumpUserGestureInFrameLoadCallbacks.
Now only dump the gesture status in "DidStartProvisionalLoad" callback.
https://bugs.webkit.org/show_bug.cgi?id=47849
* DumpRenderTree/LayoutTestController.cpp:
(LayoutTestController::LayoutTestController):
(dumpUserGestureInFrameLoadCallbacksCallback):
(LayoutTestController::staticFunctions):
* DumpRenderTree/LayoutTestController.h:
(LayoutTestController::dumpUserGestureInFrameLoadCallbacks):
(LayoutTestController::setDumpUserGestureInFrameLoadCallbacks):
* DumpRenderTree/chromium/LayoutTestController.cpp:
(LayoutTestController::LayoutTestController):
(LayoutTestController::dumpUserGestureInFrameLoadCallbacks):
(LayoutTestController::reset):
* DumpRenderTree/chromium/LayoutTestController.h:
(LayoutTestController::shouldDumpUserGestureInFrameLoadCallbacks):
(LayoutTestController::setShouldDumpUserGestureInFrameLoadCallbacks):
* DumpRenderTree/chromium/TestShell.h:
(TestShell::shouldDumpUserGestureInFrameLoadCallbacks):
* DumpRenderTree/chromium/WebViewHost.cpp:
(WebViewHost::didStartProvisionalLoad):
(WebViewHost::printFrameUserGestureStatus):
* DumpRenderTree/chromium/WebViewHost.h:
* DumpRenderTree/mac/FrameLoadDelegate.mm:
(-[WebFrame _drt_printFrameUserGestureStatus]):
(-[FrameLoadDelegate webView:didStartProvisionalLoadForFrame:]):
(-[FrameLoadDelegate webView:didCommitLoadForFrame:]):
(-[FrameLoadDelegate webView:didFailProvisionalLoadWithError:forFrame:]):
(-[FrameLoadDelegate webView:didFinishLoadForFrame:]):
(-[FrameLoadDelegate webView:didFailLoadWithError:forFrame:]):
(-[FrameLoadDelegate webView:windowScriptObjectAvailable:]):
(-[FrameLoadDelegate webView:didReceiveTitle:forFrame:]):
* DumpRenderTree/qt/LayoutTestControllerQt.cpp:
(LayoutTestController::reset):
(LayoutTestController::dumpUserGestureInFrameLoadCallbacks):
* DumpRenderTree/qt/LayoutTestControllerQt.h:
2010-10-25 Sheriff Bot <webkit.review.bot@gmail.com>
Unreviewed, rolling out r70442.
http://trac.webkit.org/changeset/70442
https://bugs.webkit.org/show_bug.cgi?id=48248
http locking doesn't work on Windows (Requested by Ossy on
#webkit).
* Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2010-10-25 Gabor Rapcsanyi <rgabor@inf.u-szeged.hu>
Reviewed by Ojan Vafai.
Make http locking default in NRWT.
https://bugs.webkit.org/show_bug.cgi?id=48053
* Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2010-10-25 David Kilzer <ddkilzer@apple.com>
<http://webkit.org/b/48185> build-webkit: add support for --progress-tag switch
(Re-landing r70413 after Qt minimal buildfix: r70440.)
Reviewed by Kenneth Rohde Christiansen.
* Scripts/build-webkit: Added support for --progress-tag switch.
It should have been added with r57051.
2010-10-24 Adam Barth <abarth@webkit.org>
Reviewed by David Kilzer.
webkit-patch upload fails when the patch removes a file
https://bugs.webkit.org/show_bug.cgi?id=48187
We need to use "--" to separate file names from the rest of the
command.
* Scripts/webkitpy/common/checkout/scm.py:
2010-10-24 Sheriff Bot <webkit.review.bot@gmail.com>
Unreviewed, rolling out r70413.
http://trac.webkit.org/changeset/70413
https://bugs.webkit.org/show_bug.cgi?id=48210
It broke Qt minimal build (Requested by Ossy on #webkit).
* Scripts/build-webkit:
2010-10-24 Yi Shen <yi.4.shen@nokia.com>
Reviewed by Andreas Kling.
[Qt][QtTestBrowser] Toggle use of QGraphicsView messes up the menu
https://bugs.webkit.org/show_bug.cgi?id=48141
* QtTestBrowser/launcherwindow.cpp:
(LauncherWindow::init):
(LauncherWindow::initializeView):
(LauncherWindow::toggleWebView):
2010-10-24 Daniel Bates <dbates@rim.com>
Reviewed by David Kilzer.
Fix Perl uninitialized warnings in VCSUtils::svnStatus()
and VCSUtils::removeEOL().
https://bugs.webkit.org/show_bug.cgi?id=48196
VCSUtils::svnStatus() concatenates the output of svn status with
a new line character and svn status may return no output (say for
a file that has not been added, deleted, or modified). We should
only concatenate the output of svn status if there is some.
Also, VCSUtils::removeEOL() should ensure that its argument
is initialized before performing a string substitution.
* Scripts/VCSUtils.pm:
- Modified removeEOL() to return "" if its argument is undefined.
- Exported removeEOL() so that it can be tested.
* Scripts/webkitperl/VCSUtils_unittest/removeEOL.pl: Added.
2010-10-24 David Kilzer <ddkilzer@apple.com>
<http://webkit.org/b/48185> build-webkit: add support for --progress-tag switch
Reviewed by Kenneth Rohde Christiansen.
* Scripts/build-webkit: Added support for --progress-tag switch.
It should have been added with r57051.
2010-10-24 David Kilzer <ddkilzer@apple.com>
<http://webkit.org/b/48184> build-webkit: reformat support variables for better maintainability
Reviewed by Daniel Bates.
* Scripts/build-webkit: Alphabetized the @features array based
on the name of the feature. Reformatted the list of support
variables so that they match the order of @features, and so that
each has its own line. This makes it easy to add new variables
in the correct order.
2010-10-23 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
EWS never removes invalid patch ids
https://bugs.webkit.org/show_bug.cgi?id=48173
This is just sticking another finger in the dam.
However this adds more unit testing which will help
us make sure we're always releasing patches once we
redesign the release_patch API and call these from
a more central place.
* Scripts/webkitpy/tool/commands/queues.py:
* Scripts/webkitpy/tool/commands/queues_unittest.py:
* Scripts/webkitpy/tool/mocktool.py:
- Added the ability to request invalid patches.
Log a warning message to make sure we don't ever have
tests use invalid patch fetches by mistake.
2010-10-23 Dan Bernstein <mitz@apple.com>
Build fix. Add stub implementations for required NSDraggingInfo methods.
* DumpRenderTree/mac/DumpRenderTreeDraggingInfo.mm:
(-[DumpRenderTreeDraggingInfo draggingFormation]):
(-[DumpRenderTreeDraggingInfo setDraggingFormation:]):
(-[DumpRenderTreeDraggingInfo animatesToDestination]):
(-[DumpRenderTreeDraggingInfo setAnimatesToDestination:]):
(-[DumpRenderTreeDraggingInfo numberOfValidItemsForDrop]):
(-[DumpRenderTreeDraggingInfo setNumberOfValidItemsForDrop:]):
(-[DumpRenderTreeDraggingInfo enumerateDraggingItemsWithOptions:forView:classes:searchOptions:usingBlock:]):
2010-10-23 David Kilzer <ddkilzer@apple.com>
<http://webkit.org/b/48186> Remove unneeded WebHTMLRepresentationInternal.h header
Reviewed by Sam Weinig.
The only method defined in WebHTMLRepresentationInternal.h is
also defined in WebHTMLRepresentation.h, so use that instead.
* DumpRenderTree/mac/InternalHeaders/WebKit/WebHTMLRepresentationInternal.h: Removed.
* DumpRenderTree/mac/WebArchiveDumpSupport.mm: Switched to use
WebHTMLRepresentation.h.
2010-10-23 Sheriff Bot <webkit.review.bot@gmail.com>
Unreviewed, rolling out r70367.
http://trac.webkit.org/changeset/70367
https://bugs.webkit.org/show_bug.cgi?id=48176
It made 8-10 tests crash on Qt bot (Requested by Ossy on
#webkit).
* DumpRenderTree/LayoutTestController.cpp:
(setCustomPolicyDelegateCallback):
* DumpRenderTree/LayoutTestController.h:
* DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
(LayoutTestController::setCustomPolicyDelegate):
* DumpRenderTree/mac/LayoutTestControllerMac.mm:
(LayoutTestController::setCustomPolicyDelegate):
* DumpRenderTree/mac/PolicyDelegate.h:
* DumpRenderTree/mac/PolicyDelegate.mm:
(-[PolicyDelegate webView:decidePolicyForNavigationAction:request:frame:decisionListener:]):
(-[PolicyDelegate setPermissive:]):
(-[PolicyDelegate setControllerToNotifyDone:]):
* DumpRenderTree/win/LayoutTestControllerWin.cpp:
(LayoutTestController::setCustomPolicyDelegate):
* DumpRenderTree/wx/LayoutTestControllerWx.cpp:
(LayoutTestController::setCustomPolicyDelegate):
2010-10-22 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
EWS never releases patches which fail to apply
https://bugs.webkit.org/show_bug.cgi?id=48171
* Scripts/webkitpy/tool/commands/earlywarningsystem_unittest.py:
* Scripts/webkitpy/tool/commands/queues.py:
2010-10-22 Eric Seidel <eric@webkit.org>
Unreviewed.
Make the EWS queues restart themselves more often.
This matches the commit-queue.
* EWSTools/start-queue.sh:
2010-10-22 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
commit-queue should not pass --quiet to subcommands now that the parent command does the reporting
https://bugs.webkit.org/show_bug.cgi?id=48165
* Scripts/webkitpy/tool/bot/commitqueuetask.py:
* Scripts/webkitpy/tool/bot/commitqueuetask_unittest.py:
* Scripts/webkitpy/tool/commands/queues_unittest.py:
2010-10-22 Mario Sanchez Prada <msanchez@igalia.com>
Adding myself as a committer.
* Scripts/webkitpy/common/config/committers.py:
2010-10-20 Anders Carlsson <andersca@apple.com>
Reviewed by Adam Barth and Darin Adler.
Using the decidePolicyForMIMEType delegate message in an asynchronous manner does not work
https://bugs.webkit.org/show_bug.cgi?id=48014
<rdar://problem/8202716>
* DumpRenderTree/LayoutTestController.cpp:
(setCustomPolicyDelegateCallback):
Add callIgnoreInDecidePolicyForMIMETypeAfterOneSecond argument.
* DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
(LayoutTestController::setCustomPolicyDelegate):
Add callIgnoreInDecidePolicyForMIMETypeAfterOneSecond argument.
* DumpRenderTree/mac/LayoutTestControllerMac.mm:
(LayoutTestController::setCustomPolicyDelegate):
Add callIgnoreInDecidePolicyForMIMETypeAfterOneSecond argument.
* DumpRenderTree/mac/PolicyDelegate.h:
* DumpRenderTree/mac/PolicyDelegate.mm:
(-[PolicyDelegate webView:decidePolicyForMIMEType:request:frame:decisionListener:]):
If _callIgnoreInDecidePolicyForMIMETypeAfterOneSecond is true, call [listener ignore] after one second.
(-[PolicyDelegate setCallIgnoreInDecidePolicyForMIMETypeAfterOneSecond:]):
Update _callIgnoreInDecidePolicyForMIMETypeAfterOneSecond.
* DumpRenderTree/win/LayoutTestControllerWin.cpp:
(LayoutTestController::setCustomPolicyDelegate):
Add callIgnoreInDecidePolicyForMIMETypeAfterOneSecond argument.
* DumpRenderTree/wx/LayoutTestControllerWx.cpp:
(LayoutTestController::setCustomPolicyDelegate):
Add callIgnoreInDecidePolicyForMIMETypeAfterOneSecond argument.
2010-10-22 Sam Weinig <sam@webkit.org>
Reviewed by Anders Carlsson.
Rename the InjectedBundle for TestWebKitAPI to InjectedBundleTestWebKitAPI.bundle
* TestWebKitAPI/Configurations/InjectedBundle.xcconfig:
* TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
* TestWebKitAPI/mac/PlatformUtilitiesMac.mm:
(TestWebKitAPI::Util::createInjectedBundlePath):
2010-10-22 Dirk Pranke <dpranke@chromium.org>
Reviewed by Eric Seidel.
test-webkitpy fails when passed -v (or any other option)
Fix port/base_unittest to pass an explicit list of arguments
rather than accidentally picking up sys.argv.
https://bugs.webkit.org/show_bug.cgi?id=48071
* Scripts/webkitpy/layout_tests/port/base_unittest.py:
2010-10-22 Brian Weinstein <bweinstein@apple.com>
Windows build fix. Update the createNewPage callback to account for the new
arguments.
* MiniBrowser/win/BrowserView.cpp:
(createNewPage):
2010-10-22 Sam Weinig <sam@webkit.org>
Reviewed by Anders Carlsson.
WebKit2 needs to pass the current event modifier flags when requesting a new window
https://bugs.webkit.org/show_bug.cgi?id=48140
* MiniBrowser/mac/BrowserWindowController.m:
(createNewPage):
* WebKitTestRunner/TestController.cpp:
(WTR::createOtherPage):
Update for new signature for the WKPageCreateNewPageCallback.
2010-10-22 Eric Seidel <eric@webkit.org>
Unreviewed. Just adding a test case.
The style-queue was not recognizing new patches. It turns
out I had fixed this in an earlier change, but just not
deployed to queues.webkit.org. As part of investigating why it
was broken, I wrote a test for my previous change which I'm now landing.
* QueueStatusServer/model/queues_unittest.py:
2010-10-22 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
EWS does not need to process obsolete patches
https://bugs.webkit.org/show_bug.cgi?id=48093
This was an easy change, but to test it I had to pipe
real Attachment objects into the queue testing system.
Doing so revealed a whole bunch of bugs in our unit tests,
which I fixed as part of this patch.
* Scripts/webkitpy/tool/commands/earlywarningsystem.py:
- This is the actual code change. This will not reduce the
backlog in the EWS queues much, but it will make rejections
much quicker for obsolete patches or closed bugs.
* Scripts/webkitpy/tool/commands/earlywarningsystem_unittest.py:
- Test my new code.
- Used a real attachment object and got rid of MockPatch
- Shared code between the mac-ews and cr-mac-ews tests.
* Scripts/webkitpy/tool/commands/queues_unittest.py:
- Can't use MockPatch anymore.
- Removing MockPatch found more bugs here!
* Scripts/webkitpy/tool/commands/queuestest.py:
* Scripts/webkitpy/tool/commands/sheriffbot_unittest.py:
* Scripts/webkitpy/tool/mocktool.py:
- MockBugzilla should not be a "Mock" object. Right now tool.bugs()
is allowed, but wrong. Making it not a Mock will make tool.bugs() correctly fail.
2010-10-22 Sheriff Bot <webkit.review.bot@gmail.com>
Unreviewed, rolling out r70301.
http://trac.webkit.org/changeset/70301
https://bugs.webkit.org/show_bug.cgi?id=48126
"Lang attribute layout tests failing" (Requested by satish on
#webkit).
* DumpRenderTree/LayoutTestController.cpp:
(setMockSpeechInputResultCallback):
* DumpRenderTree/LayoutTestController.h:
* DumpRenderTree/chromium/LayoutTestController.cpp:
(LayoutTestController::setMockSpeechInputResult):
* DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
(LayoutTestController::setMockSpeechInputResult):
* DumpRenderTree/mac/LayoutTestControllerMac.mm:
(LayoutTestController::setMockSpeechInputResult):
* DumpRenderTree/qt/LayoutTestControllerQt.cpp:
(LayoutTestController::setMockSpeechInputResult):
* DumpRenderTree/qt/LayoutTestControllerQt.h:
* DumpRenderTree/win/LayoutTestControllerWin.cpp:
(LayoutTestController::setMockSpeechInputResult):
* DumpRenderTree/wx/LayoutTestControllerWx.cpp:
(LayoutTestController::setMockSpeechInputResult):
2010-10-22 Leandro Gracia Gil <leandrogracia@google.com>
Reviewed by Jeremy Orlow.
Added a second parameter to setMockSpeechInputResult for
the language used in speech input.
https://bugs.webkit.org/show_bug.cgi?id=47089
* DumpRenderTree/LayoutTestController.cpp:
(setMockSpeechInputResultCallback):
* DumpRenderTree/LayoutTestController.h:
* DumpRenderTree/chromium/LayoutTestController.cpp:
(LayoutTestController::setMockSpeechInputResult):
* DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
(LayoutTestController::setMockSpeechInputResult):
* DumpRenderTree/mac/LayoutTestControllerMac.mm:
(LayoutTestController::setMockSpeechInputResult):
* DumpRenderTree/qt/LayoutTestControllerQt.h:
* DumpRenderTree/qt/LayoutTestControllerQt.cpp:
(LayoutTestController::setMockSpeechInputResult):
* DumpRenderTree/win/LayoutTestControllerWin.cpp:
(LayoutTestController::setMockSpeechInputResult):
* DumpRenderTree/wx/LayoutTestControllerWx.cpp:
(LayoutTestController::setMockSpeechInputResult):
2010-10-21 MORITA Hajime <morrita@google.com>
Reviewed by Kent Tamura.
[Win][DRT] should have LayoutTestController.hasSpellingMarker()
https://bugs.webkit.org/show_bug.cgi?id=47885
- Implemented LayoutTestController.hasSpellingMarker(),
- Added fake spellcheck implementation.
* DumpRenderTree/win/DumpRenderTree.cpp:
(main):
* DumpRenderTree/win/EditingDelegate.cpp:
(indexOfFirstWordCharacter):
(wordLength):
(EditingDelegate::checkSpellingOfString):
* DumpRenderTree/win/EditingDelegate.h:
* DumpRenderTree/win/LayoutTestControllerWin.cpp:
(LayoutTestController::hasSpellingMarker):
2010-10-21 Mihai Parparita <mihaip@chromium.org>
Unreviewed. Re-enable test that was mistakenly disabled by r67974.
* Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests_unittest.py:
2010-10-21 Mihai Parparita <mihaip@chromium.org>
Unreviewed fix for rebaseline-chromium-webkit-tests.
Port.diff_image no longer has a tolerance parameter. Also, use
get_option('tolerance') which is safer if the options object doesn't
define a tolerance attribute (it doesn't for the one used in
rebaseline_chromium_webkit_tests).
* Scripts/webkitpy/layout_tests/port/port_testcase.py:
* Scripts/webkitpy/layout_tests/port/webkit.py:
* Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
2010-10-21 Mihai Parparita <mihaip@chromium.org>
Reviewed by Ojan Vafai.
Add support for --tolerance in NRWT
https://bugs.webkit.org/show_bug.cgi?id=47959
Add support for the --tolerance flag in NRWT. The Port.diff_image
signature shouldn't need a tolerance parameter (it's not set per test),
just have ports that use it (currently only WebKitPort) read it from
the options object.
* Scripts/webkitpy/layout_tests/port/chromium.py:
* Scripts/webkitpy/layout_tests/port/port_testcase.py:
* Scripts/webkitpy/layout_tests/port/test.py:
* Scripts/webkitpy/layout_tests/port/webkit.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2010-10-21 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
Add --suggest-reviewers option to upload to auto-suggest reviewers for your patch
https://bugs.webkit.org/show_bug.cgi?id=48088
This is a first-pass. Works, but we'll eventually
turn this on by default, refine the suggestion algorithm
and possibly move it to a different place in the upload step order.
* Scripts/webkitpy/common/checkout/api.py:
* Scripts/webkitpy/tool/commands/download_unittest.py:
* Scripts/webkitpy/tool/commands/upload.py:
* Scripts/webkitpy/tool/commands/upload_unittest.py:
* Scripts/webkitpy/tool/mocktool.py:
* Scripts/webkitpy/tool/steps/__init__.py:
* Scripts/webkitpy/tool/steps/options.py:
2010-10-21 Mario Sanchez Prada <msanchez@igalia.com>
Reviewed by Chris Fleizach.
[GTK] Segfault while testing accessibility/iframe-bastardization.html
https://bugs.webkit.org/show_bug.cgi?id=30123
Check m_element before actually using it to get the parent.
* DumpRenderTree/gtk/AccessibilityUIElementGtk.cpp:
(AccessibilityUIElement::parentElement): Add an extra check to
make sure m_element points to a valid value before actually using
it to get the parent. Also, assert ATK_IS_OBJECT(m_element).
2010-10-21 Anders Carlsson <andersca@apple.com>
Reviewed by Adam Roben.
Crash evaluating JavaScript string that throws an exception
https://bugs.webkit.org/show_bug.cgi?id=48092
<rdar://problem/8487657>
Add a test that evaluates a JavaScript string that throws an exception and check that
the callback is called.
* TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
* TestWebKitAPI/Tests/WebKit2/EvaluateJavaScript.cpp: Added.
(TestWebKitAPI::didRunJavaScript):
(TestWebKitAPI::WebKit2_EvaluateJavaScript):
* TestWebKitAPI/win/TestWebKitAPI.vcproj:
2010-10-21 Adam Roben <aroben@apple.com>
Test that the plugin's HWND is invalidated when NPN_InvalidateRect is
called
Test for <http://webkit.org/b/48086> <rdar://problem/8482944>
Silverlight doesn't repaint in WebKit2
Reviewed by Anders Carlsson.
* DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp:
(PluginTest::NPN_InvalidateRect): Added. Calls through to the browser.
(executeScript): Added. Asks the browser to evaluate the script.
(PluginTest::waitUntilDone):
(PluginTest::notifyDone):
Added. Calls through to layoutTestController.
* DumpRenderTree/TestNetscapePlugIn/PluginTest.h: Added
* NPN_InvalidateRect and waitUntilDone/notifyDone.
* DumpRenderTree/TestNetscapePlugIn/Tests/win/NPNInvalidateRectInvalidatesWindow.cpp: Added.
(TemporaryWindowMover::moveSucceeded):
(TemporaryWindowMover::TemporaryWindowMover):
(TemporaryWindowMover::~TemporaryWindowMover):
This class moves a window on-screen and shows it, then moves it back and hides it.
(NPNInvalidateRectInvalidatesWindow::NPNInvalidateRectInvalidatesWindow):
Initialize our members.
(NPNInvalidateRectInvalidatesWindow::~NPNInvalidateRectInvalidatesWindow):
Delete our window mover if it hasn't been deleted already.
(NPNInvalidateRectInvalidatesWindow::NPP_SetWindow): Subclass the
plugin HWND and move the test harness window on screen.
(NPNInvalidateRectInvalidatesWindow::wndProc): Call through to onPaint
when we get a WM_PAINT message.
(NPNInvalidateRectInvalidatesWindow::onPaint): Do the test and tell
LayoutTestController we're done.
(NPNInvalidateRectInvalidatesWindow::testInvalidateRect): Validate
ourselves, invalidate our lower-right quadrant via NPN_InvalidateRect,
then check that our HWND's invalid region is the rect that we
invalidated.
* DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
* Added NPNInvalidateRectInvalidatesWindow.
2010-10-21 Daniel Bates <dbates@rim.com>
Add Git-support to do-file-rename
https://bugs.webkit.org/show_bug.cgi?id=48015
Fix tools. Export function scmMoveOrRenameFile so that it can be called from
do-file-rename and do-webcore-rename. I inadvertently left this out of the patch.
* Scripts/VCSUtils.pm:
2010-10-21 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
mac-ews is not properly releasing patches
https://bugs.webkit.org/show_bug.cgi?id=48076
mac-ews overrides process_work_item, so it was not calling
release_work_item like the default process_work_item would.
To fix this I made all the status-reporting methods just
release the patch. I expect we'll iterate on this design further.
* Scripts/webkitpy/common/net/statusserver.py:
* Scripts/webkitpy/tool/commands/earlywarningsystem_unittest.py:
* Scripts/webkitpy/tool/commands/queues.py:
* Scripts/webkitpy/tool/commands/queues_unittest.py:
2010-10-18 Martin Robinson <mrobinson@igalia.com>
Reviewed by Gustavo Noronha Silva.
[GTK] mathml flaky tests after adding SystemFonts to the theme
https://bugs.webkit.org/show_bug.cgi?id=47727
The Fontconfig setup in the GTK+ now specifically checks for and
loads the STIX fonts for MathML tests.
* DumpRenderTree/gtk/DumpRenderTree.cpp:
(initializeFonts): Add hooks for loading the STIX fonts specifically.
2010-10-20 Adam Roben <aroben@apple.com>
Test that the UI client gets notified when WKView receives a WM_CLOSE
message
Test for <http://webkit.org/b/48044> <rdar://problem/8488446> Pressing
Ctrl+W when viewing a full-page PDF destroys the WKView but doesn't
close its parent window
Reviewed by Jon Honeycutt.
* TestWebKitAPI/Tests/WebKit2/win/WMCloseCallsUIClientClose.cpp: Added.
(TestWebKitAPI::close): Record that this function was called.
(TestWebKitAPI::WebKit2_WMCloseCallsUIClientClose): Create a WKView,
send it a WM_CLOSE message, and test that the UI client's close
callback got called. Note that this will hang if the bug is
reintroduced; <http://webkit.org/b/48043> covers making TestWebKitAPI
able to handle hangs.
* TestWebKitAPI/win/TestWebKitAPI.vcproj: Added the new test.
2010-10-20 Adam Roben <aroben@apple.com>
Make prepare-ChangeLog much faster when using git
This change also seems to make it detect renames better in some cases.
Fixes <http://webkit.org/b/48040> prepare-ChangeLog is slow when using
git
Reviewed by David Kilzer.
* Scripts/prepare-ChangeLog:
(statusCommand):
(createPatchCommand):
Use "-M -C" instead of "-C -C -M" to tell git-diff to detect renames
and copies. The two "-C"s were making it read many more files than were
necessary.
2010-10-21 Sam Weinig <sam@webkit.org>
Reviewed by Adam Roben.
Null frame passed when running alert from UserScript run at document start
<rdar://problem/8573809>
https://bugs.webkit.org/show_bug.cgi?id=48036
* DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
* TestWebKitAPI/InjectedBundleController.cpp:
(TestWebKitAPI::InjectedBundleController::didReceiveMessage):
(TestWebKitAPI::InjectedBundleController::initializeTestNamed):
* TestWebKitAPI/InjectedBundleController.h:
* TestWebKitAPI/InjectedBundleTest.h:
(TestWebKitAPI::InjectedBundleTest::initialize):
* TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
* TestWebKitAPI/Tests/WebKit2/DocumentStartUserScriptAlertCrash.cpp: Added.
(TestWebKitAPI::runJavaScriptAlert):
(TestWebKitAPI::TEST):
* TestWebKitAPI/Tests/WebKit2/DocumentStartUserScriptAlertCrash_Bundle.cpp: Added.
(TestWebKitAPI::DocumentStartUserScriptAlertCrashTest::DocumentStartUserScriptAlertCrashTest):
(TestWebKitAPI::DocumentStartUserScriptAlertCrashTest::initialize):
Add test for invoking an alert during a UserScript run at document start.
2010-10-21 Andreas Kling <kling@webkit.org>
Reviewed by Adam Roben.
TestNetscapePlugIn: Actually call NPP_SetWindow on Unix
We were just returning NPERR_NO_ERROR previously.
* DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
(webkit_test_plugin_set_window):
2010-10-21 David Kilzer <ddkilzer@apple.com>
<http://webkit.org/b/48051> Move macros from DumpRenderTreeMac.h to config.h
Reviewed by Adam Roben.
* DumpRenderTree/config.h: Moved macros to here...
* DumpRenderTree/mac/DumpRenderTreeMac.h: ...from here.
Addresses a FIXME comment from r28419.
2010-10-21 David Kilzer <ddkilzer@apple.com>
<http://webkit.org/b/48047> Fix warnings found by check-Xcode-source-file-types
Reviewed by Adam Roben.
Fixes the following warnings:
WARNING: Incorrect file type 'sourcecode.cpp.objcpp' for file 'PixelDumpSupport.cpp'.
WARNING: Incorrect file type 'sourcecode.cpp.objcpp' for file 'LayoutTestController.cpp'.
WARNING: Incorrect file type 'sourcecode.cpp.objcpp' for file 'WorkQueue.cpp'.
WARNING: Incorrect file type 'sourcecode.cpp.objcpp' for file 'cg/PixelDumpSupportCG.cpp'.
WARNING: Incorrect file type 'sourcecode.cpp.objcpp' for file 'GCController.cpp'.
5 issues found for WebKitTools/DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj.
* DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
Removed explicit file type for *.cpp files above to make them
match other C++ source files. This exposed the fact that
DumpRenderTreeMac.h was not safe to include in C++ source, which
caused PixelDumpSupport.cpp and PixelDumpSupportCG.cpp to fail
to build.
* DumpRenderTree/PixelDumpSupport.cpp: Adjusted header order.
* DumpRenderTree/mac/DumpRenderTreeMac.h: Made safe to include
in C++ source files. Included CoreFoundation/CoreFoundation.h
to make sure all CF types were defined. Removed CFStringRef
typedef.
2010-10-21 Adam Roben <aroben@apple.com>
Attempt to fix plugins/pass-different-npp-struct.html on GTK.
See <http://webkit.org/b/47690>.
* GNUmakefile.am: Added PassDifferentNPPStruct.cpp.
2010-10-21 Gabor Rapcsanyi <rgabor@inf.u-szeged.hu>
Reviewed by Csaba Osztrogonác.
[NRWT] Set ImageDiff path on Qt port
https://bugs.webkit.org/show_bug.cgi?id=48052
* Scripts/webkitpy/layout_tests/port/qt.py:
2010-10-21 Adam Barth <abarth@webkit.org>
Reviewed by Eric Seidel.
Don't CC authors of flaky tests
https://bugs.webkit.org/show_bug.cgi?id=48038
Some authors found this too annoying. We'll look for another way to
close the flaky test loop.
* Scripts/webkitpy/tool/commands/queues.py:
* Scripts/webkitpy/tool/commands/queues_unittest.py:
2010-10-21 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
commit-queue should run run-webkit-tests with --no-new-test-results
https://bugs.webkit.org/show_bug.cgi?id=47998
* Scripts/webkitpy/tool/steps/runtests.py:
* Scripts/webkitpy/tool/steps/steps_unittest.py:
2010-10-21 Gabor Rapcsanyi <rgabor@inf.u-szeged.hu>
Reviewed by Ojan Vafai.
[NRWT] Get child process number from an environment variable
https://bugs.webkit.org/show_bug.cgi?id=47981
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2010-10-20 David Kilzer <ddkilzer@apple.com>
<http://webkit.org/b/47754> New script to verify explicit source file types in Xcode project files
Reviewed by Darin Adler.
The script parses an Xcode project file and makes sure the file
extension matches the explicit file type set for all source
files. Note that the majority of source files will have their
type set by Xcode, so the script doesn't check them since there
is no need to second-guess Xcode.
* Scripts/check-Xcode-source-file-types: Added. Code borrowed
heavily from sort-Xcode-project-file.
2010-10-20 Adam Roben <aroben@apple.com>
Windows build fix
* TestWebKitAPI/win/PlatformUtilitiesWin.cpp: Added missing #include.
2010-10-20 Adam Roben <aroben@apple.com>
Qt test fix
* DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro: Added
PassDifferentNPPStruct.cpp.
2010-10-20 Sam Weinig <sam@webkit.org>
Reviewed by Adam Roben.
Add ability to test injected bundle API using TestWebKitAPI
https://bugs.webkit.org/show_bug.cgi?id=48027
* TestWebKitAPI/InjectedBundleController.cpp: Added.
* TestWebKitAPI/InjectedBundleController.h: Added.
Shared main object for bundle functionality.
* TestWebKitAPI/InjectedBundleMain.cpp: Added.
Bundle entry point.
* TestWebKitAPI/InjectedBundleTest.h: Added.
Base class for which the bundle portion of a test derives from.
* TestWebKitAPI/PlatformUtilities.h:
* TestWebKitAPI/PlatformUtilities.cpp: Added.
* TestWebKitAPI/mac/PlatformUtilitiesMac.mm:
* TestWebKitAPI/win/PlatformUtilitiesWin.cpp:
(TestWebKitAPI::Util::createInjectedBundlePath):
(TestWebKitAPI::Util::createURLForResource):
(TestWebKitAPI::Util::URLForNonExistentResource):
Add helper to create a context with the shared injected bundle,
and send the initial message to set up the test.
* TestWebKitAPI/Configurations/InjectedBundle.xcconfig: Added.
* TestWebKitAPI/InjectedBundle-Info.plist: Added.
Add mac configuration files.
* TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
Add the new files.
* TestWebKitAPI/Tests/WebKit2/InjectedBundleBasic.cpp: Added.
* TestWebKitAPI/Tests/WebKit2/InjectedBundleBasic_Bundle.cpp: Added.
Add a simple initial bundle test.
2010-10-20 Eric Seidel <eric@webkit.org>
Unreviewed. Fixing /release-patch url used by the EWS bots.
EWS should test patches with r+
https://bugs.webkit.org/show_bug.cgi?id=35460
* Scripts/webkitpy/common/net/statusserver.py:
- I changed the URL during development, and 404s are
intentionally silenced during release_patch.
2010-10-14 Adam Roben <aroben@apple.com>
Test that passing a different NPP struct back to the browser doesn't
cause an assertion failure
Test for <http://webkit.org/b/47690> <rdar://problem/8553020>
Assertion failure in NetscapePlugin::fromNPP when using Shockwave in
WebKit2
Reviewed by John Sullivan.
* DumpRenderTree/TestNetscapePlugIn/Tests/PassDifferentNPPStruct.cpp: Copied from WebKitTools/DumpRenderTree/TestNetscapePlugIn/Tests/win/GetValueNetscapeWindow.cpp.
(PassDifferentNPPStruct::PassDifferentNPPStruct): Initialize our
members.
(PassDifferentNPPStruct::NPP_SetWindow): Pass a different NPP to the
browser than the one it gave us in NPP_New and see if it works.
* DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
* DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
Added PassDifferentNPPStruct.
2010-10-19 Adam Roben <aroben@apple.com>
Gently nudge old-run-webkit-tests toward working with Win32 Perl
This makes old-run-webkit-tests able to build DRT and find all the
tests to run. It even invokes DRT and passes it the list of tests. But
DRT ends up hung blocking on I/O.
Fixes <http://webkit.org/b/47961> Get old-run-webkit-tests mostly
working with Win32 Perl
Reviewed by David Kilzer.
* Scripts/old-run-webkit-tests:
- Use File::Spec instead of manually concatenating paths
- Use dirname instead of manually stripping off the base name
- Use isCygwin/isWindows/isAppleWinWebKit more judiciously
- Explicitly invoke Perl when running Perl scripts
- Quote paths when using them in regular expressions to allow them
to include characters that have special meanings in regular
expressions
* Scripts/run-webkit-tests: Use File::Spec instead of manually
concatenating paths.
* Scripts/webkitdirs.pm:
- Remove the unused $windowsTmpPath variable
- Use isCygwin/isWindows/isAppleWinWebKit more judiciously
- Only pass paths to cygpath when using Cygwin Perl
- Only use pdevenv when using Cygwin Perl, for now
2010-10-20 Daniel Bates <dbates@rim.com>
Reviewed by Martin Robinson.
Add Git-support to do-file-rename
https://bugs.webkit.org/show_bug.cgi?id=48015
Also, abstracts the SCM move/rename functionality in do-file-rename and
do-webcore-rename into a common function VCSUtils::scmMoveOrRenameFile().
Currently, do-file-rename is hard coded to assume the SCM is Subversion.
Instead, we should abstract the rename logic to be SCM-independent. This
will allow us to add Git support as well move such functionality into
our SCM library VCSUtils, where it can be shared by do-webcore-rename.
* Scripts/VCSUtils.pm:
- Added function scmMoveOrRenameFile.
* Scripts/do-file-rename: Modified to call VCSUtils::scmMoveOrRenameFile().
* Scripts/do-webcore-rename: Ditto.
2010-10-20 Adam Roben <aroben@apple.com>
Fix old-run-webkit-tests when there's a space in the path to DRT
Reviewed by Jon Honeycutt.
* Scripts/old-run-webkit-tests: Quote the path to DRT before executing
it.
2010-10-20 Kenneth Russell <kbr@google.com>
Reviewed by Dimitri Glazkov.
Regression in chromium_gpu_unittests after r70175
https://bugs.webkit.org/show_bug.cgi?id=48008
* Scripts/webkitpy/layout_tests/port/chromium_gpu_unittest.py:
2010-10-20 Kenneth Russell <kbr@google.com>
Reviewed by James Robinson.
chromium_gpu port of new-run-webkit-tests must do Linux -> Win expectations fallback
https://bugs.webkit.org/show_bug.cgi?id=48005
* Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
2010-10-20 Adam Barth <abarth@webkit.org>
Reviewed by Eric Seidel.
webkit-patch doesn't get along with git rm
https://bugs.webkit.org/show_bug.cgi?id=47940
Turns out we need to pass "--" to tell git this is a path.
* Scripts/webkitpy/common/checkout/scm.py:
2010-10-20 Robert Hogan <robert@webkit.org>
Reviewed by Antonio Gomes.
[Qt] Fix layoutTestController.getJsObjectCount
Qt Bridge doesn't know size_t so pass result as unsigned int.
Unskip fast/dom/gc-10.html
https://bugs.webkit.org/show_bug.cgi?id=47931
* DumpRenderTree/qt/GCControllerQt.cpp:
(GCController::getJSObjectCount):
* DumpRenderTree/qt/GCControllerQt.h:
2010-10-20 Sheriff Bot <webkit.review.bot@gmail.com>
Unreviewed, rolling out r70149.
http://trac.webkit.org/changeset/70149
https://bugs.webkit.org/show_bug.cgi?id=47989
"Build breaks in mac and win" (Requested by satish on
#webkit).
* DumpRenderTree/LayoutTestController.cpp:
(setMockSpeechInputResultCallback):
* DumpRenderTree/LayoutTestController.h:
* DumpRenderTree/chromium/LayoutTestController.cpp:
(LayoutTestController::setMockSpeechInputResult):
* DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
(LayoutTestController::setMockSpeechInputResult):
2010-10-20 Leandro Gracia Gil <leandrogracia@google.com>
Reviewed by Jeremy Orlow.
Added a second parameter to setMockSpeechInputResult for
the language used in speech input.
https://bugs.webkit.org/show_bug.cgi?id=47089
* DumpRenderTree/LayoutTestController.cpp:
(setMockSpeechInputResultCallback):
* DumpRenderTree/LayoutTestController.h:
* DumpRenderTree/chromium/LayoutTestController.cpp:
(LayoutTestController::setMockSpeechInputResult):
* DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
(LayoutTestController::setMockSpeechInputResult):
2010-10-19 Adam Roben <aroben@apple.com>
Teach update-webkit-support-libs about the new versioning of
WebKitSupportLibrary
Fixes <http://webkit.org/b/47915> update-webkit-support-libs should
check version numbers instead of modification times
Reviewed by Sam Weinig.
* Scripts/update-webkit-support-libs: Fetch the expected version
number from developer.apple.com, then compare with the version number
of the extracted library and of the zipped library to see if anything
needs to be done. Removed code that tracked the modified timestamp of
the library, as it is no longer needed.
2010-10-19 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
EWS should test patches with r+
https://bugs.webkit.org/show_bug.cgi?id=35460
* QueueStatusServer/handlers/submittoews.py:
* QueueStatusServer/model/queues.py:
* Scripts/webkitpy/tool/bot/patchcollection.py: Removed.
* Scripts/webkitpy/tool/bot/patchcollection_unittest.py: Removed.
* Scripts/webkitpy/tool/commands/queues.py:
2010-10-19 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
Teach feeder-queue how to feed the EWS bots
https://bugs.webkit.org/show_bug.cgi?id=47943
queues.webkit.org already knew how to accept EWS submissions
via /submit-to-ews. This teaches the feeder queue how to post
to that page with any new r? patches it sees.
* QueueStatusServer/model/activeworkitems_unitest.py: Added.
- More unit testing is always a good thing.
* Scripts/webkitpy/common/net/bugzilla.py:
* Scripts/webkitpy/common/net/statusserver.py:
* Scripts/webkitpy/tool/bot/feeders.py:
* Scripts/webkitpy/tool/commands/queues.py:
* Scripts/webkitpy/tool/commands/queues_unittest.py:
* Scripts/webkitpy/tool/mocktool.py:
2010-10-19 Eric Seidel <eric@webkit.org>
Unreviewed, just fixing test-webkitpy. I'm really on a roll today.
commit-queue gets stuck when release-patch returns 404
https://bugs.webkit.org/show_bug.cgi?id=47935
Fix test-webkitpy and unittest NetworkTransaction.
* Scripts/webkitpy/common/net/networktransaction.py:
* Scripts/webkitpy/common/net/networktransaction_unittest.py:
2010-10-19 Eric Seidel <eric@webkit.org>
Unreviewed.
commit-queue gets stuck when release-patch returns 404
https://bugs.webkit.org/show_bug.cgi?id=47935
Turns out ClientForm gets upset if passed an int() instead of a string type.
Yay for untestable code.
* Scripts/webkitpy/common/net/statusserver.py:
2010-10-19 Eric Seidel <eric@webkit.org>
Unreviewed. Will get Adam's commentary after his meeting
for now this gets the commit-cluster back running.
commit-queue gets stuck when release-patch returns 404
https://bugs.webkit.org/show_bug.cgi?id=47935
I taught NetworkTransaction the basics of 404 handling.
We'll want to go back and teach it how to handle urllib2 404's too
and then deploy it to the places that want it.
* QueueStatusServer/handlers/releasepatch.py:
* Scripts/webkitpy/common/net/buildbot.py:
* Scripts/webkitpy/common/net/networktransaction.py:
* Scripts/webkitpy/common/net/statusserver.py:
2010-10-19 Eric Seidel <eric@webkit.org>
Unreviewed. Fixing typos in my previous commit.
Make patch release explicit and not a magic part of "retry" status
https://bugs.webkit.org/show_bug.cgi?id=47909
All of these typos again due to our inability to unit
test much of this code. I added one unit test where
possible. activeworkitems_unittest.py will be in a separate patch.
* QueueStatusServer/handlers/releasepatch.py:
* QueueStatusServer/main.py:
* QueueStatusServer/model/activeworkitems.py:
* QueueStatusServer/model/workitems.py:
* QueueStatusServer/model/workitems_unittest.py:
* QueueStatusServer/templates/releasepatch.html:
* Scripts/webkitpy/common/net/statusserver.py:
2010-10-19 Tony Chang <tony@chromium.org>
Reviewed by Kent Tamura.
[chromium] Use webkit's TestNetscapePlugIn in DRT mac
https://bugs.webkit.org/show_bug.cgi?id=47850
* DumpRenderTree/chromium/TestNetscapePlugIn/Info.plist: Added. Forked
because we're going to name our plugin WebKitTestNetscapePlugIn
temporarily until the chromium forked plugin goes away.
2010-10-19 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
Make patch release explicit and not a magic part of "retry" status
https://bugs.webkit.org/show_bug.cgi?id=47909
This moves us another step closer to running r+ patches on the EWS bots.
Currently all bots just spam /update-work-items with their list of current
work items. queues.webkit.org uses that data for display. As part of making
the EWS run r+ patches, we're moving the official list of patches-to-process
into the server, and feeding them out to bots one at a time. We need to be
able to remove patches from the queues one at a time instead of just spamming
/update-work-items with a new complete list. That's what this patch adds.
* QueueStatusServer/handlers/nextpatch.py:
* QueueStatusServer/handlers/queuestatus.py:
* QueueStatusServer/handlers/releasepatch.py: Added.
* QueueStatusServer/handlers/statusbubble_unittest.py:
- Fix a typo causing test failure. This was not caught by the bots
because they don't have AppEngineLauncher installed and thus don't run
the QueueStatusServer tests.
* QueueStatusServer/handlers/updatestatus.py:
* QueueStatusServer/model/activeworkitems.py:
* QueueStatusServer/templates/releasepatch.html: Added.
* Scripts/webkitpy/common/net/statusserver.py:
* Scripts/webkitpy/tool/commands/queues.py:
* Scripts/webkitpy/tool/mocktool.py:
2010-10-19 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
cr-mac bubble has caused status bubbles to wrap
https://bugs.webkit.org/show_bug.cgi?id=47928
We now have too many EWSes to fit in the bugs.webkit.org
status-bubble iframe when more than a couple EWS builds are pending.
To fix this I've reduced the space taken up by queue position,
and also moved cr-mac to the end of the list (since it's going to be
triple-digits for a while).
* QueueStatusServer/model/queues.py:
* QueueStatusServer/templates/statusbubble.html:
2010-10-19 Kenneth Russell <kbr@google.com>
Reviewed by David Levin.
chromium_gpu port of new-run-webkit-tests must search chromium-gpu directory for expectations
https://bugs.webkit.org/show_bug.cgi?id=47874
* Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
* Scripts/webkitpy/layout_tests/port/chromium_gpu_unittest.py:
2010-10-19 Tony Chang <tony@chromium.org>
Reviewed by Kent Tamura.
[chromium] DumpRenderTree shouldn't put '.' in include path
https://bugs.webkit.org/show_bug.cgi?id=47877
Fix include paths.
* DumpRenderTree/chromium/AccessibilityController.cpp:
* DumpRenderTree/chromium/AccessibilityUIElement.cpp:
* DumpRenderTree/chromium/AccessibilityUIElement.h:
* DumpRenderTree/chromium/CppBoundClass.cpp:
* DumpRenderTree/chromium/CppVariant.cpp:
* DumpRenderTree/chromium/CppVariant.h:
* DumpRenderTree/chromium/DRTDevToolsAgent.cpp:
* DumpRenderTree/chromium/DRTDevToolsAgent.h:
* DumpRenderTree/chromium/DRTDevToolsCallArgs.h:
* DumpRenderTree/chromium/DRTDevToolsClient.cpp:
* DumpRenderTree/chromium/DRTDevToolsClient.h:
* DumpRenderTree/chromium/EventSender.cpp:
* DumpRenderTree/chromium/EventSender.h:
* DumpRenderTree/chromium/LayoutTestController.cpp:
* DumpRenderTree/chromium/LayoutTestController.h:
* DumpRenderTree/chromium/MockSpellCheck.cpp:
* DumpRenderTree/chromium/NotificationPresenter.cpp:
* DumpRenderTree/chromium/NotificationPresenter.h:
* DumpRenderTree/chromium/PlainTextController.cpp:
* DumpRenderTree/chromium/Task.cpp:
* DumpRenderTree/chromium/TestNavigationController.h:
* DumpRenderTree/chromium/TestShell.cpp:
* DumpRenderTree/chromium/TestWebWorker.h:
* DumpRenderTree/chromium/TextInputController.cpp:
* DumpRenderTree/chromium/WebPreferences.cpp:
* DumpRenderTree/chromium/WebPreferences.h:
* DumpRenderTree/chromium/WebThemeEngineDRT.cpp:
* DumpRenderTree/chromium/WebThemeEngineDRT.h:
* DumpRenderTree/chromium/WebViewHost.cpp:
* DumpRenderTree/chromium/WebViewHost.h:
2010-10-19 Sam Weinig <sam@webkit.org>
Reviewed by Dan Bernstein.
Need to include WKErrorRef in the WKPageDidFail... functions
https://bugs.webkit.org/show_bug.cgi?id=47871
Update tools for new parameter in failure callbacks.
* MiniBrowser/mac/BrowserWindowController.m:
(didFailProvisionalLoadWithErrorForFrame):
(didFailLoadWithErrorForFrame):
(-[BrowserWindowController updateProvisionalURLForFrame:]):
* TestWebKitAPI/Tests/WebKit2/FailedLoad.cpp:
(TestWebKitAPI::didFailProvisionalLoadWithErrorForFrame):
* WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
(WTR::InjectedBundlePage::didFailProvisionalLoadWithErrorForFrame):
(WTR::InjectedBundlePage::didFailLoadWithErrorForFrame):
* WebKitTestRunner/InjectedBundle/InjectedBundlePage.h:
2010-10-19 Luiz Agostini <luiz.agostini@openbossa.org>
Reviewed by Antti Koivisto.
[Qt] WebKit2 MacOS build fix
https://bugs.webkit.org/show_bug.cgi?id=47897
Qt WebKit2 MacOS build fix.
* WebKitTestRunner/PlatformWebView.h:
* WebKitTestRunner/qt/TestControllerQt.cpp:
2010-10-19 Adam Barth <abarth@webkit.org>
Reviewed by Eric Seidel.
webkit-patch stats the filesystem too many times
https://bugs.webkit.org/show_bug.cgi?id=47883
This patch attempts to cache the list of changed files more agressively
and to use that list to compute the diff instead of stating the file
system again.
* Scripts/webkitpy/common/checkout/api.py:
* Scripts/webkitpy/common/checkout/scm.py:
* Scripts/webkitpy/tool/mocktool.py:
* Scripts/webkitpy/tool/steps/abstractstep.py:
* Scripts/webkitpy/tool/steps/editchangelog.py:
* Scripts/webkitpy/tool/steps/preparechangelog.py:
2010-10-19 David Kilzer <ddkilzer@apple.com>
<http://webkit.org/b/47741> Make sort-Xcode-project-file a little more friendly
Reviewed by Darin Adler.
* Scripts/sort-Xcode-project-file:
- Don't print an error message about missing arguments when
-h|--help is used.
- Allow Xcode project files to be specified as Project.xcodeproj
instead of Project.xcodeproj/project.pbxproj.
2010-10-18 Jocelyn Turcotte <jocelyn.turcotte@nokia.com>
Reviewed by Andreas Kling.
[Qt] QtTestBrowser: Prevent calling load() directly from loadFinished() in robot mode.
https://bugs.webkit.org/show_bug.cgi?id=47809
Connecting a call to load from the loadFinished signal can cause
re-entrance crashes in WebCore. This patch uses a timer to do so,
also giving some time to subsequent frames to finish loading.
* QtTestBrowser/urlloader.cpp:
(UrlLoader::UrlLoader):
(UrlLoader::loadNext):
(UrlLoader::checkIfFinished):
(UrlLoader::frameLoadStarted):
(UrlLoader::frameLoadFinished):
* QtTestBrowser/urlloader.h:
2010-10-19 Sergio Villar Senín <svillar@igalia.com>
Reviewed by Xan Lopez.
[GTK] http/history tests are failing
https://bugs.webkit.org/show_bug.cgi?id=36173
Clear the history each time a test is run. Return the actual
history item count when calling
LayoutTestController::webHistoryItemCount
* DumpRenderTree/gtk/DumpRenderTree.cpp:
(resetDefaultsToConsistentValues):
* DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
(LayoutTestController::webHistoryItemCount):
2010-10-18 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
Add Chromium Mac EWS to the list of queues at queues.webkit.org
https://bugs.webkit.org/show_bug.cgi?id=47878
* QueueStatusServer/model/queues.py:
2010-10-18 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
queues.webkit.org/next_patch is always 404
https://bugs.webkit.org/show_bug.cgi?id=47881
With the addition of the Queue class, I changed most of the
code to lookup WorkItems using get_or_insert with a key_name
instead of WorkItems.all().filter(queue_name=).
Because the new get_or_insert code uses an explicit key_name
(which is obviously different from the previously autogenerated
ones), there were new WorkItem records created for each queue.
However, some parts of the code still use WorkItems.all().filter,
thus some parts were getting the new record and some parts the old record.
The same basic bug was occurring with ActiveWorkItems, because I
changed the key_name for that class as well.
To fix this I've moved more of the code over to using Queue.*work_items.
I've also enabled the datastore_admin (new in GAE 1.3.8) so that
we can go delete the old WorkItems records.
I also changed remote_api to use the new builtin: syntax (also added in GAE 1.3.8).
* QueueStatusServer/app.yaml:
* QueueStatusServer/handlers/queuestatus.py:
* QueueStatusServer/handlers/recentstatus.py:
* QueueStatusServer/handlers/updatestatus.py:
* QueueStatusServer/handlers/updateworkitems.py:
2010-10-18 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
Correct a bunch of typos in QueueStatusServer
https://bugs.webkit.org/show_bug.cgi?id=47880
These are all due to our complete lack of unit testing in QueueStatusServer.
I added a couple unit tests to cover a few of these fixes, but most of these
are still only caught by actually running the application.
* QueueStatusServer/handlers/nextpatch.py:
* QueueStatusServer/handlers/statusbubble.py:
* QueueStatusServer/handlers/statusbubble_unittest.py: Added.
* QueueStatusServer/handlers/updateworkitems.py:
* QueueStatusServer/model/attachment.py:
* QueueStatusServer/model/queuepropertymixin.py:
* QueueStatusServer/model/queuepropertymixin_unittest.py:
* QueueStatusServer/model/workitems.py:
* QueueStatusServer/model/workitems_unittest.py: Added.
2010-10-18 Adam Barth <abarth@webkit.org>
Disable this test because it's failing on the bots and the authors
aren't around to fix it.
* Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2010-10-18 Adam Barth <abarth@webkit.org>
Reviewed by Eric Seidel.
CC authors of flaky tests when the commit-queue hits a flaky test
https://bugs.webkit.org/show_bug.cgi?id=47872
* Scripts/webkitpy/common/checkout/api.py:
* Scripts/webkitpy/common/net/layouttestresults.py:
* Scripts/webkitpy/tool/commands/queues.py:
* Scripts/webkitpy/tool/commands/queues_unittest.py:
* Scripts/webkitpy/tool/mocktool.py:
2010-10-18 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
Make it possible to run a chromium-mac-ews builder
https://bugs.webkit.org/show_bug.cgi?id=47876
Since we can't run Mac OS X in a VM, we need to only run committer patches.
There was a multiple inheritance problem which was holding this patch back,
but I decided to just ignore the problem and go with a functional hack for now.
* Scripts/webkitpy/tool/commands/earlywarningsystem.py:
* Scripts/webkitpy/tool/commands/earlywarningsystem_unittest.py:
2010-10-18 Kenneth Russell <kbr@google.com>
Reviewed by Tony Chang.
Regression in run_webkit_tests_unittest from r70017
https://bugs.webkit.org/show_bug.cgi?id=47875
* Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2010-10-18 Adam Barth <abarth@webkit.org>
Reviewed by Eric Seidel.
Implement webkit-patch suggest-reviewers
https://bugs.webkit.org/show_bug.cgi?id=47866
* Scripts/webkitpy/common/checkout/api.py:
- The main logic. We look at the last five changes to each
modified (non-ChangeLog) file and collect up the reviewers of
those changes as well as the authors of those changes who are
reviewers.
* Scripts/webkitpy/common/checkout/api_unittest.py:
- Test the logic with some fun mocks.
* Scripts/webkitpy/common/checkout/scm.py:
- Fix a bug when you have local git commits.
* Scripts/webkitpy/common/checkout/scm_unittest.py:
- Test that the bug is fixed.
* Scripts/webkitpy/tool/commands/queries.py:
- Add the query.
2010-10-18 Kenneth Russell <kbr@google.com>
Reviewed by Eric Seidel.
new-run-webkit-tests produces corrupt PNG baselines on Windows
https://bugs.webkit.org/show_bug.cgi?id=47867
* Scripts/webkitpy/layout_tests/port/base.py:
2010-10-18 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
Make it possible to submit patches to the EWS bots
https://bugs.webkit.org/show_bug.cgi?id=47869
* QueueStatusServer/handlers/nextpatch.py:
- Move more logic into Queue, so that it can be shared with SubmitToEWS.
* QueueStatusServer/handlers/queuestatus.py:
- Fix two typos from a previous commit.
* QueueStatusServer/handlers/submittoews.py: Added.
* QueueStatusServer/handlers/updatestatus.py:
- Use the new is_retry_request method to share this (hacky) code with SubmitToEWS
* QueueStatusServer/main.py:
- Add /submit-to-ews
* QueueStatusServer/model/queuepropertymixin.py:
- Fix circular imports caused by adding Queue.work_items()
* QueueStatusServer/model/queues.py:
- Add work_items() and active_work_items()
* QueueStatusServer/model/queuestatus.py:
* QueueStatusServer/model/workitems.py:
- Add transaction-safe add/remove methods.
* QueueStatusServer/templates/submittoews.html: Added.
2010-10-18 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
Deploy Queue class in more places throughout QueueStatusServer
https://bugs.webkit.org/show_bug.cgi?id=47855
I also caught two typos from the previous change. Unfortunately
I don't yet know how to unittest request handlers yet.
* QueueStatusServer/handlers/dashboard.py:
* QueueStatusServer/handlers/statusbubble.py:
* QueueStatusServer/handlers/updateworkitems.py:
* QueueStatusServer/model/activeworkitems.py:
* QueueStatusServer/model/attachment.py:
* QueueStatusServer/model/queuepropertymixin.py: Added.
* QueueStatusServer/model/queuepropertymixin_unittest.py: Added.
* QueueStatusServer/model/queuestatus.py:
* QueueStatusServer/model/workitems.py:
2010-10-18 Adam Barth <abarth@webkit.org>
Reviewed by Eric Seidel.
scm.py should be able tell us what revisions made changes to a given file
https://bugs.webkit.org/show_bug.cgi?id=47863
Look again, your SCM.py can now log files.
* Scripts/webkitpy/common/checkout/scm.py:
* Scripts/webkitpy/common/checkout/scm_unittest.py:
2010-10-18 Dirk Pranke <dpranke@chromium.org>
Reviewed by Eric Siedel.
new-run-webkit-tests: clean up the options-parsing code in the port
classes.
This change modifies the Port interface to have a get_option() and
set_option_default() method for accessing the options argument
passed to the constructor. If the constructor is not passed an
options argument, we default to a MockOptions() argument from
mocktool, which has the same semantics we want.
Note that there is a disadvantage to port.get_option('foo') over
port._options.foo, which is that you lose some of the checking
for whether 'foo' is set (typos result in the default value, not
an exception being raised. This is desired in this case, since the
Port class is not allowed to assume that options does have any
particular values set, and so this change ensures that all of
the subclasses are following the same, intended, logic.
Arguably this is the wrong semantics to have, and the Port
classes should be able to assume a default set of
attributes/arguments, but that change will need to wait for a
different CL where we can modify new-run-webkit-tests to pull a
list of arguments from the port factory routines.
Also, add unit tests for webkitpy.tool.mocktool.MockOptions .
https://bugs.webkit.org/show_bug.cgi?id=47510
* Scripts/webkitpy/layout_tests/port/base.py:
* Scripts/webkitpy/layout_tests/port/base_unittest.py:
* Scripts/webkitpy/layout_tests/port/chromium.py:
* Scripts/webkitpy/layout_tests/port/chromium_gpu_unittest.py:
* Scripts/webkitpy/layout_tests/port/chromium_linux.py:
* Scripts/webkitpy/layout_tests/port/chromium_mac.py:
* Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
* Scripts/webkitpy/layout_tests/port/chromium_win.py:
* Scripts/webkitpy/layout_tests/port/dryrun.py:
* Scripts/webkitpy/layout_tests/port/factory_unittest.py:
* Scripts/webkitpy/layout_tests/port/mac_unittest.py:
* Scripts/webkitpy/layout_tests/port/port_testcase.py:
* Scripts/webkitpy/layout_tests/port/test.py:
* Scripts/webkitpy/layout_tests/port/webkit.py:
* Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests_unittest.py:
* Scripts/webkitpy/tool/mocktool_unittest.py: Added.
2010-10-18 Dirk Pranke <dpranke@chromium.org>
Reviewed by Eric Seidel.
Re-submit a revised version of r69638 - enabling new-run-webkit-tests
under cygwin. The initial version had a bug in base:uri_to_test_name
that was causing tests to fail. This version corrects that bug, but
also makes the code safer by calling cygpath more reliably, and
leaving a long-running cygpath process open.
This patch also corrects a couple of minor bugs in http_lock_unittest,
chromium_unittest, and dedpulicate_tests_unittest that showed up
while testing this.
https://bugs.webkit.org/show_bug.cgi?id=47220
* Scripts/webkitpy/common/system/path.py:
* Scripts/webkitpy/common/system/path_unittest.py:
* Scripts/webkitpy/layout_tests/deduplicate_tests_unittest.py:
* Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
* Scripts/webkitpy/layout_tests/port/base.py:
* Scripts/webkitpy/layout_tests/port/base_unittest.py:
* Scripts/webkitpy/layout_tests/port/chromium.py:
* Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
* Scripts/webkitpy/layout_tests/port/http_lock_unittest.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2010-10-18 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
Add Queue class and add minimal unittesting of QueueStatusServer code
https://bugs.webkit.org/show_bug.cgi?id=47847
* QueueStatusServer/handlers/dashboard.py:
* QueueStatusServer/handlers/queuestatus.py:
* QueueStatusServer/handlers/recentstatus.py:
* QueueStatusServer/handlers/statusbubble.py:
* QueueStatusServer/handlers/updateworkitems.py:
* QueueStatusServer/model/attachment.py:
* QueueStatusServer/model/queues.py:
* QueueStatusServer/model/queues_unittest.py: Added.
* QueueStatusServer/model/svnrevision.py:
* Scripts/test-webkitpy:
* Scripts/webkitpy/test/main.py:
2010-10-18 Anders Carlsson <andersca@apple.com>
Fix build.
* DumpRenderTree/TestNetscapePlugIn/main.cpp:
(handleEventCarbon):
2010-10-18 Stuart Morgan <stuartmorgan@chromium.org>
Reviewed by Eric Seidel.
Switch to using the new Carbon NPAPI event declarations, and remove
the old ones.
https://bugs.webkit.org/show_bug.cgi?id=40784
* DumpRenderTree/TestNetscapePlugIn/main.cpp:
(handleEventCarbon):
2010-10-18 David Levin <levin@chromium.org>
Reviewed by Adam Barth.
check-webkit-style should treat the GObject binding directory like other GTK directories.
https://bugs.webkit.org/show_bug.cgi?id=47796
* Scripts/webkitpy/style/checker.py: Added the GObject binding directory
with the other gtk directories (and fixed typo).
2010-10-18 David Levin <levin@chromium.org>
Reviewed by Oliver Hunt.
check-webkit-style needs to ignore underscores in opcode names and vm_throw
https://bugs.webkit.org/show_bug.cgi?id=47789
* Scripts/webkitpy/style/checker.py: Added the exception for the assembler directory.
* Scripts/webkitpy/style/checkers/cpp.py: Added special cased names.
* Scripts/webkitpy/style/checkers/cpp_unittest.py: Added unit tests for the special cases.
2010-10-18 MORITA Hajime <morrita@google.com>
Reviewed by Kent Tamura.
TextInputController.hasSpellingMarkers() should be owned by LayoutTestController
https://bugs.webkit.org/show_bug.cgi?id=47659
Moved hasSpellingMarkers() from TextInputController to
LayoutTestController. Currently the implementation is available
only for Mac and for Chromium.
* DumpRenderTree/LayoutTestController.cpp:
(hasSpellingMarkerCallback):
(LayoutTestController::staticFunctions):
* DumpRenderTree/LayoutTestController.h:
* DumpRenderTree/chromium/LayoutTestController.cpp:
(LayoutTestController::LayoutTestController):
(LayoutTestController::hasSpellingMarker):
* DumpRenderTree/chromium/LayoutTestController.h:
* DumpRenderTree/chromium/TextInputController.cpp:
(TextInputController::TextInputController):
(TextInputController::makeAttributedString):
* DumpRenderTree/chromium/TextInputController.h:
* DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
(LayoutTestController::hasSpellingMarker):
* DumpRenderTree/mac/LayoutTestControllerMac.mm:
(LayoutTestController::hasSpellingMarker):
* DumpRenderTree/mac/TextInputController.m:
(+[TextInputController isSelectorExcludedFromWebScript:]):
(+[TextInputController webScriptNameForSelector:]):
* DumpRenderTree/qt/LayoutTestControllerQt.cpp:
(LayoutTestController::hasSpellingMarker):
* DumpRenderTree/qt/LayoutTestControllerQt.h:
* DumpRenderTree/win/LayoutTestControllerWin.cpp:
(LayoutTestController::hasSpellingMarker):
* DumpRenderTree/wx/LayoutTestControllerWx.cpp:
(LayoutTestController::hasSpellingMarker):
2010-10-18 Adam Barth <abarth@webkit.org>
Reviewed by David Levin.
commit-queue's flaky test notice is very Pythony
https://bugs.webkit.org/show_bug.cgi?id=47790
This patch make the list of tests delimted by \n instead of just
converting the array to a string.
* Scripts/webkitpy/tool/commands/queues.py:
* Scripts/webkitpy/tool/commands/queues_unittest.py:
2010-10-14 Antonio Gomes <agomes@rim.com>
Reviewed by Martin Robinson and Xan Lopez.
[Gtk]: DRT does not support frame flattening testing
https://bugs.webkit.org/show_bug.cgi?id=38650
Implement DRT's support for toggling on/off frame flattening
support.
* DumpRenderTree/gtk/DumpRenderTree.cpp:
(resetDefaultsToConsistentValues):
* DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
(LayoutTestController::setFrameFlatteningEnabled):
2010-10-15 Dirk Pranke <dpranke@chromium.org>
Reviewed by Eric Siedel.
mocktool.MockOptions is inheriting from Mock, which has the side
effect of defaulting any attribute to another MockObject. So,
MockOptions().foo would always evaluate to true. This was
covering over bugs in the unit tests, and is probably the wrong
default behavior for anything attempting to mock out the options
argument returned from optparse.parse_args().
This patch changes the default behavior. The new MockOptions()
class takes an optional list of keyword parameters to set; this
patch doesn't use that feature but the fix for bug 47510 will.
Also, this patch just fills in the default values necessary to
get all of the tests to pass; I didn't stare at each test by
hand to determine the "right" values. We can either fix that in
subsequent patches or let me know if we want to do that now (and
give me some guidance on what those values might want to be).
https://bugs.webkit.org/show_bug.cgi?id=47709
* Scripts/webkitpy/tool/commands/commandtest.py:
* Scripts/webkitpy/tool/commands/download_unittest.py:
* Scripts/webkitpy/tool/commands/upload_unittest.py:
* Scripts/webkitpy/tool/mocktool.py:
* Scripts/webkitpy/tool/steps/steps_unittest.py:
* Scripts/webkitpy/tool/steps/updatechangelogswithreview_unittest.py:
2010-10-15 Simon Fraser <simon.fraser@apple.com>
Fix the build; need to add new slot to PageUIClient callbacks.
* MiniBrowser/mac/BrowserWindowController.m:
(-[BrowserWindowController awakeFromNib]):
* WebKitTestRunner/TestController.cpp:
(WTR::createOtherPage):
(WTR::TestController::initialize):
2010-10-15 Simon Fraser <simon.fraser@apple.com>
Add Matt Delaney to committers.py.
* Scripts/webkitpy/common/config/committers.py:
2010-10-15 Tony Chang <tony@chromium.org>
Reviewed by Kent Tamura.
[chromium] compile TestNetscapePlugIn on Chromium mac
https://bugs.webkit.org/show_bug.cgi?id=47633
* DumpRenderTree/DumpRenderTree.gypi: files to compile
* DumpRenderTree/TestNetscapePlugIn/main.cpp: Use ifdef because gcc was complaining
(NP_GetEntryPoints):
(NPP_New):
(NPP_Destroy):
(NPP_HandleEvent):
(NPP_GetValue):
* DumpRenderTree/chromium/TestNetscapePlugIn/ForwardingHeaders/WebKit/npapi.h: Added.
* DumpRenderTree/chromium/TestNetscapePlugIn/ForwardingHeaders/WebKit/npfunctions.h: Added.
* DumpRenderTree/chromium/TestNetscapePlugIn/ForwardingHeaders/WebKit/npruntime.h: Added.
2010-10-15 Leandro Pereira <leandro@profusion.mobi>
Reviewed by Gustavo Noronha Silva.
build-webkit: Should die when make fails when build a CMake project
https://bugs.webkit.org/show_bug.cgi?id=47726
* Scripts/webkitdirs.pm: If make fails, die immediately.
2010-10-15 Sheriff Bot <webkit.review.bot@gmail.com>
Unreviewed, rolling out r69809.
http://trac.webkit.org/changeset/69809
https://bugs.webkit.org/show_bug.cgi?id=47725
Broke chromium mac compile (Requested by japhet on #webkit).
* DumpRenderTree/DumpRenderTree.gypi:
* DumpRenderTree/TestNetscapePlugIn/main.cpp:
(NP_GetEntryPoints):
(NPP_New):
(NPP_Destroy):
(NPP_HandleEvent):
(NPP_GetValue):
* DumpRenderTree/chromium/TestNetscapePlugIn/ForwardingHeaders/WebKit/npapi.h: Removed.
* DumpRenderTree/chromium/TestNetscapePlugIn/ForwardingHeaders/WebKit/npfunctions.h: Removed.
* DumpRenderTree/chromium/TestNetscapePlugIn/ForwardingHeaders/WebKit/npruntime.h: Removed.
2010-10-15 Adam Barth <abarth@webkit.org>
Reviewed by Adam Barth.
test-webkitpy fails on Linux
https://bugs.webkit.org/show_bug.cgi?id=47713
The old code failed on Linux because the MacPort tries to read
something out of platform that doesn't make sense on Linux.
* Scripts/webkitpy/tool/mocktool.py:
2010-10-14 Balazs Kelemen <kbalazs@webkit.org>
Reviewed by Kenneth Rohde Christiansen.
[Qt] WTR is sloooow
https://bugs.webkit.org/show_bug.cgi?id=47695
* WebKitTestRunner/qt/TestControllerQt.cpp:
Change the timer interval of RunUntilConditionLoop from
50 milliseconds to 1 to avoid wasting time after the
test had been finished.
2010-10-14 Eric Seidel <eric@webkit.org>
Unreviewed, just fixing an exception seen on the commit-queue.
I should have unit tested this function before.
* Scripts/webkitpy/tool/commands/queues.py:
* Scripts/webkitpy/tool/commands/queues_unittest.py:
2010-10-14 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
commit-queue should not fail patches due to flaky tests
https://bugs.webkit.org/show_bug.cgi?id=47647
This patch makes it so that the *same* flaky test has to fail
twice in a row to have a false negative from a flaky test.
If different flaky tests fail (or if a test fails and then passes
in a second run) then we will warn in the bug that we encountered
a flaky test.
This patch grew to include moving port off of steps onto tool
(which Adam wrote and then I integrated), as well as removing the
use of tool from CommitQueueTask.
* Scripts/webkitpy/common/config/ports.py:
- Added a layout_test_results_path method. This covers old-run-webkit-tests
but doesn't cover NRWT. This is probably not the long term solution, but
putting this knowledge on port makes more sense than in LayoutTestResults.
* Scripts/webkitpy/common/net/buildbot.py:
- LayoutTestResults shouldn't know how to fetch from the network, make
the Build code do that instead.
* Scripts/webkitpy/common/net/buildbot_unittest.py:
- Code style fix.
* Scripts/webkitpy/common/net/layouttestresults.py:
- Remove code for reading from the network.
* Scripts/webkitpy/common/net/layouttestresults_unittest.py:
- Test the new entrypoint.
* Scripts/webkitpy/tool/bot/commitqueuetask.py:
- Make the delegate interface explicit.
- Remove the _tool member, since using the delegate for
everything is cleaner.
- Teach the testing logic how to deal with flaky tests.
* Scripts/webkitpy/tool/bot/commitqueuetask_unittest.py:
- Update to match the CommitQueueTask changes.
* Scripts/webkitpy/tool/commands/queues.py:
- Use the new CommitQueueTaskDelegate interface.
* Scripts/webkitpy/tool/commands/queues_unittest.py:
- Fix the SecondThoughtsCommitQueue test which was broken.
- Add a new test to make sure the flaky test reporting works.
* Scripts/webkitpy/tool/main.py:
- Store the port on the tool object.
* Scripts/webkitpy/tool/mocktool.py:
- Add a port() accessor to MockTool
* Scripts/webkitpy/tool/steps/abstractstep.py:
- Move port() off of Step and onto Tool.
* Scripts/webkitpy/tool/steps/build.py:
* Scripts/webkitpy/tool/steps/preparechangelog.py:
* Scripts/webkitpy/tool/steps/runtests.py:
* Scripts/webkitpy/tool/steps/steps_unittest.py:
- Two tests with the same name! only the latter was being run.
* Scripts/webkitpy/tool/steps/update.py:
2010-10-14 Dirk Pranke <dpranke@chromium.org>
Reviewed by Eric Seidel.
new-run-webkit-tests will now handle missing Ruby installs (or
missing PrettyPatch scripts) more cleanly - previously this
would be detected when we actually tried to create the diff, and
the error message was obscure. Now we'll log a warning up front
and otherwise be silent.
This change also refactors some global variables to be class or
instance variables to be slightly more testable and more
modular. There are no cases where we create lots of port objects
and can't afford to test for configurations repeatedly, so
there's no performance concern here.
https://bugs.webkit.org/show_bug.cgi?id=47466
* Scripts/webkitpy/layout_tests/port/base.py:
* Scripts/webkitpy/layout_tests/port/base_unittest.py:
* Scripts/webkitpy/layout_tests/port/chromium.py:
2010-10-08 Martin Robinson <mrobinson@igalia.com>
Reviewed by Xan Lopez.
[GTK] Form controls do not respect GTK+ font size
https://bugs.webkit.org/show_bug.cgi?id=47134
* DumpRenderTree/gtk/DumpRenderTree.cpp:
(initializeGtkFontSettings): Initialize the font control size when running
DumpRenderTree to a standard value.
2010-10-14 Tony Chang <tony@chromium.org>
Reviewed by Kent Tamura.
[chromium] compile TestNetscapePlugIn on Chromium mac
https://bugs.webkit.org/show_bug.cgi?id=47633
* DumpRenderTree/DumpRenderTree.gypi: files to compile
* DumpRenderTree/TestNetscapePlugIn/main.cpp: Use ifdef because gcc was complaining
(NP_GetEntryPoints):
(NPP_New):
(NPP_Destroy):
(NPP_HandleEvent):
(NPP_GetValue):
* DumpRenderTree/chromium/TestNetscapePlugIn/ForwardingHeaders/WebKit/npapi.h: Added.
* DumpRenderTree/chromium/TestNetscapePlugIn/ForwardingHeaders/WebKit/npfunctions.h: Added.
* DumpRenderTree/chromium/TestNetscapePlugIn/ForwardingHeaders/WebKit/npruntime.h: Added.
2010-10-14 Victor Wang <victorw@chromium.org>
Reviewed by Nate Chapin.
[Chromium] Fix rebaseline_chromium_webkit_tests to use 0 tolerance image diff.
https://bugs.webkit.org/show_bug.cgi?id=47686
* Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
2010-10-14 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
test-webkitpy spews Delegate terminated queue messages
https://bugs.webkit.org/show_bug.cgi?id=47678
* Scripts/webkitpy/tool/bot/queueengine_unittest.py:
2010-10-14 Adam Roben <aroben@apple.com>
Test that pressing the Alt key generates a WM_SYSCOMMAND message
Test for <http://webkit.org/b/47671> <rdar://problem/8435594> Pressing
the Alt key when MiniBrowser's WKView is focused doesn't send focus to
the menu bar
Reviewed by Steve Falkenburg.
* TestWebKitAPI/Configurations/TestWebKitAPICommon.vsprops: Added
TestWebKitAPI/win to the include path so that WindowMessageObserver
can be found.
* TestWebKitAPI/PlatformWebView.h: Added simulateAltKeyPress and
Windows-specific members.
(TestWebKitAPI::PlatformWebView::setParentWindowMessageObserver):
Added this simple setter.
* TestWebKitAPI/Tests/WebKit2/win/AltKeyGeneratesWMSysCommand.cpp: Added.
(TestWebKitAPI::WMSysCommandObserver::WMSysCommandObserver):
Initialize our lone data member.
(TestWebKitAPI::WMSysCommandObserver::windowDidReceiveWMSysCommand):
Simple getter.
(TestWebKitAPI::WMSysCommandObserver::windowReceivedMessage): Record
when a WM_SYSCOMMAND message is received.
(TestWebKitAPI::didNotHandleKeyEventCallback): Record when a
WM_SYSKEYUP message is not handled.
(TestWebKitAPI::WebKit2_AltKeyGeneratesWMSysCommand): Simulate
pressing the Alt key and check that a WM_SYSCOMMAND message got sent
to the WKView's parent window.
* TestWebKitAPI/win/PlatformWebViewWin.cpp:
(TestWebKitAPI::PlatformWebView::registerWindowClass): Made this a
member function so it could access PlatformWebView::wndProc. Use
PlatformWebView::wndProc as the WNDPROC.
(TestWebKitAPI::PlatformWebView::PlatformWebView): Initialize our new
data member and pass the this pointer to CreateWindowEx so we can
store it on the HWND.
(TestWebKitAPI::PlatformWebView::simulateSpacebarKeyPress): Moved some
constants from here to the top of the file.
(TestWebKitAPI::PlatformWebView::simulateAltKeyPress): Added. Sends
the same messages that Notepad receives when you press the Alt key.
(TestWebKitAPI::PlatformWebView::wndProc): Added.
- When WM_CREATE is received, we set the PlatformWebView instance
pointer as a property on the HWND so we can access it later.
- For other messages, we try to get the PlatformWebView instance
pointer from the HWND property.
- When WM_NCDESTROY is received (which is the last message we will
receive), we remove the PlatformWebView instance property.
- Pass messages to the parent window's message observer, if there is
one.
- Pass all messages through to ::DefWindowProcW.
* TestWebKitAPI/win/TestWebKitAPI.vcproj: Added WindowMessageObserver
and AltKeyGeneratesWMSysCommand.
* TestWebKitAPI/win/WindowMessageObserver.h: Added. This class can be
used to observe messages sent to one or more windows.
2010-10-01 Martin Robinson <mrobinson@igalia.com>
Reviewed by Holger Freyther.
[GTK] REGRESSION: FreeType backend does not respect XSettings font settings after r68558
https://bugs.webkit.org/show_bug.cgi?id=47033
* DumpRenderTree/gtk/DumpRenderTree.cpp:
(initializeGtkFontSettings): Added this method which initializes XSettings
font settings to consistent values before running a test. For the one test
in which we need subpixel aliasing turned on, do that.
(initializeFonts): Accepts a testURL parameter now and delegates to initializeGtkFontSettings.
(runTest): Pass the testURL to initializeFonts.
2010-10-14 Adam Barth <abarth@webkit.org>
Reviewed by Eric Seidel.
ChromiumXVFBPort.run_webkit_tests_command has infinite recursion
https://bugs.webkit.org/show_bug.cgi?id=47655
* Scripts/webkitpy/common/config/ports.py:
* Scripts/webkitpy/common/config/ports_unittest.py:
2010-10-13 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
beat diff_parser with the ugly stick
https://bugs.webkit.org/show_bug.cgi?id=47626
* Scripts/webkitpy/common/checkout/diff_parser.py:
* Scripts/webkitpy/style/patchreader.py:
* Scripts/webkitpy/style/patchreader_unittest.py:
2010-10-13 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
Break LayoutTestResults out into its own file
https://bugs.webkit.org/show_bug.cgi?id=47637
* Scripts/webkitpy/common/net/buildbot.py:
* Scripts/webkitpy/common/net/buildbot_unittest.py:
* Scripts/webkitpy/common/net/layouttestresults.py: Added.
* Scripts/webkitpy/common/net/layouttestresults_unittest.py: Added.
2010-10-13 Adam Barth <abarth@webkit.org>
Unreviewed.
Chromium port can't run JavaScriptCore tests
https://bugs.webkit.org/show_bug.cgi?id=47654
This lets webkit-patch build-and-test work on Chromium.
* Scripts/webkitpy/common/config/ports.py:
* Scripts/webkitpy/tool/steps/runtests.py:
2010-10-13 Adam Barth <abarth@webkit.org>
Reviewed by Eric Seidel.
Make --port a global option and pass the port information to the commit-queue subprocess
https://bugs.webkit.org/show_bug.cgi?id=47650
This patch paves the way to run the commit-queue on a non-Mac port.
* Scripts/webkitpy/tool/commands/queues.py:
* Scripts/webkitpy/tool/commands/queues_unittest.py:
* Scripts/webkitpy/tool/commands/queuestest.py:
* Scripts/webkitpy/tool/main.py:
* Scripts/webkitpy/tool/steps/options.py:
* Scripts/webkitpy/tool/steps/preparechangelog.py:
* Scripts/webkitpy/tool/steps/runtests.py:
* Scripts/webkitpy/tool/steps/update.py:
2010-10-13 Adam Barth <abarth@webkit.org>
Reviewed by Eric Seidel.
Introduce the ChromiumXVFBPort for running commit-queue on EC2
https://bugs.webkit.org/show_bug.cgi?id=47653
I'm not entirely sure this is the best way to do this, but we need to
run the tests under XVFB on EC2 because the EC2 instances don't have a
real monitor hooked up. This patch adds a ChromiumXVFBPort that runs
that way. The idea is that XVFB is like a platform for the Chromium
port, but we don't have a real notion of platform separate from port.
* Scripts/webkitpy/common/config/ports.py:
2010-10-13 Sergio Villar Senin <svillar@igalia.com>
Reviewed by Martin Robinson.
[GTK] Add HTTP caching support
https://bugs.webkit.org/show_bug.cgi?id=44261
Add include paths for the new soup HTTP cache code.
* GNUmakefile.am: Add paths for the new soup HTTP cache code.
2010-10-13 Yi Shen <yi.4.shen@nokia.com>
Reviewed by Antonio Gomes.
[Qt] QtTestBrowser shows two Url input fields
https://bugs.webkit.org/show_bug.cgi?id=47613
* QtTestBrowser/mainwindow.cpp:
(MainWindow::MainWindow):
(MainWindow::buildUI):
2010-10-13 Yi Shen <yi.4.shen@nokia.com>
Reviewed by Tony Chang.
[gdb] Add pretty-print supports for UString, Identifier and JSString
https://bugs.webkit.org/show_bug.cgi?id=47601
* gdb/webkit.py:
2010-10-13 Adam Barth <abarth@webkit.org>
Reviewed by Eric Seidel.
Add list of Ubuntu packages needed by the EWS bots
https://bugs.webkit.org/show_bug.cgi?id=47628
I've been copy/pasting this list between bots, but it's better to have
this checked in.
* EWSTools/ubuntu-ews-packages: Added.
2010-10-13 Leandro Pereira <leandro@profusion.mobi>
Reviewed by Csaba Osztrogonác.
[EFL] Adds a build slave.
https://bugs.webkit.org/show_bug.cgi?id=47290
* BuildSlaveSupport/build.webkit.org-config/config.json: Add an
entry for a release build of the EFL port.
* BuildSlaveSupport/build.webkit.org-config/master.cfg: Add platform
flag to build the EFL port.
2010-10-13 Adam Barth <abarth@webkit.org>
Reviewed by Csaba Osztrogonác.
webkit-patch build shouldn't need --build to actually build!
https://bugs.webkit.org/show_bug.cgi?id=47438
Yeah, requiring --build for the build command is really dumb. We did
this originally because the build step is usually optional in other
commands. We don't have a good way of reversing the default for an
option in one command. This approach is slightly hacky since --build
still shows up as an option on the help page, but at least it makes
progress. Passing --build is harmless, so the EWS bots shouldn't
explode because of this change.
* Scripts/webkitpy/tool/bot/commitqueuetask.py:
* Scripts/webkitpy/tool/bot/commitqueuetask_unittest.py:
* Scripts/webkitpy/tool/commands/download.py:
* Scripts/webkitpy/tool/commands/earlywarningsystem.py:
* Scripts/webkitpy/tool/commands/queues_unittest.py:
2010-10-13 Sheriff Bot <webkit.review.bot@gmail.com>
Unreviewed, rolling out r69638.
http://trac.webkit.org/changeset/69638
https://bugs.webkit.org/show_bug.cgi?id=47595
"Broke win and chromium-win bots" (Requested by dglazkov on
#webkit).
* Scripts/webkitpy/common/system/path.py:
* Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
* Scripts/webkitpy/layout_tests/port/base.py:
* Scripts/webkitpy/layout_tests/port/base_unittest.py:
* Scripts/webkitpy/layout_tests/port/chromium.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2010-10-13 John Knottenbelt <jknotten@chromium.org>
Reviewed by Steve Block.
First step towards client-based Geolocation in Chromium. Build
fixes for CLIENT_BASED_GEOLOCATION preprocessor feature define.
https://bugs.webkit.org/show_bug.cgi?id=47586
* DumpRenderTree/chromium/WebViewHost.cpp:
* DumpRenderTree/chromium/WebViewHost.h:
2010-10-13 Adam Roben <aroben@apple.com>
Test that pressing the spacebar in a text field does not scroll the
document
Test for <http://webkit.org/b/47544> <rdar://problem/8540645>
REGRESSION: Pressing spacebar in a text field in WebKit2 does not
insert a space, scrolls the page instead
Reviewed by Sam Weinig.
* TestWebKitAPI/PlatformUtilities.h: Added isKeyDown.
* TestWebKitAPI/PlatformWebView.h: Added simulateSpacebarKeyPress.
* TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
* TestWebKitAPI/win/TestWebKitAPI.vcproj:
* TestWebKitAPI/win/copy-resources.cmd:
Added new files.
* TestWebKitAPI/Tests/WebKit2/SpacebarScrolling.cpp: Added.
(TestWebKitAPI::JavaScriptCallbackContext::JavaScriptCallbackContext):
We use this to track what the result of calling into JavaScript was.
(TestWebKitAPI::didFinishLoadForFrame): Records when the page
finishes loading.
(TestWebKitAPI::didNotHandleKeyEventCallback): Records when a key down
event is not handled.
(TestWebKitAPI::javaScriptCallback): Records that JavaScript finished
executing and whether the result matched our expectation.
(TestWebKitAPI::wk): Turns a UTF-8 C string into a WKStringRef.
(TestWebKitAPI::runJSTest): Calls into JS, waits for the call to
complete, and returns whether we got back the expected result.
(TestWebKitAPI::WebKit2_SpacebarScrolling): Tests that pressing
spacebar inside a text field does not scroll the document and that
pressing it outside the text field does scroll the document.
* TestWebKitAPI/Tests/WebKit2/spacebar-scrolling.html: Added.
* TestWebKitAPI/mac/PlatformUtilitiesMac.mm:
(TestWebKitAPI::Util::isKeyDown): Checks the event's type.
* TestWebKitAPI/mac/PlatformWebViewMac.mm:
(TestWebKitAPI::PlatformWebView::simulateSpacebarKeyPress): Copied
code from DRT's EventSendingController.
* TestWebKitAPI/win/PlatformUtilitiesWin.cpp:
(TestWebKitAPI::Util::isKeyDown): Checks the message's type.
* TestWebKitAPI/win/PlatformWebViewWin.cpp:
(TestWebKitAPI::PlatformWebView::simulateSpacebarKeyPress): Send the
same messages that get sent when you press spacebar in Notepad.
2010-10-13 Adam Roben <aroben@apple.com>
Fix a couple of issues with the TestWebKitAPI build
* TestWebKitAPI/Configurations/TestWebKitAPICFLite.vsprops:
* TestWebKitAPI/Configurations/TestWebKitAPICoreFoundation.vsprops:
Fixed a typo.
* TestWebKitAPI/win/TestWebKitAPIGenerated.vcproj: Use common.vsprops
to set our output and intermediate directories so that we don't spew
files into the source tree.
2010-10-13 Balazs Kelemen <kbalazs@webkit.org>
Reviewed by Csaba Osztrogonác.
WTR should accept relative paths
https://bugs.webkit.org/show_bug.cgi?id=47486
* WebKitTestRunner/StringFunctions.h:
* WebKitTestRunner/TestInvocation.cpp:
(WTR::createWKURL): Moved from StringFunctions.h since it is
used only here. Extend relative paths to absolute.
2010-10-12 Adam Roben <aroben@apple.com>
Make TestWebKitAPI work on Windows
Fixes <http://webkit.org/b/47552> <rdar://problem/8541708>.
Reviewed by Sam Weinig.
* Scripts/build-api-tests: Build TestWebKitAPI.sln on Windows.
* Scripts/run-api-tests: Fix the PATH so that TestWebKitAPI can be
run on Windows.
(runTest): Added code to run TestWebKitAPI.exe on Windows and to die
on unsupported platforms.
(populateTests): Added code to run TestWebKitAPI.exe on Windows and to
die on other platforms. Extracted some formerly-Mac-specific code to
be cross-platform and made it handle any style of line-endings.
* TestWebKitAPI/Configurations/TestWebKitAPICFLite.vsprops: Added.
Links against CFLite.
* TestWebKitAPI/Configurations/TestWebKitAPICommon.vsprops: Added.
Contains most properties for the project.
* TestWebKitAPI/Configurations/TestWebKitAPICoreFoundation.vsprops:
Added. Links against CoreFoundation.
* TestWebKitAPI/TestWebKitAPIPrefix.h: Added a Windows-specific
section and moved the cross-platform section after the
platform-specific parts.
* TestWebKitAPI/win/PlatformUtilitiesWin.cpp: Added.
(TestWebKitAPI::Util::run): Runs a normal message pump until we're
done.
(TestWebKitAPI::Util::cf): Turns a UTF-8 C string into a CFString.
(TestWebKitAPI::Util::createURLForResource): Uses CFBundle to get the
resource path.
(TestWebKitAPI::Util::URLForNonExistentResource): Creates a bogus
WKURL.
* TestWebKitAPI/win/PlatformWebViewWin.cpp: Added.
(TestWebKitAPI::registerWindowClass):
(TestWebKitAPI::PlatformWebView::PlatformWebView):
(TestWebKitAPI::PlatformWebView::~PlatformWebView):
(TestWebKitAPI::PlatformWebView::page):
This was mostly copied from WebKitTestRunner's PlatformWebView.
* TestWebKitAPI/win/TestWebKitAPI.sln: Added. Builds both
TestWebKitAPI and TestWebKitAPIGeneratd. This is used by the
build-api-tests script.
* TestWebKitAPI/win/TestWebKitAPI.vcproj: Added. Builds
TestWebKitAPI.exe.
* TestWebKitAPI/win/TestWebKitAPIGenerated.vcproj: Added. Just calls
through to copy-resources.cmd.
* TestWebKitAPI/win/copy-resources.cmd: Added. Copies resources into
or deletes resources from TestWebKitAPI.resources.
* TestWebKitAPI/win/main.cpp: Added.
(main): Calls through to TestsController.
2010-10-13 Gyuyoung Kim <gyuyoung.kim@samsung.com>
Reviewed by Kenneth Rohde Christiansen.
[EFL] Support viewport configuration and add new arguments for WebKit EFL
https://bugs.webkit.org/show_bug.cgi?id=47084
Opera spec regarding to viewport meta tag was adjusted to WebCore. So, EFL port
needs to be modified according to the changes.
* EWebLauncher/main.c:
(on_viewport_changed):
2010-10-12 Dirk Pranke <dpranke@chromium.org>
Reviewed by Eric Seidel.
This patch enables new-run-webkit-tests (in particular the
chromium-win port) to run under Cygwin as well as Win32. Mostly
this just required some conversions from cygwin paths to Win32
paths when we spawn off Win32 binaries like test_shell.
https://bugs.webkit.org/show_bug.cgi?id=47220
* Scripts/webkitpy/common/system/path.py:
- Expose the cygpath() function for path conversion
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
* Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
- shift filename->uri conversion in the TestInfo objects to the
dump_render_tree thread
* Scripts/webkitpy/layout_tests/port/base.py:
* Scripts/webkitpy/layout_tests/port/chromium.py:
- use cygpath()
2010-10-12 Yuta Kitamura <yutak@chromium.org>
Unreviewed. Add Yuta Kitamura (yutak) to the committers list.
* Scripts/webkitpy/common/config/committers.py:
2010-10-12 Sam Weinig <sam@webkit.org>
Reviewed by Dan Bernstein.
When a provisional load fails, the provisional URL returned
from WKFrameCopyProvisionalURL should be empty
<rdar://problem/8540878>
https://bugs.webkit.org/show_bug.cgi?id=47546
Add test.
* TestWebKitAPI/Tests/WebKit2/FailedLoad.cpp:
(TestWebKitAPI::didFailProvisionalLoadWithErrorForFrame):
2010-10-12 Anders Carlsson <andersca@apple.com>
Reviewed by Sam Weinig.
Start fleshing out find page overlays
https://bugs.webkit.org/show_bug.cgi?id=47559
* MiniBrowser/mac/BrowserWindowController.m:
(-[BrowserWindowController find:]):
Pass kWKFindOptionsShowOverlay to WKPageFindString.
2010-10-12 Robert Hogan <robert@webkit.org>
Reviewed by Antonio Gomes.
[Qt] editing/input/emacs-ctrl-o.html
ctrl-o is bound to a sequence of commands on Mac, namely:
insertParagraphSeparator then moveLeft. In the DRT, we just interpret
ctrl-o in EventSenderQt as '\n', i.e. insertParagraphSeparator. As a
result we only issue one command, so don't generate the change in caret
position that results in the extra editing delegate messages.
So for the sake of not managing suspicious-looking differences, just
make DRT issue the appropriate edit commands to pass the test.
https://bugs.webkit.org/show_bug.cgi?id=47548
* DumpRenderTree/qt/EventSenderQt.cpp:
(EventSender::keyDown):
2010-10-12 Anders Carlsson <andersca@apple.com>
Reviewed by Sam Weinig.
Add a simple Find UI to MiniBrowser
https://bugs.webkit.org/show_bug.cgi?id=47553
* MiniBrowser/mac/BrowserWindow.xib:
Add Find panel.
* MiniBrowser/mac/BrowserWindowController.h:
* MiniBrowser/mac/BrowserWindowController.m:
(-[BrowserWindowController performFindPanelAction:]):
Show the find panel.
(-[BrowserWindowController find:]):
Tell the WKPageRef to find the given string.
2010-10-12 Tony Chang <tony@chromium.org>
Unreviewed, fixing DRT compile on Windows after r69586.
* DumpRenderTree/chromium/WebThemeEngineDRT.h:
2010-10-12 Tony Chang <tony@chromium.org>
Unreviewed, trying to fix NRWT on Windows.
* Scripts/webkitpy/layout_tests/port/http_lock.py:
2010-10-12 Balazs Kelemen <kbalazs@webkit.org>
Reviewed by Kenneth Rohde Christiansen.
[Qt][WTR] Checking the path of the injected bundle is bogus
https://bugs.webkit.org/show_bug.cgi?id=47541
* WebKitTestRunner/qt/TestControllerQt.cpp:
(WTR::isExistingLibrary): Added (static helper).
Extend the path with the appropriate suffix(es) on the
platform and check that the file exists.
(WTR::TestController::initializeInjectedBundlePath):
There were two bugs with the path checking:
- the condition should have been inverted
- QLibrary::fileName does not give back the filename
with the library suffix so we cannot check
that existance of the file this way.
2010-10-12 Gabor Rapcsanyi <rgabor@inf.u-szeged.hu>
Reviewed by Tony Chang.
Implement http locking in NRWT.
https://bugs.webkit.org/show_bug.cgi?id=47072
* Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
* Scripts/webkitpy/layout_tests/port/base.py:
* Scripts/webkitpy/layout_tests/port/http_lock.py: Added.
* Scripts/webkitpy/layout_tests/port/http_lock_unittest.py: Added.
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2010-10-12 Mark Rowe <mrowe@apple.com>
Fix the 32-bit WebKit2 build.
* MiniBrowser/mac/BrowserWindowController.m:
(getWindowFrame): Use NSRect rather than CGRect since that's what -frame returns.
(setWindowFrame): Use NSMakeRect rather than CGRectMake since that's what -setFrame:display: expects.
2010-10-11 Tony Chang <tony@chromium.org>
Reviewed by Adam Barth.
change the order of mrobinson's email addresses so bugs.webkit.org suggests the right one
https://bugs.webkit.org/show_bug.cgi?id=47513
bugs.webkit.org suggests mrobinson@igalia.com as an autocomplete,
which doesn't actually work.
* Scripts/webkitpy/common/config/committers.py:
2010-10-11 Prasad Tammana <prasadt@chromium.org>
Reviewed by Dmitry Titov.
Added support for showModalDialog on chromium port.
Implement showModalDialog for Layout tests for chromium port
https://bugs.webkit.org/show_bug.cgi?id=46759
* DumpRenderTree/chromium/WebViewHost.cpp:
(WebViewHost::closeWidgetSoon): Quit current message loop if in a modal loop.
(WebViewHost::runModal): Start a nested message loop and remember that.
(WebViewHost::WebViewHost): Initialize m_inModalLoop.
* DumpRenderTree/chromium/WebViewHost.h: Add m_inModalLoop.
2010-10-11 Martin Robinson <mrobinson@igalia.com>
Reviewed by Gustavo Noronha Silva.
[GTK] Cleanup font selection code for the Freetype backend
https://bugs.webkit.org/show_bug.cgi?id=47503
Update DRT to set values for the cursive and fantasy font families, so
that we can generate consistent results for tests that use these families.
Add some missing information to fonts.conf.
* DumpRenderTree/gtk/DumpRenderTree.cpp:
(resetDefaultsToConsistentValues): Always set the fantasy and cursive font
family names.
* DumpRenderTree/gtk/fonts/fonts.conf: Add aliases for cursive and fantasy fonts
to our default serif font. Eventually we may want to import or require some "real"
fonts, but this should at least ensure consistent rendering during test runs. Also
add aliases for some other commonly used fonts in tests (Arial and Lucida Grande).
Finally, add a missing oblique specification for DejaVu Serif, which was resulting
in some incorrect baselines for the synthetic oblique test.
2010-10-11 Adam Barth <abarth@webkit.org>
Unreviewed.
Fix sheriffbot not to crash when there are failing tests. It turns out
Python sets don't support the + operator. We need to use union
instead. There's a more elegant way to do this in Python 2.6, but we
need this code to work in 2.5.
* Scripts/webkitpy/common/net/failuremap.py:
* Scripts/webkitpy/common/net/failuremap_unittest.py:
2010-10-10 Robert Hogan <robert@webkit.org>
Unreviewed, fix failing test from r69468.
[Qt] Put all DRT-created pages in a page group
Fix on r69468.
Because we don't delete closed pages immediately in DRT we need
to remove them from the page group explicitly instead.
Fixes failure of fast/events/popup-blocked-from-fake-user-gesture.html
https://bugs.webkit.org/show_bug.cgi?id=47469
* DumpRenderTree/qt/DumpRenderTreeQt.cpp:
(WebCore::DumpRenderTree::windowCloseRequested):
2010-10-10 Leandro Pereira <leandro@profusion.mobi>
Reviewed by Adam Barth.
commit-queue: Add EFL-EWS status bubble to Bugzilla, now that the
EFL-EWS bot is up and running.
https://bugs.webkit.org/show_bug.cgi?id=47277
* QueueStatusServer/handlers/statusbubble.py:
2010-10-10 Robert Hogan <robert@webkit.org>
Reviewed by Antonio Gomes.
[Qt] Put all DRT-created pages in a page group
https://bugs.webkit.org/show_bug.cgi?id=47469
* DumpRenderTree/qt/DumpRenderTreeQt.cpp:
(WebCore::DumpRenderTree::DumpRenderTree):
(WebCore::DumpRenderTree::createWindow):
2010-10-09 Sam Weinig <sam@webkit.org>
Reviewed by Dan Bernstein.
WKFrameGetFrameLoadState() returns kWKFrameLoadStateCommitted after the load has been stopped
<rdar://problem/8173667>
https://bugs.webkit.org/show_bug.cgi?id=47461
* TestWebKitAPI/PlatformUtilities.h:
* TestWebKitAPI/Test.h:
* TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
* TestWebKitAPI/Tests/WebKit2/FailedLoad.cpp: Added.
(TestWebKitAPI::didFailProvisionalLoadWithErrorForFrame):
(TestWebKitAPI::TEST):
* TestWebKitAPI/mac/PlatformUtilitiesMac.mm:
(TestWebKitAPI::Util::URLForNonExistentResource):
Adds a test for the frame load state after a failed provisional load. I wanted to test
the frame load state after a committed load failed, but I don't believe that is possible
to do without hooking up the http server to serve a long loading page.
2010-10-09 Sam Weinig <sam@webkit.org>
Reviewed by Dan Bernstein.
Add WebKit2 version of runBeforeUnloadConfirmPanelWithMessage
<rdar://problem/8447690>
https://bugs.webkit.org/show_bug.cgi?id=47459
* MiniBrowser/mac/BrowserWindowController.m:
(runBeforeUnloadConfirmPanel):
(-[BrowserWindowController awakeFromNib]):
* WebKitTestRunner/TestController.cpp:
(WTR::createOtherPage):
(WTR::TestController::initialize):
2010-10-09 Sam Weinig <sam@webkit.org>
Reviewed by Dan Bernstein.
Need implementation of ChromeClient windowRect related functions.
<rdar://problem/8469476>
https://bugs.webkit.org/show_bug.cgi?id=47386
* MiniBrowser/mac/BrowserWindowController.m:
(getWindowFrame):
(setWindowFrame):
(-[BrowserWindowController awakeFromNib]):
* WebKitTestRunner/PlatformWebView.h:
* WebKitTestRunner/TestController.cpp:
(WTR::TestController::initialize):
* WebKitTestRunner/mac/PlatformWebViewMac.mm:
(WTR::PlatformWebView::windowFrame):
(WTR::PlatformWebView::setWindowFrame):
* WebKitTestRunner/qt/PlatformWebViewQt.cpp:
(WTR::PlatformWebView::windowFrame):
(WTR::PlatformWebView::setWindowFrame):
* WebKitTestRunner/win/PlatformWebViewWin.cpp:
(WTR::PlatformWebView::windowFrame):
(WTR::PlatformWebView::setWindowFrame):
2010-10-08 Adam Barth <abarth@webkit.org>
Unreviewed (Eric is on vacation in this change is trivial).
Add a "clean" command to webkit-patch to clean the working directory
https://bugs.webkit.org/show_bug.cgi?id=47436
This command is useful when using SVN because "svn revert" leaves
unversioned files behind.
* Scripts/webkitpy/tool/commands/download.py:
2010-10-08 Andras Becsi <abecsi@webkit.org>
Unreviewed build fix.
[Qt] Add missing limits.h include since the Qt version currently
ran on the bot does need that.
* WebKitTestRunner/InjectedBundle/qt/ActivateFontsQt.cpp:
2010-10-08 Balazs Kelemen <kbalazs@webkit.org>
Reviewed by Csaba Osztrogonác.
[Qt] Turn on building WTR
https://bugs.webkit.org/show_bug.cgi?id=47349
* Scripts/webkitdirs.pm:
2010-10-08 Balazs Kelemen <kbalazs@webkit.org>
Reviewed by Andreas Kling.
[Qt] TestController needs its platform dependent methods
https://bugs.webkit.org/show_bug.cgi?id=47413
* WebKitTestRunner/qt/TestControllerQt.cpp:
(WTR::TestController::initializeInjectedBundlePath):
If the WTR_INJECTEDBUNDLE_PATH environmental variable is
set then use that otherwise use the path where the bundle
lives in a normal trunk build.
(WTR::TestController::initializeTestPluginDirectory):
Set the bundle's value. It is not used currently.
(WTR::TestController::platformInitializeContext):
* WebKitTestRunner/qt/WebKitTestRunner.pro:
2010-10-08 Jocelyn Turcotte <jocelyn.turcotte@nokia.com>
Reviewed by Kenneth Rohde Christiansen.
WTR: Prevent sending the Done message twice on test timeout.
https://bugs.webkit.org/show_bug.cgi?id=47410
When InjectedBundle::done() calls stopLoading, this may trigger
InjectedBundlePage::didFailLoadWithErrorForFrame which calls
InjectedBundle::done() itself later in the stack.
This would output the timeout failure message twice and confuse
the run-webkit-tests script.
This patch adds a third state, Stopping, which prevents WebCore
errors to trigger done() when testing is over.
* WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
(WTR::InjectedBundle::done):
* WebKitTestRunner/InjectedBundle/InjectedBundle.h:
2010-10-08 Andras Becsi <abecsi@webkit.org>
Reviewed by Kenneth Rohde Christiansen.
[Qt] Add an activateFonts() implementation similar to DRT's initializeFonts().
Original code by Simon Hausmann.
https://bugs.webkit.org/show_bug.cgi?id=47402
* WebKitTestRunner/InjectedBundle/qt/ActivateFontsQt.cpp: Added.
(WTR::activateFonts):
2010-10-08 Andras Becsi <abecsi@webkit.org>
Reviewed by Csaba Osztrogonác.
[Qt] Add support for Qt's WebKitTestRunner to old-run-webkit-tests.
https://bugs.webkit.org/show_bug.cgi?id=47401
* Scripts/build-webkittestrunner:
* Scripts/old-run-webkit-tests:
2010-10-07 Jocelyn Turcotte <jocelyn.turcotte@nokia.com>
Reviewed by Kenneth Rohde Christiansen.
[Qt] WTR first compile: fix compile issues.
https://bugs.webkit.org/show_bug.cgi?id=47343
* WebKitTestRunner/InjectedBundle/Bindings/JSWrappable.h:
* WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
* WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
* WebKitTestRunner/InjectedBundle/LayoutTestController.h:
* WebKitTestRunner/TestController.cpp:
* WebKitTestRunner/TestInvocation.cpp:
2010-10-07 Antonio Gomes <agomes@rim.com>
Reviewed by Simon Fraser.
[Mac] [DRT] implement setSpatialNavigationEnabled
https://bugs.webkit.org/show_bug.cgi?id=47291
Implemented LayoutTestController::setSpatialNavigationEnabled for Mac's DRT
so it can track regression on the existing implementation and future improvements
we are making.
* DumpRenderTree/mac/LayoutTestControllerMac.mm:
(LayoutTestController::setSpatialNavigationEnabled):
2010-10-07 Dirk Pranke <dpranke@chromium.org>
Reviewed by Adam Barth.
rebaseline-chromium-webkit-tests relied on the filename_to_uri()
hook in the Port infrastructure to generate URIs for the files
in its summary HTML report; however, that method is supposed to only
be used for test files (and should really be renamed), so this would
crash.
This change adds a new "path" module to the system package with a
routine called abspath_to_uri() that handles converting paths to
file: URIs independently of anything in the layout_tests package,
and changes the code to use this. At some point in the near future
the layout_tests/port/* code should use this as well.
This change also deletes a bunch of unused code and fixes some
comments in rebaseline_chromium_webkit_tests.py.
https://bugs.webkit.org/show_bug.cgi?id=47319
* Scripts/webkitpy/common/system/path.py: Added.
* Scripts/webkitpy/common/system/path_unittest.py: Added.
* Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
* Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests_unittest.py:
2010-10-07 Martin Robinson <mrobinson@igalia.com>
Reviewed by Gustavo Noronha Silva.
[GTK] The FreeType backend does not respect the FC_EMBOLDEN property
https://bugs.webkit.org/show_bug.cgi?id=46216
Fix my latest commit, as it contained some bad changes from a merge gone wrong.
* DumpRenderTree/gtk/DumpRenderTree.cpp:
(initializeFonts): Also initialize the DevaVu Sans font.
2010-10-07 Daniel Cheng <dcheng@chromium.org>
Reviewed by Tony Chang.
[chromium] Minor naming cleanup in WebDragData, part 2
https://bugs.webkit.org/show_bug.cgi?id=47227
Update DRT to use the renamed methods.
* DumpRenderTree/chromium/EventSender.cpp:
(EventSender::beginDragWithFiles):
* DumpRenderTree/chromium/WebViewHost.cpp:
(addDRTFakeFileToDataObject):
2010-10-07 Anders Carlsson <andersca@apple.com>
Reviewed by Sam Weinig.
Add a WKPageFindClient, hook up WKPageCountStringMatches
https://bugs.webkit.org/show_bug.cgi?id=47373
* TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
Remove hard coded paths.
* TestWebKitAPI/Tests/WTF/VectorBasic.cpp:
(TestWebKitAPI::TEST):
Fix expected result.
* TestWebKitAPI/Tests/WebKit2/Find.cpp: Added.
(TestWebKitAPI::didFinishLoadForFrame):
(TestWebKitAPI::didCountStringMatches):
(TestWebKitAPI::TEST):
* TestWebKitAPI/Tests/WebKit2/find.html: Added.
Add test for WKPageCountStringMatches.
2010-09-21 Martin Robinson <mrobinson@igalia.com>
Reviewed by Chris Fleizach.
[GTK] The FreeType backend does not respect the FC_EMBOLDEN property
https://bugs.webkit.org/show_bug.cgi?id=46216
* DumpRenderTree/gtk/fonts/fonts.conf: Add a setting which enables FC_EMBOLDEN
for DejaVu Serif when there is no bold version available. This is necessary to
properly test the property in layout tests.
2010-10-07 Jocelyn Turcotte <jocelyn.turcotte@nokia.com>, Andras Becsi <abecsi@webkit.org>, Balazs Kelemen <kbalazs@webkit.org>
Reviewed by Kenneth Rohde Christiansen.
[Qt] Add WTR's InjectedBundle build files.
https://bugs.webkit.org/show_bug.cgi?id=47333
* Scripts/webkitdirs.pm:
* WebKitTestRunner/DerivedSources.pro: Added.
* WebKitTestRunner/InjectedBundle/qt/InjectedBundle.pro: Added.
* WebKitTestRunner/WebKitTestRunner.pro: Added.
* WebKitTestRunner/qt/DerivedSources.pro: Removed.
Content merged in ../DerivedSources.pro
2010-10-07 Sheriff Bot <webkit.review.bot@gmail.com>
Unreviewed, rolling out r69315.
http://trac.webkit.org/changeset/69315
https://bugs.webkit.org/show_bug.cgi?id=47363
Forgot to add the new files (Requested by kbalazs on #webkit).
* Scripts/webkitdirs.pm:
* WebKitTestRunner/qt/DerivedSources.pro: Added.
2010-10-07 Sam Weinig <sam@webkit.org>
Reviewed by Anders Carlsson.
Change API tester to ask the executable for the list of all
tests instead of relying on the directory structure.
https://bugs.webkit.org/show_bug.cgi?id=47359
* Scripts/run-test-webkit-api: Added.
* Scripts/webkitdirs.pm:
Add additional script to just launch the api tester with the correct settings.
* Scripts/run-api-tests:
Change to use the new --dump-tests options to build the list of tests and change
output to reflect a (suite, testcase) tuple for each test. Also adds some color.
* TestWebKitAPI/Test.h:
(TestWebKitAPI::Test::Register::Register):
* TestWebKitAPI/Tests/WTF/VectorBasic.cpp:
(TestWebKitAPI::TEST):
* TestWebKitAPI/Tests/WebKit2/FrameMIMETypeHTML.cpp:
(TestWebKitAPI::TEST):
* TestWebKitAPI/Tests/WebKit2/FrameMIMETypePNG.cpp:
(TestWebKitAPI::TEST):
* TestWebKitAPI/Tests/WebKit2/PageLoadBasic.cpp:
(TestWebKitAPI::TEST):
* TestWebKitAPI/Tests/WebKit2/WKString.cpp:
(TestWebKitAPI::TEST):
* TestWebKitAPI/Tests/WebKit2/WKStringJSString.cpp:
(TestWebKitAPI::TEST):
Give each test a suite name in addition to the test name.
* TestWebKitAPI/TestsController.cpp:
(TestWebKitAPI::TestsController::dumpTestNames):
(TestWebKitAPI::TestsController::runTestNamed):
* TestWebKitAPI/TestsController.h:
* TestWebKitAPI/mac/main.mm:
(main):
Add option to print all registered tests.
2010-10-07 Jocelyn Turcotte <jocelyn.turcotte@nokia.com>, Andras Becsi <abecsi@webkit.org>, Balazs Kelemen <kbalazs@webkit.org>
Reviewed by Kenneth Rohde Christiansen.
[Qt] Add WTR's InjectedBundle build files.
https://bugs.webkit.org/show_bug.cgi?id=47333
* Scripts/webkitdirs.pm:
* WebKitTestRunner/DerivedSources.pro: Added.
* WebKitTestRunner/InjectedBundle/qt/InjectedBundle.pro: Added.
* WebKitTestRunner/WebKitTestRunner.pro: Added.
* WebKitTestRunner/qt/DerivedSources.pro: Removed.
Content merged in ../DerivedSources.pro
2010-10-07 Balazs Kelemen <kbalazs@webkit.org>
Unreviewed. Do a renaming that was recommended by the reviewer
(Kenneth) but I forgot to do before landing (http://trac.webkit.org/changeset/69253)
* WebKitTestRunner/qt/TestControllerQt.cpp:
(WTR::RunUntilConditionLoop::start):
(WTR::RunUntilConditionLoop::RunUntilConditionLoop):
(WTR::TestController::runUntil):
2010-10-07 Balazs Kelemen <kbalazs@webkit.org>
Reviewed by Kenneth Rohde Christiansen.
[Qt] Watchdog timer implementation for WTR
https://bugs.webkit.org/show_bug.cgi?id=47337
* WebKitTestRunner/InjectedBundle/LayoutTestController.h: Factor out
the definition of the timer type to a typedef.
* WebKitTestRunner/InjectedBundle/qt/LayoutTestControllerQt.cpp: Added.
Timer implementation with QTimer and a helper QObject class.
(WTR::WatchdogTimerHelper::instance):
(WTR::WatchdogTimerHelper::timerFired):
(WTR::WatchdogTimerHelper::WatchdogTimerHelper):
(WTR::LayoutTestController::platformInitialize):
(WTR::LayoutTestController::invalidateWaitToDumpWatchdogTimer):
(WTR::LayoutTestController::initializeWaitToDumpWatchdogTimerIfNeeded):
2010-10-07 Balazs Kelemen <kbalazs@webkit.org>
Unreviewed buildfix for 69297 again
* WebKitTestRunner/TestController.cpp:
(WTR::blankURL):
2010-10-07 Balazs Kelemen <Balazs Kelemen>
Unreviewed trivial build fix for r69297
* WebKitTestRunner/TestController.cpp:
(WTR::blankURL):
2010-10-07 Balazs Kelemen <kbalazs@webkit.org>
Reviewed by Kenneth Rohde Christiansen.
WebKitTestRunner should be portable
https://bugs.webkit.org/show_bug.cgi?id=45393
Use only the WebKit API for working with urls.
* WebKitTestRunner/StringFunctions.h:
(WTR::createWKURL):
* WebKitTestRunner/TestController.cpp:
(WTR::blankURL):
(WTR::TestController::resetStateToConsistentValues):
(WTR::TestController::didFinishLoadForFrame):
2010-10-07 Carlos Garcia Campos <cgarcia@igalia.com>
Reviewed by Xan Lopez.
[GTK] Use gtk_widget_draw() instead of gtk_widget_get_snapshot() in PixelDumpSupportGtk when building with gtk3
https://bugs.webkit.org/show_bug.cgi?id=47332
* DumpRenderTree/gtk/PixelDumpSupportGtk.cpp:
(createBitmapContextFromWebView):
2010-10-06 Chris Guillory <chris.guillory@google.com>
Reviewed by Chris Fleizach.
Remove unused accessibility functions from webkit.
https://bugs.webkit.org/attachment.cgi?bugid=46707
* DumpRenderTree/chromium/WebViewHost.cpp:
(WebViewHost::postAccessibilityNotification):
* DumpRenderTree/chromium/WebViewHost.h:
2010-10-06 Lucas Forschler <lforschler@apple.com>
Reviewed by Adam Roben.
Fix the sunspider-compare-results to actually use the passed in value for $root.
* Scripts/sunspider-compare-results:
2010-10-06 Balazs Kelemen <kbalazs@webkit.org>
Reviewed by Kenneth Rohde Christiansen.
[Qt] Runloop implementation for WTR
https://bugs.webkit.org/show_bug.cgi?id=47280
* WebKitTestRunner/qt/TestControllerQt.cpp:
Implemented TestController::runUntil by a timerEvent
and a QEventLoop. We step into the event loop from runUntil.
While we are waiting in the loop a timerEvent is periodically
checking the value of the condition. Once the condition has
becoming true the timerEvent wakes us up.
(WTR::RunUntilLoop::start):
(WTR::RunUntilLoop::RunUntilLoop):
(WTR::RunUntilLoop::run):
(WTR::RunUntilLoop::timerEvent):
(WTR::TestController::platformInitialize):
(WTR::TestController::runUntil):
* WebKitTestRunner/qt/main.cpp:
Start the main event loop first and creating the TestController later.
(Launcher::Launcher):
(Launcher::~Launcher):
(Launcher::launch): Creating the TestController.
(main): Setting up a timer for calling Launcher::launch from
the main event loop.
2010-10-06 Dirk Pranke <dpranke@chromium.org>
Reviewed by Kenneth Russell.
new-run-webkit-tests: fix typo in chromium-gpu that was trying to
enable 'accelerated-composting' instead of 'accelerated-compositing'.
https://bugs.webkit.org/show_bug.cgi?id=47312
* Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
* Scripts/webkitpy/layout_tests/port/chromium_gpu_unittest.py:
2010-10-06 Balazs Kelemen <kbalazs@webkit.org>
Reviewed by Kenneth Rohde Christiansen.
[Qt] PlatformWebView implementation for WebKitTestRunner
https://bugs.webkit.org/show_bug.cgi?id=47276
The implementation follows the way how
we use the API in MiniBrowser.
* WebKitTestRunner/PlatformWebView.h:
* WebKitTestRunner/qt/PlatformWebViewQt.cpp:
(WTR::WebView::wkView):
(WTR::WebView::~WebView):
(WTR::WebView::WebView):
(WTR::PlatformWebView::PlatformWebView):
(WTR::PlatformWebView::~PlatformWebView):
(WTR::PlatformWebView::resizeTo):
(WTR::PlatformWebView::page):
(WTR::PlatformWebView::focus):
* WebKitTestRunner/qt/WebKitTestRunner.pro:
2010-10-06 Balazs Kelemen <kbalazs@webkit.org>
Reviewed by Kenneth Rohde Christiansen.
[Qt] Add WebKitTestRunner's build files
https://bugs.webkit.org/show_bug.cgi?id=44155
Parts was taken by Zoltan Horvath's patch.
* Scripts/generate-forwarding-headers.pl: Moved from WebKit2.
* Scripts/webkitdirs.pm: Make the generated files needed by WTR.
* WebKitTestRunner/PlatformWebView.h: Addeed typedefs for
PlatformWKView and PlatformWindow (void* for now).
Buildfix the case when __APPLE__ is not defined.
* WebKitTestRunner/qt/DerivedSources.pro: Added.
* WebKitTestRunner/qt/PlatformWebViewQt.cpp: Added.
Empty stub implementation.
(WTR::registerWindowClass):
(WTR::PlatformWebView::PlatformWebView):
(WTR::PlatformWebView::~PlatformWebView):
(WTR::PlatformWebView::resizeTo):
(WTR::PlatformWebView::page):
(WTR::PlatformWebView::focus):
* WebKitTestRunner/qt/TestControllerQt.cpp: Added.
Empty stub implementation.
(WTR::registerWindowClass):
(WTR::TestController::runUntil):
(WTR::TestController::platformInitialize):
(WTR::TestController::initializeInjectedBundlePath):
(WTR::TestController::initializeTestPluginDirectory):
(WTR::TestController::platformInitializeContext):
* WebKitTestRunner/qt/WebKitTestRunner.pro: Added.
* WebKitTestRunner/qt/main.cpp: Added.
(main):
2010-10-06 Tony Chang <tony@chromium.org>
Unreviewed, rolling out r69202.
http://trac.webkit.org/changeset/69202
https://bugs.webkit.org/show_bug.cgi?id=46937
Broke compile of test_shell
* DumpRenderTree/chromium/EventSender.cpp:
(EventSender::beginDragWithFiles):
* DumpRenderTree/chromium/WebViewHost.cpp:
(addDRTFakeFileToDataObject):
2010-10-06 Daniel Cheng <dcheng@chromium.org>
Reviewed by Tony Chang.
[chromium] Minor naming cleanup in WebDragData, part 2
https://bugs.webkit.org/show_bug.cgi?id=46937
Update DRT to use the renamed methods.
* DumpRenderTree/chromium/EventSender.cpp:
(EventSender::beginDragWithFiles):
* DumpRenderTree/chromium/WebViewHost.cpp:
(addDRTFakeFileToDataObject):
2010-10-06 David Kilzer <ddkilzer@apple.com>
<http://webkit.org/b/47270> Move WebArchive serialization code into its own file
Reviewed by Adam Roben.
This is the first step in making webarchive tests work on
Windows.
* DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj: Added
new files to the project.
* DumpRenderTree/mac/DumpRenderTree.mm: Removed code that moved
to WebArchiveDumpSupport.mm.
* DumpRenderTree/mac/WebArchiveDumpSupport.h: Added.
* DumpRenderTree/mac/WebArchiveDumpSupport.mm: Copied from WebKitTools/DumpRenderTree/mac/DumpRenderTree.mm.
(serializeWebArchiveToXML):
2010-10-05 Daniel Bates <dbates@rim.com>
Reviewed by David Kilzer.
Add infrastructure to towards detecting change log diffs that aren't at the top of the ChangeLog
https://bugs.webkit.org/show_bug.cgi?id=46058
Make VCSUtils::fixChangeLogPatch() return a reference to a hash
structure so as to support returning additional information
about a change log diff.
Currently, VCSUtils::fixChangeLogPatch() returns a string that
represents the change log diff. Towards supporting the return
of additional information, such as whether the change log diff
inserts an entry at the top of the ChangeLog file, we need to
make VCSUtils::fixChangeLogPatch() return a reference to hash
structure.
* Scripts/VCSUtils.pm:
- Modified fixChangeLogPatch() to return a reference to a
hash structure.
- Added documentation to fixChangeLogPatch().
- Modified call site in mergeChangeLogs() as necessary.
* Scripts/svn-apply:
- Modified call site in patch() as necessary.
* Scripts/svn-create-patch:
- Modified call site in generateDiff() as necessary.
* Scripts/svn-unapply:
- Modified call site in patch() as necessary.
* Scripts/webkitperl/VCSUtils_unittest/fixChangeLogPatch.pl:
- Modified the unit tests as necessary.
2010-10-05 Tony Chang <tony@chromium.org>
Reviewed by Darin Adler.
fix the link to the expected image on windows
https://bugs.webkit.org/show_bug.cgi?id=47228
* Scripts/webkitpy/layout_tests/test_types/test_type_base.py: On
windows, the file must be opened in binary mode when writing
binary data.
2010-10-05 Adam Barth <abarth@webkit.org>
Update expected result of unittest to match Tony's change below.
* Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
2010-10-05 Tony Chang <tony@chromium.org>
Reviewed by Ojan Vafai (over the shoulder).
Paper over errors in image_diff so we don't crash the whole test run.
* Scripts/webkitpy/layout_tests/port/chromium.py:
2010-10-05 Tony Chang <tony@chromium.org>
Unreviewed, make shutil.rmtree more resiliant to errors
since windows was raising WindowsError: The process cannot access the file because it
is being used by another process.
* Scripts/webkitpy/layout_tests/port/chromium.py:
2010-10-05 Tony Chang <tony@chromium.org>
Reviewed by Ojan Vafai.
[chromium] fix image diffing in NRWT
https://bugs.webkit.org/show_bug.cgi?id=47128
* Scripts/webkitpy/layout_tests/port/chromium.py: Stop using
NamedTemporaryFile since it doesn't work on Windows.
* Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
2010-10-05 Kenneth Russell <kbr@google.com>
Reviewed by Tony Chang.
[chromium] Implement layerTreeAsText in DumpRenderTree
https://bugs.webkit.org/show_bug.cgi?id=47216
Plumbed Frame::layerTreeAsText through Chromium's WebKit API to
make it callable from DumpRenderTree.
No new tests; verified with existing compositor layout tests.
* DumpRenderTree/chromium/LayoutTestController.cpp:
(LayoutTestController::LayoutTestController):
(LayoutTestController::layerTreeAsText):
* DumpRenderTree/chromium/LayoutTestController.h:
2010-10-05 Brent Fulgham <bfulgham@webkit.org>
Unreviewed build correction.
* DumpRenderTree/win/ImageDiff.vcproj: Use WinCairo debug
property sheet so proper libraries are linked.
2010-10-05 Robert Hogan <robert@webkit.org>
Reviewed by Andreas Kling.
[Qt] Fix url conversion in QWebHistory
Converting from KURL to WFT::String to QUrl does not
permit proper percent encoding later.
https://bugs.webkit.org/show_bug.cgi?id=47048
* DumpRenderTree/qt/DumpRenderTreeQt.cpp:
(WebCore::dumpHistoryItem):
2010-10-05 Anders Carlsson <andersca@apple.com>
Reviewed by Sam Weinig.
REGRESSION (r68966?): All dumpAsText test fail on WebKit2
https://bugs.webkit.org/show_bug.cgi?id=47188
<rdar://problem/8514104>
When constructing the std::string, use the real string length instead of the maximum buffer size.
* TestWebKitAPI/PlatformUtilities.h:
(TestWebKitAPI::Util::toSTD):
* WebKitTestRunner/StringFunctions.h:
(WTR::toSTD):
2010-10-04 Sam Weinig <sam@webkit.org>
Reviewed by Mark Rowe.
Add MIMEType accessor to WKFrame
<rdar://problem/8347683>
https://bugs.webkit.org/show_bug.cgi?id=47138
* TestWebKitAPI/PlatformUtilities.h:
(TestWebKitAPI::Util::toSTD):
* TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
* TestWebKitAPI/Tests/WebKit2/FrameMIMETypeHTML.cpp: Added.
(TestWebKitAPI::didStartProvisionalLoadForFrame):
(TestWebKitAPI::didCommitLoadForFrame):
(TestWebKitAPI::didFinishLoadForFrame):
(TestWebKitAPI::TEST):
* TestWebKitAPI/Tests/WebKit2/FrameMIMETypePNG.cpp: Added.
(TestWebKitAPI::didStartProvisionalLoadForFrame):
(TestWebKitAPI::didCommitLoadForFrame):
(TestWebKitAPI::didFinishLoadForFrame):
(TestWebKitAPI::TEST):
* TestWebKitAPI/Tests/WebKit2/PageLoadBasic.cpp:
(TestWebKitAPI::TEST):
* TestWebKitAPI/Tests/WebKit2/basic-1.html: Removed.
* TestWebKitAPI/Tests/WebKit2/icon.png: Added.
* TestWebKitAPI/Tests/WebKit2/simple.html: Added.
* TestWebKitAPI/mac/PlatformUtilitiesMac.mm:
(TestWebKitAPI::Util::createURLForResource):
2010-10-05 Gabor Rapcsanyi <rgabor@inf.u-szeged.hu>
Reviewed by Ojan Vafai.
[NRWT] Rename current_dir to current_group because it's not a directory
https://bugs.webkit.org/show_bug.cgi?id=47169
* Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2010-10-04 Dirk Pranke <dpranke@chromium.org>
Unreviewed, build fix.
r69065 broke test-webkitpy by trying to create a '/tmp-X' directory,
which can't always be done. This test uses tempfile.mkdtemp(), which
should be safe.
* Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2010-10-04 Dirk Pranke <dpranke@chromium.org>
Unreviewed, build fix.
Add missing "from __future__ import with_statement" that was
breaking test-webkitpy on the Leopard bots (broke in r69040).
* Scripts/webkitpy/layout_tests/port/google_chrome.py:
2010-10-04 Sheriff Bot <webkit.review.bot@gmail.com>
Unreviewed, rolling out r69066.
http://trac.webkit.org/changeset/69066
https://bugs.webkit.org/show_bug.cgi?id=47163
Made NRWT reliability worse. (Requested by tkent on #webkit).
* Scripts/webkitpy/layout_tests/port/chromium.py:
* Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
2010-10-04 Gabor Rapcsanyi <rgabor@inf.u-szeged.hu>
Reviewed by Tony Chang.
[NRWT] Reverse the http tests to alphabetical order.
https://bugs.webkit.org/show_bug.cgi?id=47075
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2010-10-04 Robert Hogan <robert@webkit.org>
Reviewed by Andreas Kling.
[Qt] Fix timeout on http/tests/navigation/post-goback2.html and postredirect-goback2.html
We need to queue back and forward navigations in the DRT from the
LayoutTestController so that maybeDump() knows about them.
* DumpRenderTree/qt/LayoutTestControllerQt.cpp:
(LayoutTestController::queueBackNavigation):
(LayoutTestController::queueForwardNavigation):
2010-10-04 Tony Chang <tony@chromium.org>
Reviewed by Kent Tamura.
[chromium] fix image diffing in NRWT
https://bugs.webkit.org/show_bug.cgi?id=47128
* Scripts/webkitpy/layout_tests/port/chromium.py: Flush data to the
temp file and check the image_diff error code more carefully
* Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
2010-10-04 Dirk Pranke <dpranke@chromium.org>
Reviewed by Adam Barth.
aroben's change in r68792 actually broke new-run-webkit-tests when
running the DRT code path. His change was intended to fix the
way we were converting windows paths to URIs when running under
Cygwin (the paths were getting one too many "/" on the front).
However, the change ended up breaking the chromium_win port, which
had slightly different logic.
This patch removes the port-specific code and adds tests to make
sure we're getting the behavior we expect. The Port object no longer
exposes a get_absolute_path() method that can be used outside of
of converting test filenames, because it's unreliable otherwise
(we don't have the right context to know which conversion is intended).
https://bugs.webkit.org/show_bug.cgi?id=47140
* Scripts/webkitpy/layout_tests/port/base.py:
* Scripts/webkitpy/layout_tests/port/base_unittest.py:
* Scripts/webkitpy/layout_tests/port/chromium_win.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2010-10-04 Dirk Pranke <dpranke@chromium.org>
Unreviewed, build fix.
Handle crash introduced in r69040 if we are not running in a
Chromium checkout.
* Scripts/webkitpy/layout_tests/port/google_chrome_unittest.py:
2010-10-04 Dirk Pranke <dpranke@chromium.org>
Reviewed by Tony Chang.
Add a way for us to have test expectations that are specific to the
official builds of Google Chrome (as opposed to Chromium). This change
looks for an additional "test_expectations_chrome.txt" file in
Chromium's repository (webkit/tools/layout_tests), and uses the
concatenation of that file and the regular test_expectations.txt
file for test overrides.
https://bugs.webkit.org/show_bug.cgi?id=46854
* Scripts/webkitpy/layout_tests/port/google_chrome.py:
* Scripts/webkitpy/layout_tests/port/google_chrome_unittest.py:
2010-10-04 Simon Fraser <simon.fraser@apple.com>
Color tests in the list based on existing pass/fail result.
Fix the summary data to account for htmlOnly and nonHTML tests.
* CSSTestSuiteHarness/harness/harness.css:
(#test-list > option.pass):
(#test-list > option.fail):
(#test-list > option.skipped):
* CSSTestSuiteHarness/harness/harness.js:
(Test):
(Test.prototype.statusForFormat):
(TestSuite.prototype.fillTestList):
(TestSuite.prototype.updateTestList):
(TestSuite.prototype.recordResult):
(TestSuite.prototype.markTestCompleted):
(TestSuite.prototype.countTestsWithFlag):
(TestSuite.prototype.queryDatabaseForSummary.this.db.transaction):
(TestSuite.prototype.queryDatabaseForSummary):
2010-10-04 Chris Fleizach <cfleizach@apple.com>
Reviewed by Beth Dakin.
AX: doAXRangeForLine does not work
https://bugs.webkit.org/show_bug.cgi?id=47101
DRT support to handle NSAccessibilityRangeForLineParameterizedAttribute.
* DumpRenderTree/AccessibilityUIElement.cpp:
(rangeForLineCallback):
(AccessibilityUIElement::rangeForLine):
(AccessibilityUIElement::getJSClass):
* DumpRenderTree/AccessibilityUIElement.h:
* DumpRenderTree/mac/AccessibilityUIElementMac.mm:
(AccessibilityUIElement::rangeForLine):
2010-10-04 Andrey Kosyakov <caseq@chromium.org>
Unreviewed. Adding myself to the committers list.
* Scripts/webkitpy/common/config/committers.py:
2010-10-04 Sam Weinig <sam@webkit.org>
Reviewed by Anders Carlsson.
Add ability to set the CacheModel in Webkit2
https://bugs.webkit.org/show_bug.cgi?id=47066
* MiniBrowser/mac/AppDelegate.m:
(-[BrowserAppDelegate init]): Opt minibrowser into
a PrimaryWebBrowser cache model.
2010-10-03 Simon Fraser <simon.fraser@apple.com>
Make sure to enter all tests when creating the database
for the first time.
When migrating to a new version of the suite, be sure to
sync up the database and testinfo.data by removing old
tests, and inserting new ones.
* CSSTestSuiteHarness/harness/harness.js:
(TestSuite.prototype.openDatabase.creation):
(TestSuite.prototype.databaseCreated):
(TestSuite.prototype.populateDatabaseFromTestInfoData):
(TestSuite.prototype.insertTest):
2010-10-03 Simon Fraser <simon.fraser@apple.com>
More work on treating HTML4 and XHTML1 independently; when
changing the format, rebuild the test list, and update the
numbers in the chapter popup.
* CSSTestSuiteHarness/harness/harness.js:
(Test):
(Test.prototype.runForFormat):
(Test.prototype.completedForFormat):
(ChapterSection):
(ChapterSection.prototype.countTests):
(Chapter):
(Chapter.prototype.description):
(Chapter.prototype.countTests):
(Chapter.prototype.testCount):
(Chapter.prototype.untestedCount):
(TestSuite.prototype.buildChapters):
(TestSuite.prototype.loadCurrentTest):
(TestSuite.prototype.updateProgressLabel):
(TestSuite.prototype.processFlags):
(TestSuite.prototype.formatChanged):
2010-10-03 Simon Fraser <simon.fraser@apple.com>
Update test suite version 20101001.
Handle database migration, and delete tests from the db that are
not present in testinfo.data.
Load about:blank into the test frame before the test url, to make
missing tests more obvious.
* CSSTestSuiteHarness/harness/harness.js:
(TestSuite.prototype.loadTest):
(TestSuite.prototype.openDatabase.creation):
(TestSuite.prototype.openDatabase.migration1_0To1_1):
(TestSuite.prototype.openDatabase.if.return):
(TestSuite.prototype.databaseReady):
(TestSuite.prototype.populateDatabaseFromTestInfoData):
(TestSuite.prototype.syncDatabaseWithTestInfoData.this.db.transaction):
(TestSuite.prototype.syncDatabaseWithTestInfoData):
2010-10-02 Simon Fraser <simon.fraser@apple.com>
Add the ability to jump to a specific test.
* CSSTestSuiteHarness/harness/harness.html:
* CSSTestSuiteHarness/harness/harness.js:
(TestSuite.prototype.goToTestByName):
(TestSuite.prototype.switchToFormat):
2010-10-02 Simon Fraser <simon.fraser@apple.com>
For a ref test, load the ref in the same format (HTML4 vs XHTML1)
as the test.
* CSSTestSuiteHarness/harness/harness.js:
(TestSuite.prototype.loadRef):
2010-10-02 Simon Fraser <simon.fraser@apple.com>
Update the UI to reflect the fact that each test needs to be tested
in both HTML4 and XHTML1 format.
* CSSTestSuiteHarness/harness/harness.js:
(Test):
(Chapter.prototype.description):
(Chapter.prototype.untestedCount):
(TestSuite.prototype.fillChapterPopup):
(TestSuite.prototype.updateChapterPopup):
(TestSuite.prototype.fillTestList):
(TestSuite.prototype.updateTestList):
(TestSuite.prototype.goToNextIncompleteTest):
(TestSuite.prototype.firstIncompleteTestIndex):
(TestSuite.prototype.recordResult):
(TestSuite.prototype.formatChanged):
(TestSuite.prototype.markTestCompleted):
(TestSuite.prototype.resetTestStatus):
2010-10-01 Simon Fraser <simon.fraser@apple.com>
Add the ability to remove results for re-testing, and
to import results.
* CSSTestSuiteHarness/harness/harness.css:
(.custom button):
(#overlay):
(#overlay.visible):
(.overlay-contents):
(.overlay-contents textarea):
(.overlay-contents .buttons):
(.overlay-contents .note):
(.overlay-contents .buttons button):
* CSSTestSuiteHarness/harness/harness.html:
* CSSTestSuiteHarness/harness/harness.js:
(TestSuite.prototype.importResults):
(TestSuite.prototype.clearResults):
(TestSuite.prototype.markTestCompleted):
(TestSuite.prototype.resetTestStatus):
(TestSuite.prototype.storeTestResult):
(TestSuite.prototype.importTestResults):
(TestSuite.prototype.clearTestResults):
2010-10-01 Sam Weinig <sam@webkit.org>
Reviewed by Jon Honeycutt.
Deploy the new WKString functions to remove most uses of CF from
the WebKitTestRunner.
* WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
(WTR::InjectedBundle::didReceiveMessage):
(WTR::InjectedBundle::done):
* WebKitTestRunner/StringFunctions.h:
(WTR::toWK):
(WTR::toJS):
(WTR::toSTD):
(WTR::operator<<):
(WTR::copyURLString):
* WebKitTestRunner/TestController.cpp:
(WTR::TestController::resetStateToConsistentValues):
(WTR::TestController::didFinishLoadForFrame):
* WebKitTestRunner/TestInvocation.cpp:
(WTR::TestInvocation::invoke):
(WTR::TestInvocation::didReceiveMessageFromInjectedBundle):
2010-10-01 Mihai Parparita <mihaip@chromium.org>
Unreviewed TestResultsServer change.
Add files introduced by http://crrev.com/61273 to the list of dashboard
files to update.
* TestResultServer/handlers/dashboardhandler.py:
2010-10-01 Brian Weinstein <bweinstein@apple.com>
Build Fix for Windows.
* WebKitTestRunner/win/WebKitTestRunner.vcproj: Don't copy ForwardingHeaders from
WebKitTestRunner/ForwardingHeaders, instead, add $(WebKitOutputDir)/include/WebCore/
ForwardingHeaders to the include path.
2010-10-01 Sam Weinig <sam@webkit.org>
Reviewed by Anders Carlsson.
Add SPI to convert a WKStringRef to a JSStringRef and vice-versa.
* TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
* TestWebKitAPI/Tests/WebKit2/WKString.cpp:
(TestWebKitAPI::TEST):
* TestWebKitAPI/Tests/WebKit2/WKStringJSString.cpp: Added.
(TestWebKitAPI::TEST):
2010-10-01 Sam Weinig <sam@webkit.org>
Reviewed by Kenneth Rohde Christiansen.
Add additional WKString API
https://bugs.webkit.org/show_bug.cgi?id=46958
Add basic WKStringRef tests.
* TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
* TestWebKitAPI/Tests/WebKit2/WKString.cpp: Added.
2010-10-01 Sam Weinig <sam@webkit.org>
Reviewed by Dan Bernstein.
Add makefile for TestWebKitAPI and call it from the base makefile.
* Makefile:
* TestWebKitAPI/Makefile: Added.
2010-10-01 Sam Weinig <sam@webkit.org>
Reviewed by Adam Roben.
Remove unnecessary ForwardingHeaders from test runners.
https://bugs.webkit.org/show_bug.cgi?id=47010
* TestWebKitAPI/Configurations/Base.xcconfig:
* TestWebKitAPI/ForwardingHeaders: Removed.
* WebKitTestRunner/Configurations/Base.xcconfig:
* WebKitTestRunner/ForwardingHeaders: Removed.
2010-09-30 Adam Barth <abarth@webkit.org>
Reviewed by Eric Seidel.
WebSocket tests are flaky
https://bugs.webkit.org/show_bug.cgi?id=46956
Update these tools to understand how to run the WebSocket tests off the
Apache server.
* Scripts/old-run-webkit-tests:
* Scripts/webkitpy/layout_tests/port/base.py:
* Scripts/webkitpy/layout_tests/port/websocket_server.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2010-10-01 Adam Roben <aroben@apple.com>
Encode Executive command arguments using UTF-8 on Cygwin
Cygwin's Python's os.execv doesn't support unicode command arguments.
Cygwin's execv expects arguments to be encoded using the current code
page. But code pages are limited in what characters they can handle,
and our tests include characters that the English code page can't
handle. So for now we'll just encode everything in UTF-8 on Cygwin,
which can handle all characters but might confuse some commands, for
expediency's sake. I'm sure we'll run into cases where UTF-8 isn't
good enough, but we can deal with that when the problem arises.
Reviewed by Adam Barth.
Fixes <http://webkit.org/b/46892> <rdar://problem/8496639>
webkitpy.common.system.executive_unittest.ExecutiveTest.test_run_command_with_unicode
fails on Windows
* Scripts/webkitpy/common/system/executive.py:
(Executive._run_command_with_teed_output):
(Executive.run_command):
On Cygwin, encode arguments using UTF-8.
2010-10-01 Sam Weinig <sam@webkit.org>
Reviewed by Adam Roben.
Add simple API tester for WebKit2
https://bugs.webkit.org/show_bug.cgi?id=46953
This adds a very simple testing infrastructure for testing APIs exposed
through the WebKit project. It consists of two parts: 1) a project that
builds the all the tests 2) a script that searches the tests directory and
calls the tester once for each test.
This adds the infrastructure and two tests:
- Tests/WTF/Vector1.cpp - A proof of concept test of WTF data-structures.
- Tests/WebKit2/BasicTest1.cpp - A proof of concept test of WebKit2 API.
This currently only works on the mac, but is designed to be easily ported
to any platform.
* Scripts/run-api-tests: Added.
* Scripts/build-api-tests: Added.
Scripts to build/run the tests.
* TestWebKitAPI: Added.
* TestWebKitAPI/Configurations: Added.
* TestWebKitAPI/Configurations/Base.xcconfig: Added.
* TestWebKitAPI/Configurations/DebugRelease.xcconfig: Added.
* TestWebKitAPI/Configurations/TestWebKitAPI.xcconfig: Added.
* TestWebKitAPI/ForwardingHeaders: Added.
* TestWebKitAPI/ForwardingHeaders/wtf: Added.
* TestWebKitAPI/PlatformUtilities.h: Added.
* TestWebKitAPI/PlatformWebView.h: Added.
(TestWebKitAPI::PlatformWebView::platformView):
* TestWebKitAPI/StringFunctions.h: Added.
* TestWebKitAPI/Test.h: Added.
(TestWebKitAPI::Test::~Test):
(TestWebKitAPI::Test::name):
(TestWebKitAPI::Test::Register::Register):
(TestWebKitAPI::Test::Register::create):
(TestWebKitAPI::Test::Test):
* TestWebKitAPI/TestWebKitAPI.xcodeproj: Added.
* TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj: Added.
* TestWebKitAPI/TestWebKitAPIPrefix.h: Added.
* TestWebKitAPI/TestsController.cpp: Added.
(TestWebKitAPI::TestsController::shared):
(TestWebKitAPI::TestsController::TestsController):
(TestWebKitAPI::TestsController::runTestNamed):
(TestWebKitAPI::TestsController::testFailed):
(TestWebKitAPI::TestsController::registerCreateTestFunction):
* TestWebKitAPI/TestsController.h: Added.
* TestWebKitAPI/mac: Added.
* TestWebKitAPI/mac/PlatformUtilitiesMac.mm: Added.
(TestWebKitAPI::Util::run):
(TestWebKitAPI::Util::createURLForResource):
* TestWebKitAPI/mac/PlatformWebViewMac.mm: Added.
(TestWebKitAPI::PlatformWebView::PlatformWebView):
(TestWebKitAPI::PlatformWebView::resizeTo):
(TestWebKitAPI::PlatformWebView::~PlatformWebView):
(TestWebKitAPI::PlatformWebView::page):
(TestWebKitAPI::PlatformWebView::focus):
* TestWebKitAPI/mac/main.mm: Added.
Infrastructure.
* TestWebKitAPI/Tests: Added.
* TestWebKitAPI/Tests/WTF: Added.
* TestWebKitAPI/Tests/WTF/Vector1.cpp: Added.
* TestWebKitAPI/Tests/WebKit2: Added.
* TestWebKitAPI/Tests/WebKit2/BasicTest1.cpp: Added.
(TestWebKitAPI::State::State):
* TestWebKitAPI/Tests/WebKit2/basic-1.html: Added.
Proof of concept tests.
2010-10-01 Adam Roben <aroben@apple.com>
Don't assume AccessibleObjectFromEvent succeeds
Fixes <http://webkit.org/b/44136> <rdar://problem/8321684> Crash in
DumpRenderTree!notificationListenerProc when running
plugins/access-after-page-destroyed.html
Reviewed by Sam Weinig.
* DumpRenderTree/win/AccessibilityControllerWin.cpp:
(notificationListenerProc): Check both the return value of
AccessibleObjectFromEvent and the object it returns, as MSDN
recommends.
2010-10-01 Gabor Rapcsanyi <rgabor@inf.u-szeged.hu>
Reviewed by Tony Chang.
[NRWT] Put the http and websocket tests first in the test list.
https://bugs.webkit.org/show_bug.cgi?id=46453
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2010-10-01 Fady Samuel <fsamuel@chromium.org>
Unreviewed, adding myself to the committer list.
* Scripts/webkitpy/common/config/committers.py:
2010-10-01 Adam Roben <aroben@apple.com>
Test NPN_GetValue(NPNVnetscapeWindow)
Test for <http://webkit.org/b/46726> <rdar://problem/8486319>
Right-clicking on windowless Flash plugin in WebKit2 makes a context
menu appear in the bottom-right corner of the screen
Reviewed by Anders Carlsson.
* DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp:
(PluginTest::NPN_GetValue): Added. Calls through to the browser.
* DumpRenderTree/TestNetscapePlugIn/PluginTest.h: Added NPN_GetValue.
* DumpRenderTree/TestNetscapePlugIn/Tests/win/GetValueNetscapeWindow.cpp: Added.
(GetValueNetscapeWindow::GetValueNetscapeWindow): Initialize members.
(GetValueNetscapeWindow::NPP_SetWindow): Test that
NPN_GetValue(NPNVnetscapeWindow) returns a valid HWND and that it
isn't our HWND.
* DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
Added GetValueNetscapeWindow.
2010-10-01 Andreas Kling <andreas.kling@nokia.com>
Unreviewed, adding my webkit.org identity to reviewer list.
* Scripts/webkitpy/common/config/committers.py:
2010-09-30 Adam Barth <abarth@webkit.org>
Reviewed by Eric Seidel.
The WebSocket server should flush its logs
https://bugs.webkit.org/show_bug.cgi?id=46946
The WebSocket server logs are truncated because the driver just
terminates the child. It should run Python without buffering so we can
see the end of the log.
* Scripts/webkitpy/layout_tests/port/websocket_server.py:
2010-09-30 Simon Fraser <simon.fraser@apple.com>
Add a button to jump to the next untested test.
* CSSTestSuiteHarness/harness/harness.css:
(.test-type):
(.name > button):
* CSSTestSuiteHarness/harness/harness.html:
* CSSTestSuiteHarness/harness/harness.js:
(Chapter.prototype.description):
(Chapter.prototype.untestedCount):
(TestSuite.prototype.testInfoDataLoaded):
(TestSuite.prototype.fillChapterPopup):
(TestSuite.prototype.updateChapterPopup):
(TestSuite.prototype.buildTestListForChapter):
(TestSuite.prototype.goToNextIncompleteTest):
(TestSuite.prototype.firstIncompleteTestIndex):
(TestSuite.prototype.testCompletionStateChanged):
2010-09-29 Jon Honeycutt <jhoneycutt@apple.com>
WebKit2 on Windows should use Windows fonts for the various standard
font families
https://bugs.webkit.org/show_bug.cgi?id=43499
<rdar://problem/8272758>
Reviewed by Adam Roben.
* WebKitTestRunner/TestController.cpp:
(WTR::TestController::resetStateToConsistentValues):
Use the Mac fonts for running tests. This matches DRT behavior.
2010-09-30 Adam Barth <abarth@webkit.org>
Third attempt. We need a better integration test environment.
* Scripts/webkitpy/common/net/failuremap.py:
* Scripts/webkitpy/common/net/failuremap_unittest.py:
2010-09-30 Robert Hogan <robert@webkit.org>
Reviewed by Andreas Kling.
[Qt] Support for PlainTextController
Unskip editing/text-iterator/basic-iteration.html
https://bugs.webkit.org/show_bug.cgi?id=38805
* DumpRenderTree/qt/DumpRenderTree.pro:
* DumpRenderTree/qt/DumpRenderTreeQt.cpp:
(WebCore::DumpRenderTree::DumpRenderTree):
(WebCore::DumpRenderTree::initJSObjects):
* DumpRenderTree/qt/DumpRenderTreeQt.h:
* DumpRenderTree/qt/PlainTextControllerQt.cpp: Added.
(PlainTextController::PlainTextController):
(PlainTextController::plainText):
* DumpRenderTree/qt/PlainTextControllerQt.h: Added.
2010-09-30 Adam Barth <abarth@webkit.org>
Unreviewed. Second attempt to fix sheriffbot.
* Scripts/webkitpy/common/net/buildbot.py:
2010-09-30 Adam Barth <abarth@webkit.org>
Unreviewed.
This patch should stop the sheriffbot from throwing an exception.
* Scripts/webkitpy/common/net/buildbot.py:
2010-09-30 Brian Weinstein <bweinstein@apple.com>
Reviewed by Adam Roben.
The same bots shouldn't run both Windows Debug Tests and Windows Release Tests.
https://bugs.webkit.org/show_bug.cgi?id=46926
* BuildSlaveSupport/build.webkit.org-config/config.json: Have two bots running Windows
Release Tests, and two bots running Windows Debug Tests, instead of four running both.
2010-09-30 Adam Barth <abarth@webkit.org>
Reviewed by Eric Seidel.
webkit-patch find-flaky-tests should print results URLs
https://bugs.webkit.org/show_bug.cgi?id=46917
Printing out the results URL makes find-flaky-tests more actionable
because you can see what happens when a test fails. The output from
the script is still pretty noisy, but it seems useful.
* Scripts/webkitpy/tool/commands/queries.py:
2010-09-30 Chris Marrin <cmarrin@apple.com>
Reviewed by Simon Fraser.
Make 2D accelerated canvas rendering build on Mac
https://bugs.webkit.org/show_bug.cgi?id=46007
Added ACCELERATED_2D_CANVAS to build-webkit
* Scripts/build-webkit:
2010-09-29 Kenneth Russell <kbr@google.com>
Reviewed by Dimitri Glazkov.
[chromium] Add accelerated compositing support to DumpRenderTree and test_shell
https://bugs.webkit.org/show_bug.cgi?id=46849
Added offscreen code path for WebGraphicsContext3DDefaultImpl which
works with the compositor integration in both DumpRenderTree and
test_shell, since both pass a non-null WebCanvas* to WebViewImpl::paint
and thereby trigger the compositor's readback code path. Added support
for --enable-accelerated-compositing to DumpRenderTree.
Tested in both test_shell and DumpRenderTree on Linux, the latter by
modifying a compositing layout test, dumping the pixels and verifying
that they matched the output when the compositor was active.
* DumpRenderTree/chromium/DumpRenderTree.cpp:
(main):
* DumpRenderTree/chromium/TestShell.cpp:
(TestShell::TestShell):
(TestShell::resetWebSettings):
* DumpRenderTree/chromium/TestShell.h:
(TestShell::setAcceleratedCompositingEnabled):
* DumpRenderTree/chromium/WebPreferences.cpp:
(WebPreferences::reset):
(WebPreferences::applyTo):
* DumpRenderTree/chromium/WebPreferences.h:
* DumpRenderTree/chromium/WebViewHost.cpp:
(WebViewHost::scheduleComposite):
* DumpRenderTree/chromium/WebViewHost.h:
2010-09-30 Adam Roben <aroben@apple.com>
Ensure that QueueEngine cleans up its log files when its delegate
tells it to stop processing work
Reviewed by Adam Barth.
Fixes <http://webkit.org/b/46891> <rdar://problem/8496638> Many tests
in webkitpy.tool.bot.queueengine_unittest.QueueEngineTest crash on
Windows
* Scripts/webkitpy/tool/bot/queueengine.py:
(QueueEngine.run): Stop ourselves normally (including cleaning up log
files) when the delegate tells us to stop processing work.
* Scripts/webkitpy/tool/bot/queueengine_unittest.py:
(LoggingDelegate.__init__): Moved code here from
RaisingDelegate.__init__.
(LoggingDelegate.expeced_callbacks): Added the stop_work_queue
callback.
(LoggingDelegate.stop_work_queue): Moved here from RaisingDelegate.
(RaisingDelegate.__init__): Removed code that LoggingDelegate takes
care of for us now.
(QueueEngineTest.test_trivial): Make sure we got the expected stop
message.
(QueueEngineTest.test_not_safe_to_proceed): Changed to explicitly
remove the callbacks that are related to processing a single work
item, rather than removing all callbacks after a certain point, as
there are now more callbacks we expect to receive at the end.
2010-09-30 Adam Roben <aroben@apple.com>
Fix path -> URL conversion on Cygwin
Reviewed by Adam Barth.
Fixes <http://webkit.org/b/46890> <rdar://problem/8496637> Many tests
in webkitpy.layout_tests.run_webkit_tests_unittest assert on Windows
* Scripts/webkitpy/layout_tests/port/base.py:
(Port.filename_to_uri): Treat Cygwin like other UNIX-y platforms by
assuming paths already have a leading slash.
2010-09-30 Adam Roben <aroben@apple.com>
Teach WindowGeometryInitializedBeforeSetWindow that NPP_SetWindow can
be called more than once
The NPP_SetWindow call made when the plugin is torn down was confusing
us, and was causing us to spew an error into the next test.
Fixes <http://webkit.org/b/46900> <rdar://problem/8496888> REGRESSION
(r68520): plugins/access-after-page-destroyed.html failing on Windows
Reviewed by Anders Carlsson.
* DumpRenderTree/TestNetscapePlugIn/Tests/win/WindowGeometryInitializedBeforeSetWindow.cpp:
(WindowGeometryInitializedBeforeSetWindow::WindowGeometryInitializedBeforeSetWindow):
Initialize member.
(WindowGeometryInitializedBeforeSetWindow::NPP_SetWindow): Bail if
this isn't the first time this is called.
2010-09-30 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
webkit-patch upload warns that I'm using Xcode when I'm not
https://bugs.webkit.org/show_bug.cgi?id=46869
String.find returns -1 when not found, so switched to "in".
* Scripts/webkitpy/common/system/user.py:
* Scripts/webkitpy/common/system/user_unittest.py:
2010-09-30 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
webkit-patch failure-reason dies if non-trunk commits are in the blame range
https://bugs.webkit.org/show_bug.cgi?id=46866
I also made failure-reason use RegressionWindow in a cleaner way.
* Scripts/webkitpy/tool/commands/queries.py:
2010-09-29 Adam Barth <abarth@webkit.org>
Reviewed by Eric Seidel.
webkit-patch find-flaky-tests
https://bugs.webkit.org/show_bug.cgi?id=46876
This command helps us find flaky tests so we can squash them.
* Scripts/webkitpy/tool/commands/queries.py:
2010-09-29 Simon Fraser <simon.fraser@apple.com>
Fix export to use '?' instead of 'null' for
tests with no results.
* CSSTestSuiteHarness/harness/harness.js:
2010-09-29 Simon Fraser <simon.fraser@apple.com>
Add a warning when a test requires special steps.
Add a Print Preview button for 'paged' tests that
brings up the print dialog, allowing the user to
judge paged media tests.
* CSSTestSuiteHarness/harness/harness.css:
* CSSTestSuiteHarness/harness/harness.html:
* CSSTestSuiteHarness/harness/harness.js:
(TestSuite.prototype.loadTest):
(TestSuite.prototype.processFlags):
2010-09-29 Adam Barth <abarth@webkit.org>
Reviewed by Eric Seidel.
Remove revisions_casuing_failures
https://bugs.webkit.org/show_bug.cgi?id=46872
This function exists only to be unit tested.
* Scripts/webkitpy/common/net/failuremap.py:
* Scripts/webkitpy/common/net/failuremap_unittest.py:
* Scripts/webkitpy/tool/mocktool.py:
2010-09-29 Adam Barth <abarth@webkit.org>
Reviewed by Eric Seidel.
SheriffBot should post the list of failing tests to bugs
https://bugs.webkit.org/show_bug.cgi?id=46871
* Scripts/webkitpy/common/net/buildbot.py:
* Scripts/webkitpy/common/net/failuremap.py:
* Scripts/webkitpy/common/net/regressionwindow.py:
* Scripts/webkitpy/tool/bot/sheriff.py:
* Scripts/webkitpy/tool/bot/sheriff_unittest.py:
* Scripts/webkitpy/tool/commands/earlywarningsystem_unittest.py:
* Scripts/webkitpy/tool/commands/queues_unittest.py:
* Scripts/webkitpy/tool/commands/sheriffbot.py:
* Scripts/webkitpy/tool/commands/sheriffbot_unittest.py:
* Scripts/webkitpy/tool/commands/upload_unittest.py:
* Scripts/webkitpy/tool/mocktool.py:
2010-09-29 Brent Fulgham <bfulgham@webkit.org>
Reviewed by Martin Robinson, Eric Seidel, and Mark Rowe.
[WinCairo] Adds a build slave.
https://bugs.webkit.org/show_bug.cgi?id=46360
* BuildSlaveSupport/build.webkit.org-config/config.json: Add
entries for a debug build of the WinCairo port.
* BuildSlaveSupport/build.webkit.org-config/master.cfg: Add
platform flag to configuration for wincairo (like gtk, chromium, etc.)
* Scripts/build-webkit: Change 'cairo-win32' labeling to 'wincairo'
* Scripts/webkitdirs.pm: Change 'cairo-win32' labeling to 'wincairo'
2010-09-29 Adam Barth <abarth@webkit.org>
Reviewed by Eric Seidel.
Move more SheriffBot smarts into FailureMap
https://bugs.webkit.org/show_bug.cgi?id=46703
This patch pushes the FailureMap model object further into the
SheriffBot machine. In addition, it moves a couple operations on this
object from SheriffBot itself to the model.
Eventually, FailureMap will be the canonical context object for
SheriffBot operations. FailureMap represents a map of the current
failures on the bots that might require remediation.
* Scripts/webkitpy/common/net/failuremap.py:
* Scripts/webkitpy/common/net/regressionwindow.py:
* Scripts/webkitpy/tool/commands/queries.py:
* Scripts/webkitpy/tool/commands/sheriffbot.py:
* Scripts/webkitpy/tool/mocktool.py:
2010-09-29 Tony Chang <tony@chromium.org>
Reviewed by James Robinson.
[chromium] enable -Werror for DRT and webkit_unit_tests on Linux
https://bugs.webkit.org/show_bug.cgi?id=46829
* DumpRenderTree/chromium/DRTDevToolsClient.cpp:
(DRTDevToolsClient::DRTDevToolsClient):
* DumpRenderTree/chromium/TestShell.cpp:
(TestShell::TestShell):
* DumpRenderTree/chromium/WebViewHost.cpp:
(WebViewHost::postAccessibilityNotification):
(WebViewHost::WebViewHost):
2010-09-29 Chris Fleizach <cfleizach@apple.com>
Reviewed by Darin Adler.
AX: implement CSS3 Speech "speak"
https://bugs.webkit.org/show_bug.cgi?id=46827
* DumpRenderTree/AccessibilityUIElement.cpp:
(speakCallback):
(AccessibilityUIElement::speak):
(AccessibilityUIElement::getJSClass):
* DumpRenderTree/AccessibilityUIElement.h:
* DumpRenderTree/mac/AccessibilityUIElementMac.mm:
(AccessibilityUIElement::speak):
2010-09-29 Mihai Parparita <mihaip@chromium.org>
Reviewed by Adam Barth.
Record bot ID when updating queue status
https://bugs.webkit.org/show_bug.cgi?id=46764
Since we now have multiple bots handling patches from the commit queue,
we need to differentiate status updates from them, so that we can group
the queue status page in a less confusing way.
* QueueStatusServer/handlers/updatestatus.py: Accept bot_id as input
* QueueStatusServer/model/queuestatus.py: Add bot_id property
* QueueStatusServer/templates/updatestatus.html: Add bot_id input
* Scripts/webkitpy/common/net/bugzilla_unittest.py: Extract MockBrowser
* Scripts/webkitpy/common/net/statusserver.py: Add bot_id
* Scripts/webkitpy/common/net/statusserver_unittest.py: Added.
* Scripts/webkitpy/tool/main.py: Add --bot-id
* Scripts/webkitpy/tool/mocktool.py: Add MockBrowser
2010-09-29 Simon Fraser <simon.fraser@apple.com>
No review.
References always refert to a file in .xht format, even
for HTML tests, so fix their loading.
* CSSTestSuiteHarness/harness/harness.js:
(TestSuite.prototype.loadRef):
2010-09-28 Johnny Ding <jnd@chromium.org>
Reviewed by Adam Barth.
https://bugs.webkit.org/show_bug.cgi?id=41292
Add a new parameter to the test plugin to allow specifying a script
and a mouse/keyboard event. The specified script will be evaluated
in the browser when the specified event is received by the plugin.
* DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
(pluginAllocate):
* DumpRenderTree/TestNetscapePlugIn/PluginObject.h:
* DumpRenderTree/TestNetscapePlugIn/main.cpp:
(NPP_New):
(handleEventCarbon):
(handleEventCocoa):
2010-09-28 Simon Fraser <simon.fraser@apple.com>
Fix export by adding missing argument.
* CSSTestSuiteHarness/harness/harness.js:
(TestSuite.prototype.showResultsForAllTests):
2010-09-28 Kevin Ollivier <kevino@theolliviers.com>
[wx] Build fix, add missing import.
* wx/packaging/build-win-installer.py:
2010-09-28 Simon Fraser <simon.fraser@apple.com>
No review
Show which tests have been run by dimming them out in the test list.
Adjust some element sizes.
* CSSTestSuiteHarness/harness/harness.css:
(#test-list > option.untested):
(#test-list > option.completed):
(#output):
(.output-options):
* CSSTestSuiteHarness/harness/harness.js:
(Test):
(TestSuite.prototype.fillTestList):
(TestSuite.prototype.updateTestList):
(TestSuite.prototype.setSelectedChapter):
(TestSuite.prototype.recordResult):
(TestSuite.prototype.markTestCompleted):
(TestSuite.prototype.testCompletionStateChanged):
(TestSuite.prototype.loadTestStatus):
2010-09-28 Simon Fraser <simon.fraser@apple.com>
No review.
Implement export of various queries on the test database, sharing
code with that which displays results inline.
* CSSTestSuiteHarness/harness/harness.js:
():
(TestSuite.prototype.exportResultsCompletion.var):
(TestSuite.prototype.exportResultsCompletion):
(TestSuite.prototype.showResultsForCompletedTests):
(TestSuite.prototype.exportResultsForCompletedTests):
(TestSuite.prototype.showResultsForAllTests):
(TestSuite.prototype.exportResultsForAllTests):
(TestSuite.prototype.exportResultsForTestsNotRun):
(TestSuite.prototype.exportResultsForTestsWithStatus):
(TestSuite.prototype.exportResultsForTestsWithMismatchedResults):
2010-09-28 Simon Fraser <simon.fraser@apple.com>
No review.
Work around uncertainty about the order of database
transactions when creating the database, so that we don't
try to query the table before it has been created.
* CSSTestSuiteHarness/harness/harness.js:
(TestSuite):
(TestSuite.prototype.databaseCreated):
(TestSuite.prototype.storeTestResult):
(TestSuite.prototype.populateDatabaseFromTestInfoData):
(TestSuite.prototype.queryDatabaseForTestsWithStatus):
(TestSuite.prototype.queryDatabaseForTestsWithMixedStatus):
(TestSuite.prototype.queryDatabaseForCompletedTests):
(TestSuite.prototype.queryDatabaseForTestsNotRun):
2010-09-28 Tony Chang <tony@chromium.org>
Reviewed by Eric Seidel.
add python keyring support to webkit-patch
https://bugs.webkit.org/show_bug.cgi?id=41269
* Scripts/webkitpy/common/net/credentials.py: Add the ability to read passwords using
the python keyring module
* Scripts/webkitpy/common/net/credentials_unittest.py:
* Scripts/webkitpy/common/system/user.py: Allow confirm() to default to no and add testing params.
* Scripts/webkitpy/common/system/user_unittest.py:
* Scripts/webkitpy/tool/mocktool.py:
2010-09-28 Martin Robinson <mrobinson@igalia.com>
Reviewed by Gustavo Noronha Silva.
[Cairo] FreeType fonts should obey FontConfig hinting/anti-aliasing settings
https://bugs.webkit.org/show_bug.cgi?id=46740
* DumpRenderTree/gtk/fonts/fonts.conf: Add specialized variants of common
fonts which can be used to fully test FontConfig rendering settings.
2010-09-28 Martin Robinson <mrobinson@igalia.com>
Reviewed by Gustavo Noronha Silva.
[gtk] Fedora has a different path for the liberation fonts
https://bugs.webkit.org/show_bug.cgi?id=46709
When searching for DRT fonts, also look in the path where those fonts
are commonly found on Fedora systems.
* DumpRenderTree/gtk/DumpRenderTree.cpp:
(initializeFonts): Include logic for looking for Fedora-style font paths.
2010-09-28 Martin Robinson <mrobinson@igalia.com>
Reviewed by Andreas Kling.
run-webkit-tests needs an updated list of directories with webgl tests
https://bugs.webkit.org/show_bug.cgi?id=46747
Update the list of directories containing WebGL tests and a unit test
which depends on this list.
* Scripts/old-run-webkit-tests: Update the list of directories.
* Scripts/webkitpy/layout_tests/port/webkit.py: Ditto.
* Scripts/webkitpy/layout_tests/port/webkit_unittest.py: Update the expected result.
2010-09-28 Adam Barth <abarth@webkit.org>
Reviewed by Eric Seidel.
Introduce FailureMap to summaries the failures status of all the bots
https://bugs.webkit.org/show_bug.cgi?id=46700
This patch gives the SheriffBot a handle on the list of failing tests.
* Scripts/webkitpy/common/net/buildbot.py:
* Scripts/webkitpy/common/net/failuremap.py: Added.
* Scripts/webkitpy/tool/commands/queries.py:
* Scripts/webkitpy/tool/commands/sheriffbot.py:
* Scripts/webkitpy/tool/mocktool.py:
2010-09-28 Adam Roben <aroben@apple.com>
Update for the addition of WKPageUIClient::didNotHandleKeyEvent
Fixes <http://webkit.org/b/46660> <rdar://problem/8483465> Need API to
tell a WebKit2 client application that a key event was not handled
Reviewed by Kenneth Rohde Christiansen and Sam Weinig.
* MiniBrowser/mac/BrowserWindowController.m:
(-[BrowserWindowController awakeFromNib]):
* MiniBrowser/win/BrowserView.cpp:
(BrowserView::create):
* WebKitTestRunner/TestController.cpp:
(WTR::createOtherPage):
(WTR::TestController::initialize):
2010-09-28 Simon Fraser <simon.fraser@apple.com>
Unreviewed.
Update the CSSTestSuiteHarness to not scrape the XHTML chapter files
for the test order, but instead use information from testinfo.data.
* CSSTestSuiteHarness/harness/harness.html:
* CSSTestSuiteHarness/harness/harness.js:
2010-09-28 Kevin Ollivier <kevino@theolliviers.com>
[wx] Build fix, don't try to get the svn revision from the git repo when packaging,
it causes the process to hang on the gitorious repo.
* wx/packaging/build-mac-installer.py:
* wx/packaging/build-win-installer.py:
2010-09-28 Adam Roben <aroben@apple.com>
Test that a plugin's HWND is sized/positioned before NPP_SetWindow is
called
Reviewed by Anders Carlsson.
Test for <http://webkit.org/b/46716> <rdar://problem/8482014>
Full-page Adobe Reader does not paint until window is resized
* DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp:
(PluginTest::NPP_SetWindow): Added. Just returns NPERR_NO_ERROR at
this level.
* DumpRenderTree/TestNetscapePlugIn/PluginTest.h: Declared
NPP_SetWindow.
* DumpRenderTree/TestNetscapePlugIn/Tests/win/WindowGeometryInitializedBeforeSetWindow.cpp: Added.
(WindowGeometryInitializedBeforeSetWindow::NPP_SetWindow): Checks that
the plugin's HWND has a non-zero size and that its size/position
matches that specified in the NPWindow.
* DumpRenderTree/TestNetscapePlugIn/main.cpp:
(NPP_SetWindow): Call through to the PluginTest.
* DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
Added WindowGeometryInitializedBeforeSetWindow.cpp, and let VS reorder
the files as it saw fit.
2010-09-28 Benjamin Poulain <benjamin.poulain@nokia.com>
Reviewed by Andreas Kling.
[Qt] Remove support for Qt 4.5
https://bugs.webkit.org/show_bug.cgi?id=46718
Remove the code for versions of Qt prior to 4.6.
* DumpRenderTree/qt/EventSenderQt.cpp:
(EventSender::addTouchPoint):
(EventSender::updateTouchPoint):
(EventSender::setTouchModifier):
(EventSender::touchStart):
(EventSender::touchMove):
(EventSender::touchEnd):
(EventSender::clearTouchPoints):
(EventSender::releaseTouchPoint):
(EventSender::sendTouchEvent):
* DumpRenderTree/qt/EventSenderQt.h:
* DumpRenderTree/qt/TextInputControllerQt.cpp:
(TextInputController::setMarkedText):
* QtTestBrowser/launcherwindow.cpp:
(LauncherWindow::LauncherWindow):
(LauncherWindow::initializeView):
(LauncherWindow::createChrome):
(LauncherWindow::sendTouchEvent):
(LauncherWindow::eventFilter):
(LauncherWindow::applyZoom):
(LauncherWindow::setTouchMocking):
* QtTestBrowser/launcherwindow.h:
* QtTestBrowser/utils.cpp:
(urlFromUserInput):
* QtTestBrowser/webview.cpp:
(WebViewGraphicsBased::animatedFlip):
(WebViewGraphicsBased::animatedYFlip):
* QtTestBrowser/webview.h:
(WebViewGraphicsBased::setYRotation):
2010-09-28 İsmail Dönmez <ismail@namtrac.org>
Reviewed by Andreas Kling.
Fix DRT compilation on WinCE. Introduce a WCECOMPAT variable
which should point to wcecompat installation. Needs wcecompat
library from http://github.com/mauricek/wcecompat .
* DumpRenderTree/qt/DumpRenderTree.pro:
* DumpRenderTree/qt/DumpRenderTreeQt.cpp:
(WebCore::DumpRenderTree::resetToConsistentStateBeforeTesting):
2010-09-28 Daniel Bates <dbates@rim.com>
Rollout changeset 68493 <http://trac.webkit.org/changeset/68493>
<https://bugs.webkit.org/show_bug.cgi?id=39136>
Rollout changeset 68493 because it broke Sheriffbot's rollout feature.
In particular, this change caused Sheriffbot to raise an exception when
trying to parse the bug id on a bug page. We need to look into this
some more.
* Scripts/webkitpy/common/net/bugzilla.py:
* Scripts/webkitpy/common/net/bugzilla_unittest.py:
* Scripts/webkitpy/tool/bot/sheriff.py:
* Scripts/webkitpy/tool/commands/download.py:
* Scripts/webkitpy/tool/commands/queues.py:
* Scripts/webkitpy/tool/commands/upload.py:
* Scripts/webkitpy/tool/steps/closebug.py:
* Scripts/webkitpy/tool/steps/obsoletepatches.py:
* Scripts/webkitpy/tool/steps/preparechangelog.py:
* Scripts/webkitpy/tool/steps/updatechangelogswithreviewer.py:
2010-09-28 Adam Barth <abarth@webkit.org>
Reviewed by Eric Seidel.
Move RegressionWindow further up the dependency chain
https://bugs.webkit.org/show_bug.cgi?id=46698
Baby steps towards pushing this information into bug posts.
* Scripts/webkitpy/common/net/buildbot.py:
* Scripts/webkitpy/common/net/buildbot_unittest.py:
* Scripts/webkitpy/common/net/regressionwindow.py:
* Scripts/webkitpy/tool/commands/queries.py:
2010-09-27 Daniel Bates <dbates@rim.com>
Reviewed by Adam Barth.
sheriffbot can't roll out security patches
https://bugs.webkit.org/show_bug.cgi?id=39136
Make SheriffBot determine if it's authorized to view a bug
whose change it wants to rollout before it tries to rollout
the change.
Moreover, make both webkit-patch and Sheriffbot provide human-
readable error messages when they are not authorized to view
a bug and when a bug number is invalid.
Currently, Sheriffbot does not parse Bugzilla bugs for
<bug error="...">, which indicates an error when retrieving
a bug. In particular, error="NotPermitted" if a person (or bot)
is not authorized to view a bug. For such error="NotPermitted" bugs,
Sheriffbot raises an exception when parsing the bug report and
this exception does not explicitly indicate Sheriffbot's lack
of authorization. Instead, Sheriffbot should explicitly check
for the presence <bug error="..."> before operating on a bug
and error with a human-readable message if it's not permitted
to view it.
* Scripts/webkitpy/common/net/bugzilla.py: Added BugzillaError class.
* Scripts/webkitpy/common/net/bugzilla_unittest.py:
- Added unit test test_bug_parsing_for_bugzilla_not_permitted_error().
- Added unit test test_bug_parsing_for_bugzilla_not_found_error().
- Added unit test test_bug_parsing_for_bugzilla_invalid_bug_id_error().
* Scripts/webkitpy/tool/bot/sheriff.py: Modified to catch BugzillaError.
* Scripts/webkitpy/tool/commands/download.py: Ditto.
* Scripts/webkitpy/tool/commands/queues.py: Ditto.
* Scripts/webkitpy/tool/commands/upload.py: Ditto.
* Scripts/webkitpy/tool/steps/closebug.py: Ditto.
* Scripts/webkitpy/tool/steps/obsoletepatches.py: Ditto.
* Scripts/webkitpy/tool/steps/preparechangelog.py: Ditto.
* Scripts/webkitpy/tool/steps/updatechangelogswithreviewer.py: Ditto.
2010-09-27 Adam Barth <abarth@webkit.org>
Reviewed by Eric Seidel.
Expose more more failure information from Buildbot to SheriffBot
https://bugs.webkit.org/show_bug.cgi?id=46697
This patch moves the information about what tests failured closer to
SheriffBot. There are still a couple more patches to go before
SheriffBot can post this information to bugs, but this is a step in
that direction. Yay for unit tests, which caught some bugs in earlier
versions of this patch.
* Scripts/webkitpy/common/net/buildbot.py:
* Scripts/webkitpy/common/net/buildbot_unittest.py:
* Scripts/webkitpy/common/net/regressionwindow.py: Added.
* Scripts/webkitpy/tool/commands/queries.py:
2010-09-27 Eric Seidel <eric@webkit.org>
Unreviewed. Fixing 500 error seen in the status server.
It turns out that a = b = []; b.append(1); then a[0] will be 1!
This should have been obvious to me, I guess, but it was not what I was expecting.
* QueueStatusServer/model/activeworkitems.py:
2010-09-27 Adam Barth <abarth@webkit.org>
Reviewed by Eric Seidel.
commit-queue should reject patches that fail to land
https://bugs.webkit.org/show_bug.cgi?id=46694
This can happen, for example, if there's no reviewer. Without this
patch, we'll keep retrying the patch.
* Scripts/webkitpy/tool/bot/commitqueuetask.py:
* Scripts/webkitpy/tool/bot/commitqueuetask_unittest.py:
2010-09-27 Simon Fraser <simon.fraser@apple.com>
Rubber-stamped by Adele Peterson.
Add a harness for running the CSS 2.1 test suite.
This harness assumes that you've got a local a copy of the suite
adjacent to the 'harness' directory.
The harness lets you go through the tests in chapter order,
denoting pass/fail/skipped for each test.
It uses a client-side database to store the results of testing.
* CSSTestSuiteHarness/harness/harness.css: Added.
* CSSTestSuiteHarness/harness/harness.html: Added.
* CSSTestSuiteHarness/harness/harness.js: Added.
2010-09-27 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
Patch locks should expire if a patch is marked for retry
https://bugs.webkit.org/show_bug.cgi?id=46682
This was part Adam's original CommitQueueTask design,
but support for it was missing from the server.
I added the support, but triggering lock-release based on this
special "retry" status feels a bit strange so I added a FIXME.
I also changed the text in queuestatus.html to say "Lock Acquired"
since "Last Activity" isn't really true. We only update the lock
date when the patch is started, not on every status update.
I also noticed an exception in next-patch, which I fixed by re-writing
the unzip logic in activeworkitems.py again.
* QueueStatusServer/handlers/updatestatus.py:
* QueueStatusServer/model/activeworkitems.py:
* QueueStatusServer/templates/queuestatus.html:
2010-09-27 Tony Chang <tony@chromium.org>
Reviewed by David Levin.
[chromium] fix a warning when compiling DRT on 32-bit linux
https://bugs.webkit.org/show_bug.cgi?id=46641
* DumpRenderTree/chromium/TestEventPrinter.cpp:
(DRTPrinter::handleImage):
2010-09-27 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
queue-status should report when the patch was last locked to a queue
https://bugs.webkit.org/show_bug.cgi?id=46674
This isn't necessarily the best way to expose this information
but having this accessible via the web interface is very
useful until we come up with a nicer way to display this.
I also cleaned up the code in activeworkitems.py a little
to use list comprehensions and to have the code work with
pairs instead of two lists at once. Eventually I think those
item/time pairs need to be their own little helper class.
* QueueStatusServer/handlers/queuestatus.py:
* QueueStatusServer/model/activeworkitems.py:
* QueueStatusServer/templates/queuestatus.html:
2010-09-27 Tony Chang <tony@chromium.org>
Reviewed by David Levin.
[chromium] fix detection of missing fonts on DRT
https://bugs.webkit.org/show_bug.cgi?id=46651
* DumpRenderTree/chromium/TestShellGtk.cpp:
(setupFontconfig): Only print an error message if lohit isn't found in both locations.
2010-09-27 Tony Chang <tony@chromium.org>
Reviewed by Dimitri Glazkov.
add webkit unit tests to the chromium testers
https://bugs.webkit.org/show_bug.cgi?id=46669
* BuildSlaveSupport/build.webkit.org-config/master.cfg:
* Scripts/run-chromium-webkit-unit-tests: Fix a bug where we weren't using the right configuration.
2010-09-27 Kwang Yul Seo <skyul@company100.net>
Unreviewed.
Adding myself to the committers list.
* Scripts/webkitpy/common/config/committers.py:
2010-09-27 Mihai Parparita <mihaip@chromium.org>
Reviewed by Eric Seidel.
webkitpy.common.system.user_unittest.UserTest.test_prompt_with_list spams the console
https://bugs.webkit.org/show_bug.cgi?id=46634
Use OutputCapture to prevent console spam (and also check for expected
output).
* Scripts/webkitpy/common/system/user_unittest.py:
2010-09-27 Eric Seidel <eric@webkit.org>
Reviewed by James Robinson.
start-commit-queue should abort any rebases in progress
https://bugs.webkit.org/show_bug.cgi?id=46640
webkit-patch commit-queue would do this itself, but this just
helps ensure that the commit-queue script is properly updated
before we run it (in the case where a previous rebase is in
progress the git svn rebase would not work before we run the
commit-queue).
* EWSTools/start-commit-queue.sh:
2010-09-18 Martin Robinson <mrobinson@igalia.com>
Reviewed by Gustavo Noronha Silva.
[GTK] Clean up the DRT fonts.conf and switch to Liberation
https://bugs.webkit.org/show_bug.cgi?id=46038
Switch to Liberation fonts for GTK+ layout tests. This will make our
test results metric-compatible with many Windows/Chromium results. It
also simplifies the fonts.conf that we load and prepares the way for
fixing many font bugs.
* DumpRenderTree/gtk/DumpRenderTree.cpp:
(initializeFonts): Initialize a known list of fonts instead of loading
every font in the WEBKIT_TESTFONTS directory.
* DumpRenderTree/gtk/fonts.conf: Removed.
* DumpRenderTree/gtk/fonts/AHEM____.TTF: Copied from WebKitTools/DumpRenderTree/qt/fonts/AHEM____.TTF.
* DumpRenderTree/gtk/fonts/fonts.conf: Added.
* GNUmakefile.am: Modify FONTS_CONF_FILE to be FONTS_CONF_DIR, so that
we can load both fonts.conf and AHEM____.TTF from this location.
* Scripts/old-run-webkit-tests: Remove the check for WEBKIT_TESTFONTS, since
it is no longer used.
2010-09-26 Mihai Parparita <mihaip@chromium.org>
Reviewed by Ojan Vafai.
new-run-webkit-httpd fails if Perl is not installed
https://bugs.webkit.org/show_bug.cgi?id=46602
Make Port._read_configuration catch all exceptions, since trying to run
perl (because of _webkit_build_directory) when it's not installed throws
a WindowsError or OSError, not an IOError (this became an issue after
r68268, since ChromiumPort ends up calling Port.default_configuration in
its constructor, which means that we're ending up on this codepath in
non-NRWT cases too, e.g. for the Chromium NaCl tests, which use
new-run-webkit-httpd).
* Scripts/webkitpy/layout_tests/port/base.py:
* Scripts/webkitpy/layout_tests/port/base_unittest.py:
2010-09-26 Gyuyoung Kim <gyuyoung.kim@samsung.com>
Reviewed by Antonio Gomes.
[EFL] Add setting API to set a local storage database path.
https://bugs.webkit.org/show_bug.cgi?id=45446
Add a setting API to set local storage database path.
* EWebLauncher/main.c:
(on_key_down):
(browserCreate):
(main):
2010-09-26 Sam Weinig <sam@webkit.org>
Reviewed by Anders Carlsson.
Move shared WebKit2 API (used by both bundle and main API) to WebKit2/Shared/API
https://bugs.webkit.org/show_bug.cgi?id=46587
* WebKitTestRunner/InjectedBundle/InjectedBundle.h:
2010-09-26 Antonio Gomes <agomes@rim.com>
Reviewed by Kenneth Rohde Christiansen.
DRT/Mac nodesFromRect support
[Mac][DRT] Implement LayoutTestController::nodesFromRect
https://bugs.webkit.org/show_bug.cgi?id=46580
Implements LayoutTestController::nodesFromRect for Mac DRT, and
adding stubs for GTK+ and Windows.
* DumpRenderTree/LayoutTestController.cpp:
(nodesFromRectCallback):
(LayoutTestController::staticFunctions):
* DumpRenderTree/LayoutTestController.h:
(LayoutTestController::LayoutTestController::nodesFromRect):
* DumpRenderTree/mac/LayoutTestControllerMac.mm:
(LayoutTestController::nodesFromRect):
2010-09-25 Nicolas Weber <thakis@chromium.org>
Reviewed by Adam Barth.
https://bugs.webkit.org/show_bug.cgi?id=46555
Fix typo in prepare-ChangeLog's help output.
* Scripts/prepare-ChangeLog:
2010-09-24 Sam Weinig <sam@webkit.org>
Reviewed by Darin Adler.
Implement WebKit2 callback equivalent to -[WebUIDelegate mouseDidMoveOverElement:modifierFlags:]
<rdar://problem/8359279>
https://bugs.webkit.org/show_bug.cgi?id=46546
* MiniBrowser/mac/BrowserWindowController.m:
(mouseDidMoveOverElement):
(-[BrowserWindowController awakeFromNib]):
* MiniBrowser/win/BrowserView.cpp:
(mouseDidMoveOverElement):
(BrowserView::create):
* WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
(WTR::InjectedBundlePage::InjectedBundlePage):
* WebKitTestRunner/TestController.cpp:
(WTR::createOtherPage):
(WTR::TestController::initialize):
2010-09-24 Mihai Parparita <mihaip@chromium.org>
Reviewed by Adam Barth.
Allow rebaselines for webkit-patch rebaseline to be chosen
https://bugs.webkit.org/show_bug.cgi?id=46407
Instead of always rebaselining all failing tests, allow a subset to be
chosen.
* Scripts/webkitpy/common/system/user.py:
* Scripts/webkitpy/tool/commands/rebaseline.py:
2010-09-24 Ryosuke Niwa <rniwa@webkit.org>
Reviewed by Tony Chang.
[chromium] Implement TextInputController::firstRectForCharacterRange
https://bugs.webkit.org/show_bug.cgi?id=38100
Implemented TextInputController::firstRectForCharacterRange for chromium platform.
No new tests are added since we need to implement the same feature in chromium side
in order to enable any tests that uses this function.
* DumpRenderTree/chromium/TextInputController.cpp:
(TextInputController::firstRectForCharacterRange): Added.
2010-09-24 Adam Barth <abarth@webkit.org>
Reviewed by Eric Seidel.
commit-queue reports land failures as "PASS"
https://bugs.webkit.org/show_bug.cgi?id=46530
We were ignoring the return value of land instead of passing it back to
CommitQueue. Of course, this was the one case I forgot to test!
* Scripts/webkitpy/tool/bot/commitqueuetask.py:
* Scripts/webkitpy/tool/bot/commitqueuetask_unittest.py:
2010-09-24 Kenichi Ishibashi <bashi@google.com>
Reviewed by Eric Seidel.
new-run-webkit-tests prints out nothing when build-dumprendertree fails
https://bugs.webkit.org/show_bug.cgi?id=37563
Print error message when build-dumprendertree fails.
* Scripts/webkitpy/layout_tests/port/webkit.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2010-09-24 Mihai Parparita <mihaip@chromium.org>
Reviewed by Tony Chang.
NRWT doesn't respect config set with set-webkit-configuration
https://bugs.webkit.org/show_bug.cgi?id=46278
Use Port.default_configuration() instead of hardcoding Release in
ChromiumPort configuration initialization, so that we still inherit the
configuration set by set-webkit-configuration.
* Scripts/webkitpy/layout_tests/port/chromium.py:
* Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
2010-09-24 Chang Shu <chang.shu@nokia.com>
Reviewed by Csaba Osztrogonác.
[Qt] Fix the code to check file existence.
https://bugs.webkit.org/show_bug.cgi?id=46465
* QtTestBrowser/main.cpp:
(main):
2010-09-23 Tony Chang <tony@chromium.org>
Unreviewed, rolling out r68232.
http://trac.webkit.org/changeset/68232
Broken NRWT on the canary bots.
* Scripts/webkitpy/layout_tests/port/chromium.py:
2010-09-23 Dirk Pranke <dpranke@chromium.org>
Unreviewed, build fix.
r68008 broke new-run-webkit-tests in that the chromium ports no
longer respect set-webkit-configuration. The correct fix for this
is being pursued in bug 46278 (along with a bunch of unit tests),
but in the meantime I'm reverting the particular lines that broke
things. This was tested by hand.
* Scripts/webkitpy/layout_tests/port/chromium.py:
2010-09-23 Sam Weinig <sam@webkit.org>
Reviewed by Anders Carlsson.
WebKit2 API: Need way to know when a frame is removed from the hierarchy
<rdar://problem/8414062>
https://bugs.webkit.org/show_bug.cgi?id=46432
* MiniBrowser/mac/BrowserWindowController.m:
(didRemoveFrameFromHierarchy):
(-[BrowserWindowController awakeFromNib]):
* WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
(WTR::InjectedBundlePage::InjectedBundlePage):
* WebKitTestRunner/TestController.cpp:
(WTR::TestController::initialize):
2010-09-23 Tony Chang <tony@chromium.org>
Reviewed by Dimitri Glazkov.
[chromium] implement layoutTestController.sampleSVGAnimationForElementAtTime
https://bugs.webkit.org/show_bug.cgi?id=46426
* DumpRenderTree/chromium/LayoutTestController.cpp:
(LayoutTestController::LayoutTestController):
(LayoutTestController::sampleSVGAnimationForElementAtTime):
* DumpRenderTree/chromium/LayoutTestController.h:
2010-09-23 Martin Robinson <mrobinson@igalia.com>
Reviewed by Nate Chapin.
[GTK] r68199 introduced two test failures
https://bugs.webkit.org/show_bug.cgi?id=46424
Fix a regression handling preference overrides that are attached to boolean properties.
* DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
(LayoutTestController::overridePreference): g_strcasecmp returns 0 when strings match
so we must check for that when converting a string to a boolean.
2010-09-23 Dimitri Glazkov <dglazkov@chromium.org>
Reviewed by David Levin.
Fix a condition check in ServerProcess, which becomes relevant when reading binary data.
https://bugs.webkit.org/show_bug.cgi?id=46406
This breaks when the method is called with specified size of data (image data), and the
buffer hasn't yet reached this size.
* Scripts/webkitpy/layout_tests/port/server_process.py: Changed to check for values of
index larger than 0.
2010-09-23 Martin Robinson <mrobinson@igalia.com>
Reviewed by Nate Chapin.
[GTK] Some tests from r68174 fail on the GTK+ bots
https://bugs.webkit.org/show_bug.cgi?id=46396
Simplify LayoutTestController::overridePreference to make it easier for
people unfamiliar with the code to keep the list of preferences up to date.
Add the conversion for enable-hyperlink-auditing.
* DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
(LayoutTestController::overridePreference): Use a simple if-else block to determine
out the property name for overrridePreference. Also simplify the logic for setting
string properties. Add the conversion for enable-hyperlink-auditing.
2010-09-23 Tony Chang <tony@chromium.org>
Reviewed by David Levin.
[chromium] add caretBrowsingEnabled to WebSettings and DRT
https://bugs.webkit.org/show_bug.cgi?id=46388
* DumpRenderTree/chromium/LayoutTestController.cpp:
(LayoutTestController::overridePreference):
* DumpRenderTree/chromium/WebPreferences.cpp:
(WebPreferences::reset):
(WebPreferences::applyTo):
* DumpRenderTree/chromium/WebPreferences.h:
2010-09-23 Nate Chapin <japhet@chromium.org>
Reviewed by Darin Fisher.
Allow DRT to toggle hyperlink auditing (i.e., <a ping>).
https://bugs.webkit.org/show_bug.cgi?id=30458
* DumpRenderTree/chromium/LayoutTestController.cpp:
(LayoutTestController::overridePreference):
* DumpRenderTree/chromium/WebPreferences.cpp:
(WebPreferences::reset):
(WebPreferences::applyTo):
* DumpRenderTree/chromium/WebPreferences.h:
* DumpRenderTree/gtk/DumpRenderTree.cpp:
(resetDefaultsToConsistentValues):
* DumpRenderTree/qt/LayoutTestControllerQt.cpp:
(LayoutTestController::overridePreference):
2010-09-21 Stephen White <senorblanco@chromium.org>
Reviewed by David Levin.
Implement --enable-accelerated-2d-canvas flag in DumpRenderTree.
https://bugs.webkit.org/show_bug.cgi?id=46208
This flag allows the layout tests to be run with or without accelerated
2D canvas rendering.
* DumpRenderTree/chromium/DumpRenderTree.cpp:
(main):
Declare the new flag string, and check for it on startup.
* DumpRenderTree/chromium/TestShell.cpp:
(TestShell::TestShell):
Add a boolean for the new flag, in order to preserve its value over
preferences reset.
(TestShell::resetWebSettings):
Set the new flag to the stored value on reset.
* DumpRenderTree/chromium/TestShell.h:
(TestShell::setAccelerated2dCanvasEnabled):
Add an accessor for the new flag.
* DumpRenderTree/chromium/WebPreferences.cpp:
(WebPreferences::reset):
Initialize the new flag to false.
(WebPreferences::applyTo):
Copy the flag's value to the WebSettings.
* DumpRenderTree/chromium/WebPreferences.h:
Add the new flag.
2010-09-23 Tor Arne Vestbø <tor.arne.vestbo@nokia.com>
Reviewed by Simon Hausmann.
[Qt] Refactor QtWebKitPlatformPlugin interface
Make it easier to keep source-compability for the
QtWebKitPlatformPlugin interface, and run qmake
on the example (but not build) for convenience.
https://bugs.webkit.org/show_bug.cgi?id=46345
* Scripts/webkitdirs.pm:
2010-09-23 Adam Barth <abarth@webkit.org>
Reviewed by Eric Seidel.
Tweak some status messages that Eric thought were confusing
https://bugs.webkit.org/show_bug.cgi?id=46342
* Scripts/webkitpy/tool/commands/queues.py:
* Scripts/webkitpy/tool/commands/queues_unittest.py:
2010-09-22 Adam Barth <abarth@webkit.org>
Reviewed by Eric Seidel.
Fix comm-queue typo
https://bugs.webkit.org/show_bug.cgi?id=46339
We were missing a "self". The real problem is that we didn't have an
integration test for the failure case.
* Scripts/webkitpy/tool/commands/queues.py:
* Scripts/webkitpy/tool/commands/queues_unittest.py:
2010-09-22 Adam Barth <abarth@webkit.org>
Reviewed by Eric Seidel.
commit-queue should log more detailed messages to the QueueStatusServer
https://bugs.webkit.org/show_bug.cgi?id=46333
When I created CommitQueueTask, I removed most of the previous logging.
This patch adds back more detailed logging so folks can see their patch
progress through the queue.
* Scripts/webkitpy/tool/bot/commitqueuetask.py:
* Scripts/webkitpy/tool/bot/commitqueuetask_unittest.py:
* Scripts/webkitpy/tool/commands/queues.py:
* Scripts/webkitpy/tool/commands/queues_unittest.py:
2010-09-22 Mihai Parparita <mihaip@chromium.org>
Reviewed by Darin Fisher.
[Chromium] User style layout tests don't pass on Chromium
https://bugs.webkit.org/show_bug.cgi?id=46069
Fix a typo in LayoutTestController::addUserStyleSheet that was causing a
crash the Chromium DRT. Pass InjectInExistingDocuments to mimic DRT
behavior from other ports.
* DumpRenderTree/chromium/LayoutTestController.cpp:
(LayoutTestController::addUserStyleSheet):
2010-09-22 Brent Fulgham <bfulgham@webkit.org>
Reviewed by Martin Robinson.
[WinCairo] Part 2: Update WebKitTestRunner and DumpRenderTree Build.
https://bugs.webkit.org/show_bug.cgi?id=46303.
* MiniBrowser/Configurations/MiniBrowserCFLite.vsprops: Added.
* MiniBrowser/Configurations/MiniBrowserCommon.vsprops: Moved
CoreFoundation-specific stuff to new MiniBrowserCoreFoundation
property sheet.
* MiniBrowser/Configurations/MiniBrowserCoreFoundation.vsprops: Added.
* MiniBrowser/MiniBrowser.vcproj: Updated configuration to use
appropriate property sheet for the build types.
* WebKitTestRunner/win/WebKitTestRunner.vcproj: Updated the
configuration to use appropriate property sheet for CoreFoundation
and CFLite-style builds.
2010-09-22 Adam Barth <abarth@webkit.org>
Reviewed by Eric Seidel.
Optimize commit-queue performance for green trees
https://bugs.webkit.org/show_bug.cgi?id=46254
This patch redesigns the controller logic for the commit-queue. In the
new design, the controller exercises much finer-grained control over
the landing process. In particular:
- Patches that fail to apply now get rejected almost immediately.
- Patches that fail to build get rejects after two builds (instead of
three builds and one test run).
- Patches that run into a flaky test now get accepted after one build
and two test runs instead of three full build-and-test runs.
The main cost of these optimizations is that we don't find out the tree
has a failing test until the very end of the process, but if the tree
has a busted test, there's not much we can do anyway. We might as well
burn commit-queue resources spinning optimisticly.
* Scripts/webkitpy/tool/bot/commitqueuetask.py: Added.
* Scripts/webkitpy/tool/bot/commitqueuetask_unittest.py: Added.
* Scripts/webkitpy/tool/commands/queues.py:
2010-09-22 Brent Fulgham <bfulgham@webkit.org>
Reviewed by Martin Robinson.
[WinCairo] Update WebKitTestRunner and DumpRenderTree Build.
https://bugs.webkit.org/show_bug.cgi?id=46303.
* DumpRenderTree/win/DumpRenderTree.vcproj: Disable local MD5 sources
for Cairo build. Copy wtf MD5 header to ForwardingHeaders for the
WinCairo build.
* WebKitTestRunner/Configurations/InjectedBundleCFLite.vsprops: Added.
* WebKitTestRunner/win/InjectedBundle.vcproj: Update win new *_Cairo
build targets that use the new InjectedBundleCFLite.vsprops file.
2010-09-22 Dirk Pranke <dpranke@chromium.org>
Unreviewed, build fix.
Fix bustage of rebaseline-chromium-webkit-tests resulting from
r67974. Really need better unit tests for this tool :(
* Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
2010-09-22 Adam Roben <aroben@apple.com>
Unbreak test-webkitpy
* Scripts/webkitpy/test/main.py:
(Tester.run_tests): Add a line that mysteriously got deleted.
2010-09-22 Adam Roben <aroben@apple.com>
Make test-webkitpy test WebKit2's scripts
These scripts can't be in WebKitTools due to limitations of Apple's
build process. But that doesn't mean we can't test them!
Fixes <http://webkit.org/b/46297> test-webkitpy should test code in
WebKit2/Scripts
Reviewed by Adam Barth.
* Scripts/test-webkitpy:
(_clean_packages_with_test): Renamed from _clean_webkitpy_with_test.
Now takes an external_package_paths parameter and cleans both webkitpy
and any external packages.
(init): Added an external_package_paths parameter which we pass along
to _clean_packages_with_test.
(top level): Add WebKit2/Scripts/webkit2 as our only external package
and pass it along to init and Tester.run_tests.
* Scripts/webkitpy/test/main.py:
(Tester.run_tests): Added an optional external_package_paths
parameter. We modify sys.path so that the external packages can be
imported, and search for unittest files inside all external packages
in addition to inside webkitpy.
2010-09-22 Dirk Pranke <dpranke@chromium.org>
Reviewed by Tony Chang.
new-run-webkit-tests: r68063 broke linux python tests
Add a missing "from __future__ import with_statement" to this
new file; the perils of testing only on the Mac and by review :(
https://bugs.webkit.org/show_bug.cgi?id=46293
* Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
2010-09-22 Dirk Pranke <dpranke@chromium.org>
Reviewed by Tony Chang.
Add a 'chromium-gpu' set of ports that will test the accelerated
GPU paths. This patch adds:
- support for the '--accelerated-compositing' and
'accelerated-2d-canvas' flags to new-run-webkit-tests (and the
'no-' flags)
- adds a new set of Ports that will look under
platform/chromium-gpu-$OS/ for baselines before looking in the
regular chromium search path
- adds a new test_expectations.txt file in platform/chromium-gpu
that skips all but the tests we actually want to run with
acceleration.
This patch will allow us to run both with and without acceleration
and to change the defaults for both the regular and -gpu options
as the code evolves.
We plan to add both --chromium-$OS and --chromium-gpu-$OS runs to
each test bot.
https://bugs.webkit.org/show_bug.cgi?id=46225
* Scripts/webkitpy/layout_tests/port/chromium.py:
* Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
* Scripts/webkitpy/layout_tests/port/factory.py:
* Scripts/webkitpy/layout_tests/port/webkit.py:
2010-09-22 Jamey Hicks <jamey.hicks@nokia.com>
Reviewed by Kenneth Rohde Christiansen.
[Qt] Web Inspector: Remote Web Inspector support for QtWebKit
https://bugs.webkit.org/show_bug.cgi?id=43988
Runs a web debug server on port specified by QtTestBrowser
argument -remote-inspector-port. Property
_q_webInspectorServerPort of the QWebPage instance will be set
according to the argument. All pages with that property set will
be remotely inspectable.
URL for remote inspection of first QWebPage is
http://localhost:9222/webkit/inspector/inspector.html?page=1
where 1 is the number of the QWebPage instance.
The base URL yields an index page with links to the individual inspectors:
http://localhost:9222/
* QtTestBrowser/launcherwindow.cpp:
(LauncherWindow::init):
* QtTestBrowser/launcherwindow.h:
* QtTestBrowser/main.cpp:
(LauncherApplication::handleUserOptions):
2010-09-22 Adam Roben <aroben@apple.com>
Fix webkit-patch failure-reason now that build.webkit.org has been
updated
Fixes <http://webkit.org/b/46273> webkit-patch failure-reason crashes
every time
Reviewed by Eric Seidel.
* Scripts/webkitpy/common/net/buildbot.py:
(BuildBot._file_cell_text): Added. Travels down the firstChild chain
looking for an element that contains text, then returns it.
(BuildBot._parse_twisted_file_row): Use _file_cell_text to get the
text out of the cells. This way it doesn't matter whether the cells
have <b> children (as for cells in directory rows) or not (as for
cells in file rows)
(BuildBot._parse_twisted_directory_listing): Look for rows that have
the "directory" or "file" class, rather than rows with any class,
since header rows now have a class attribute.
* Scripts/webkitpy/common/net/buildbot_unittest.py:
(BuildBotTest._example_directory_listing): Updated to more closely
match the markup that build.webkit.org is producing now.
2010-09-21 Dirk Pranke <dpranke@chromium.org>
Unreviewed, build fix.
new-run-webkit-tests: fix bug introduced in r68008 where if you
specify --chromium and no --platform, and you're running on
windows, you use the 'chromium-win' port by default instead of the
version-specific port. This breaks the buildbots.
It will be good for this refactoring to settle down so I can
rewrite the logic for default ports and the unit tests to be
clearer (and the testing more comprehensive).
* Scripts/webkitpy/layout_tests/port/chromium_win.py:
2010-09-21 Dirk Pranke <dpranke@chromium.org>
Reviewed by Adam Barth.
Modify the Port interface to take only a series of keyword arguments
in the constructor, and modify Port/factory.get() to accomodate that,
and to accept user=XXX as an argument so we can pass
webkitpy.common.system.user.User objects in.
Then, modify new-run-webkit-tests and rebaseline-chromium-webkit-tests
to use the common routine in webkitpy.common.system.user.open_url()
to display HTML files.
There was a routine in the Port interface to do the same thing,
but I see no need for a port-specific hook for this, since it is
something that will always be executed by the host environment
and displaying web pages has nothing to do with running layout tests.
Note that new-run-webkit-tests used to use test_shell to display
the page; this is potentially useful so that you can actually click
from a result to the broken page; however, since DumpRenderTree
doesn't support this functionality, it will be going away eventually.
https://bugs.webkit.org/show_bug.cgi?id=46128
* Scripts/webkitpy/layout_tests/port/base.py:
* Scripts/webkitpy/layout_tests/port/base_unittest.py:
* Scripts/webkitpy/layout_tests/port/chromium.py:
* Scripts/webkitpy/layout_tests/port/chromium_linux.py:
* Scripts/webkitpy/layout_tests/port/chromium_mac.py:
* Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
* Scripts/webkitpy/layout_tests/port/chromium_win.py:
* Scripts/webkitpy/layout_tests/port/dryrun.py:
* Scripts/webkitpy/layout_tests/port/factory.py:
* Scripts/webkitpy/layout_tests/port/google_chrome.py:
* Scripts/webkitpy/layout_tests/port/google_chrome_unittest.py:
* Scripts/webkitpy/layout_tests/port/gtk.py:
* Scripts/webkitpy/layout_tests/port/mac.py:
* Scripts/webkitpy/layout_tests/port/qt.py:
* Scripts/webkitpy/layout_tests/port/test.py:
* Scripts/webkitpy/layout_tests/port/webkit.py:
* Scripts/webkitpy/layout_tests/port/win.py:
* Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2010-09-21 Lucas De Marchi <lucas.demarchi@profusion.mobi>
Reviewed by Kenneth Rohde Christiansen.
[EFL] Get fresh theme when running EWebLauncher
https://bugs.webkit.org/show_bug.cgi?id=46210
If user tried to run EWebLauncher outside of the build tree, it would
not find the theme and fallback to the installed one. However, since
EWebLauncher is not installed, we always want to take the theme from
the just compiled source code. If user had never installed WebKit's
EFL port it could even receive a segv since no theme would be found.
Now EWebLauncher does not fallback to the installed theme and CMake
gives as DATA_DIR the directory of the theme it has just built.
* CMakeListsEfl.txt: Pass the build directory as DATA_DIR to
EWebLauncher.
* EWebLauncher/main.c: Use only the theme from build director. Do not
fallback to others as this could hide real bugs.
(quit):
(browserCreate):
(findThemePath):
(main):
2010-09-21 Dirk Pranke <dpranke@chromium.org>
Unreviewed, build fix.
Don't pass the --test-shell arg to the Chromium Mac port of DRT;
it just confuses it.
https://bugs.webkit.org/show_bug.cgi?id=46230
* Scripts/webkitpy/layout_tests/port/webkit.py:
2010-09-21 Lucas Forschler <lforschler@apple.com>
Reviewed by Stephanie Lewis.
Make a new buildbot for Leopard Debug Test
* BuildSlaveSupport/build.webkit.org-config/config.json:
2010-09-21 Dirk Pranke <dpranke@chromium.org>
Unreviewed, build fix.
Fix breakage of Chromium Mac DRT port caused by r67905.
https://bugs.webkit.org/show_bug.cgi?id=46230
* Scripts/webkitpy/layout_tests/port/webkit.py:
2010-09-21 Dirk Pranke <dpranke@chromium.org>
Unreviewed, build fix.
Add missing import of 'tempfile'.
* Scripts/webkitpy/layout_tests/port/chromium.py:
2010-09-17 Dirk Pranke <dpranke@chromium.org>
Reviewed by Ojan Vafai.
new-run-webkit-tests: pull the list of tests from the Port, and
make it possible to run with no actual test files in the filesystem.
This set of changes allows us to delete the special
webkitpy/layout_tests/data directory and test the generic code without
touching the filesystem (except to write results). This speeds up
test-webkitpy substantially.
This also cleans up and removes several code paths in the generic
code and fixes some minor bugs, notably in the test_expectations
parsing, which was previously fragile if the tests weren't present.
We also change the way we diff image results to be able to do so
in memory, without relying on files. This slows down chromium
test_shell, which always writes images to files, but should speed
up DRT and ImageDiff-based implementations slightly.
Note that pulling the list of tests to run from the Port will allow
ports to specify what tests to run as an inclusive list; previously
you could only do this as an exclusive list by using a
test_expectations file or Skipped files; nothing actually uses this
feature yet and it's unclear if it should be used.
Note that there are no functional changes -- apart from now
always printing out the location of the checksum file when we are
tracing test runs -- and the total number of lines of non-test code
actually drops by two.
There is some more cleanup that can be done in the Port/Driver
interface and in some handling of filenames, but I'll save that
for another change.
https://bugs.webkit.org/show_bug.cgi?id=45801
* Scripts/webkitpy/layout_tests/data/*: Removed.
- no longer need special mock layout_tests_directory in the
repository.
* Scripts/webkitpy/layout_tests/layout_package/printing.py:
- add code to display missing text files, checksums when tracing
- update to not look at the filesystem directly.
* Scripts/webkitpy/layout_tests/layout_package/printing_unittest.py:
- add more unit tests
* Scripts/webkitpy/layout_tests/layout_package/test_expectations.py:
- remove tests_are_present flag
- update with changes in Port interface - no longer access
the filesystem directly, although we still use os.path for
filename manipulation.
* Scripts/webkitpy/layout_tests/layout_package/test_expectations_unittest.py:
- add more unit tests
* Scripts/webkitpy/layout_tests/layout_package/test_files.py: Removed.
- renamed to port/test_files.py
* Scripts/webkitpy/layout_tests/port/base.py:
- change diff_image() to expect actual image data, not filenames
- add expected_checksum(), expected_image(), expected_text() to
return the content of the files so that we don't need a filesystem
- add path_exists(), path_isdir() for filesystem-like access.
- add test_dirs() to keep clobber-test-results from needing to
actually look at a real layout tests dir
- add tests() to return the list of tests to run on the port
(calls port/test_files to do the work).
- add update_baseline() to be able to save new baselines
- add uri_to_test_name() from port/dryrun.py so we can easily check
filename_to_uri()
* Scripts/webkitpy/layout_tests/port/base_unittest.py:
- add more unit tests
* Scripts/webkitpy/layout_tests/port/chromium.py:
- change diff_image() to accept image content, not filenames.
This will be slower for test_shell/image_diff, but faster
for DRT/ImageDiff.
* Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
- add more unit tests
* Scripts/webkitpy/layout_tests/port/dryrun.py:
- simplify greatly
* Scripts/webkitpy/layout_tests/port/port_testcase.py:
- add more unit tests
* Scripts/webkitpy/layout_tests/port/test.py:
- massive rewrite to use in-script list of tests and expectations
* Scripts/webkitpy/layout_tests/port/test_files.py:
- rename from layout_package/test_files.
* Scripts/webkitpy/layout_tests/port/test_files_unittest.py:
- add unit tests
* Scripts/webkitpy/layout_tests/port/webkit.py:
- update diff_image() to take image contents, not files. Should
make things slightly faster.
* Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
- update with changes to diff_image()
* Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests_unittest.py:
- update with changes to diff_image()
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
- remove tests_are_present from test_expectations
- pull the list of tests from port.tests() instead
of calling test_files.py directly.
* Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
- update unit tests
* Scripts/webkitpy/layout_tests/test_types/image_diff.py:
- update with changes to
* Scripts/webkitpy/layout_tests/test_types/test_type_base.py:
- update with change in Port interface
* Scripts/webkitpy/layout_tests/test_types/text_diff.py:
- update with change in Port interface
* Scripts/webkitpy/style/checkers/test_expectations.py:
- remove the tests_are_present flag
2010-09-21 Anders Carlsson <andersca@apple.com>
Disable logging.
* MiniBrowser/mac/MiniBrowser_Prefix.pch:
2010-09-21 Tony Chang <tony@chromium.org>
Reviewed by Kent Tamura.
add a script for running webkit_unit_tests
https://bugs.webkit.org/show_bug.cgi?id=46014
* Scripts/run-chromium-webkit-unit-tests: Added.
2010-09-21 Fridrich Strba <fridrich.strba@bluewin.ch>
Reviewed by Martin Robinson.
Fix linking problems on Windows.
https://bugs.webkit.org/show_bug.cgi?id=45844
* GNUmakefile.am: link the executables with winmm.dll
2010-09-21 Adam Barth <abarth@webkit.org>
Reviewed by Eric Seidel.
feeder-queue svn updates more often than needed
https://bugs.webkit.org/show_bug.cgi?id=46171
We don't need to poll SVN every time we feed the feeders. Rather, we
can count on the wrapper shell script to auto-update the queue.
* Scripts/webkitpy/tool/commands/queues.py:
* Scripts/webkitpy/tool/commands/queues_unittest.py:
2010-09-21 Jocelyn Turcotte <jocelyn.turcotte@nokia.com>
Reviewed by Andreas Kling.
[Qt] Add robot loader timeout and extra time options.
https://bugs.webkit.org/show_bug.cgi?id=46172
[-robot-timeout <s>]: Load the next page after s seconds if the current
page didn't finish loading.
[-robot-extra-time <s>]: Wait s seconds after the current page finished
loading before loading the next one. This should allow some time for the
page's JavaScript to execute.
* QtTestBrowser/launcherwindow.cpp:
(LauncherWindow::applyPrefs):
* QtTestBrowser/main.cpp:
(LauncherApplication::robotTimeout):
(LauncherApplication::robotExtraTime):
(LauncherApplication::LauncherApplication):
(LauncherApplication::handleUserOptions):
(main):
* QtTestBrowser/urlloader.cpp:
(UrlLoader::UrlLoader):
(UrlLoader::loadNext):
(UrlLoader::loadUrlList):
* QtTestBrowser/urlloader.h:
2010-09-21 Pavel Podivilov <podivilov@chromium.org>
Unreviewed.
Adding myself to the committers list.
* Scripts/webkitpy/common/config/committers.py:
2010-09-20 Philippe Normand <pnormand@igalia.com>
Reviewed by Eric Carlson.
[GTK] enhanced context menu for media elements
https://bugs.webkit.org/show_bug.cgi?id=45021
EventSender::contextClick() now returns an array of js
objects. Each object has a title property and a click() method.
* DumpRenderTree/gtk/EventSender.cpp:
(getMenuItemTitleCallback):
(setMenuItemTitleCallback):
(menuItemClickCallback):
(getMenuItemClass):
(contextClickCallback):
2010-09-20 Hayato Ito <hayato@chromium.org>
Unreviewed.
Adding myself to the committers list.
* Scripts/webkitpy/common/config/committers.py:
2010-09-20 Adam Barth <abarth@webkit.org>
Reviewed by Eric Seidel.
The tool member variable should be called _tool
https://bugs.webkit.org/show_bug.cgi?id=46160
Created by find-and-replace.
* Scripts/webkitpy/tool/commands/download.py:
* Scripts/webkitpy/tool/commands/openbugs.py:
* Scripts/webkitpy/tool/commands/queries.py:
* Scripts/webkitpy/tool/commands/queues.py:
* Scripts/webkitpy/tool/commands/queues_unittest.py:
* Scripts/webkitpy/tool/commands/rebaseline.py:
* Scripts/webkitpy/tool/commands/sheriffbot.py:
* Scripts/webkitpy/tool/commands/upload.py:
* Scripts/webkitpy/tool/multicommandtool.py:
2010-09-20 Adam Barth <abarth@webkit.org>
I ran the tests before landing but ignored the fact that they failed. :(
* Scripts/webkitpy/tool/commands/queues.py:
2010-09-20 Adam Barth <abarth@webkit.org>
Reviewed by Eric Seidel.
Add a feeder queue that polls bugs.webkit.org for the commit-cluster
https://bugs.webkit.org/show_bug.cgi?id=46141
The feeder-queue polls bugs.webkit.org every 30 seconds and updates the
list of work items on the status server. The individual commit-cluster
nodes then grab the patches from the server and process them.
* Scripts/webkitpy/tool/bot/feeders.py: Added.
* Scripts/webkitpy/tool/bot/feeders_unittest.py: Added.
* Scripts/webkitpy/tool/commands/queues.py:
* Scripts/webkitpy/tool/commands/queues_unittest.py:
* Scripts/webkitpy/tool/commands/queuestest.py:
2010-09-20 Leandro Pereira <leandro@profusion.mobi>
Reviewed by Darin Adler.
build-webkit: Add support for CMake build system
https://bugs.webkit.org/show_bug.cgi?id=44979
* Scripts/build-webkit: Add "--efl" command-line option to build the
EFL port of WebKit.
* Scripts/webkitdirs.pm: Define buildCMakeProject() and
buildEflCMakeProject() subroutines.
2010-09-20 Dirk Pranke <dpranke@chromium.org>
Reviewed by Ojan Vafai.
new-run-webkit-tests: refactor command line args getting passed to DRT
This change cleans up some argument parsing between functions to get
rid of some overlapping data structures. There should be no functional
changes in this patch; it is pure refactoring in preparation for
landing the Chrome GPU port and adding a generic way to pass
args to DRT/TestShell.
https://bugs.webkit.org/show_bug.cgi?id=46135
* Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
- pass the options argument explicitly to the threads and drivers,
also consolidate the passing of options to the driver.
- pass options directly to process_output() to remove a couple
parameters (minor cleanup).
* Scripts/webkitpy/layout_tests/port/base.py:
- pass the options argument to Port.create_driver().
* Scripts/webkitpy/layout_tests/port/base_unittest.py:
- update Port.create_driver() test
* Scripts/webkitpy/layout_tests/port/chromium.py:
- pass the options argument to Port.create_driver(), and clean up
building of the cmd line for DRT.
* Scripts/webkitpy/layout_tests/port/dryrun.py:
- pass the options argument to Port.create_driver()
* Scripts/webkitpy/layout_tests/port/test.py:
- pass the options argument to Port.create_driver()
* Scripts/webkitpy/layout_tests/port/webkit.py:
- pass the options argument to Port.create_driver(), and clean up
building of the cmd line for DRT.
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
- consolidate args in _get_dump_render_tree_args and rename to
_get_test_args(); move all of the command-line args to the
Port implementations.
2010-09-20 Andrew Wilson <atwilson@chromium.org>
Revert change which was accidentally committed along with some expectation changes.
* Scripts/webkitpy/layout_tests/port/base.py:
2010-09-20 Mihai Parparita <mihaip@chromium.org>
Reviewed by Adam Barth.
Add explicit --force-patch flag to webkitpy tools
https://bugs.webkit.org/show_bug.cgi?id=46103
It wasn't obvious until I read the applypatch.py source code that
--non-interfactive implies passing --force to patch. Add --force-patch
as an alias to the flag, so that this behavior is more discoverable.
* Scripts/webkitpy/tool/steps/applypatch.py:
* Scripts/webkitpy/tool/steps/options.py:
2010-09-20 Eric Seidel <eric@webkit.org>
Unreviewed, adding a line which got dropped and is trivially correct (and tested).
commit-queue can't land patches
https://bugs.webkit.org/show_bug.cgi?id=46138
Add a line of code I forgot in my last commit.
* Scripts/webkitpy/common/net/bugzilla.py:
* Scripts/webkitpy/common/net/bugzilla_unittest.py:
2010-09-20 Adam Barth <abarth@webkit.org>
Reviewed by Eric Seidel.
The commit-cluster bots still race to lock patch_ids
https://bugs.webkit.org/show_bug.cgi?id=46130
It turns out we need to use a transaction object to make the
read/modify/write lock operation atomic. From reading the AppEngine
documentation, I think this patch should do what we want. It's hard to
test locally because the test instance isn't distributed in the same
way the production instance is.
* QueueStatusServer/handlers/nextpatch.py:
* QueueStatusServer/model/activeworkitems.py: Added.
2010-09-20 Andy Estes <aestes@apple.com>
Reviewed by Adam Barth.
REGRESSION (HTML5 Parser): Pages broken due to <tag<tag> parsing changes
https://bugs.webkit.org/show_bug.cgi?id=40961
Implement WebKitUsePreHTML5ParserQuirks preference.
* DumpRenderTree/mac/DumpRenderTree.mm:
(resetDefaultsToConsistentValues): Reset WebKitUsePreHTML5ParserQuirks
to false after a test finishes.
2010-09-20 Adam Roben <aroben@apple.com>
Windows build fix
* MiniBrowser/Configurations/MiniBrowserCommon.vsprops: Put
$(WebKitOutputDir)\include before $(WebKitLibrariesDir)\include so
we'll pick up the most recent versions of the headers.
2010-09-20 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
commit-queue should check commit-queue+ again just before committing
https://bugs.webkit.org/show_bug.cgi?id=32679
Added a _revalidate_patch check, right before landing.
Since _revalidate_patch passes the patch_id from the work item
back to bugzilla, I had to fix all of the previous queue tests to
use valid attachment ids (that's the majority of this change).
In order to validate that the bug was still open, I had to teach
bugzilla.Bug about open/closed states.
* Scripts/webkitpy/common/net/bugzilla.py:
* Scripts/webkitpy/tool/commands/earlywarningsystem_unittest.py:
* Scripts/webkitpy/tool/commands/queues.py:
* Scripts/webkitpy/tool/commands/queues_unittest.py:
* Scripts/webkitpy/tool/commands/queuestest.py:
* Scripts/webkitpy/tool/mocktool.py:
2010-09-20 Mihai Parparita <mihaip@chromium.org>
Unreviewed.
Adding myself as a comitter.
* Scripts/webkitpy/common/config/committers.py:
2010-09-20 Adam Barth <abarth@webkit.org>
Reviewed by Eric Seidel.
Two instances of commit-queue try to process the same patch
https://bugs.webkit.org/show_bug.cgi?id=46113
This patch makes next-patch atomic so that the server won't vend the
same patch twice in the same hour.
* QueueStatusServer/handlers/nextpatch.py:
2010-09-20 Adam Roben <aroben@apple.com>
Make WebKitTestRunner's wait-to-dump watchdog timer work on Windows
We were previously trying to use a CFRunLoopTimer, but since Windows
doesn't use CFRunLoop on most threads this doesn't work. Now we use a
Windows-style timer on Windows.
I also replaced all uses of "watchdog" with "watchdog timer".
Fixes <http://webkit.org/b/46101> WebKitTestRunner's wait-to-dump
watchdog timer doesn't work on Windows
Reviewed by Anders Carlsson.
* WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
(WTR::InjectedBundlePage::dump): Updated for rename.
* WebKitTestRunner/InjectedBundle/LayoutTestController.cpp: Removed
platform-specific functions.
(WTR::LayoutTestController::LayoutTestController): Added call to
platformInitialize.
(WTR::LayoutTestController::waitUntilDone): Changed to call
initializeWaitToDumpWatchdogTimerIfNeeded instead of doing the work
right in this function.
(WTR::LayoutTestController::waitToDumpWatchdogTimerFired): Updated for
rename.
* WebKitTestRunner/InjectedBundle/LayoutTestController.h: Added new
members needed to support the watchdog timer abstraction. Replaced
some "watchdog"s with "watchdog timer".
* WebKitTestRunner/InjectedBundle/mac/LayoutTestControllerMac.mm:
Added.
(WTR::LayoutTestController::platformInitialize): Does nothing on this
platform.
(WTR::LayoutTestController::invalidateWaitToDumpWatchdog): Moved here
from LayoutTestController.cpp and changed to use an early return.
(WTR::waitUntilDoneWatchdogTimerFired): Moved here from
LayoutTestController.cpp.
(WTR::LayoutTestController::initializeWaitToDumpWatchdogTimerIfNeeded):
Moved code here from LayoutTestController::waitUntilDone and changed
to use an early return.
* WebKitTestRunner/InjectedBundle/win/LayoutTestControllerWin.cpp:
Added.
(WTR::LayoutTestController::platformInitialize): Initialize our
watchdog timer.
(WTR::LayoutTestController::invalidateWaitToDumpWatchdog): Added.
Kills and clears the watchdog timer.
(WTR::waitToDumpWatchdogTimerFired): Added. Calls through to the
LayoutTestController member function of the same name.
(WTR::LayoutTestController::initializeWaitToDumpWatchdogTimerIfNeeded):
Added. Sets up the timer if it isn't already set.
* WebKitTestRunner/WebKitTestRunner.xcodeproj/project.pbxproj: Added
LayoutTestControllerMac.mm.
* WebKitTestRunner/win/InjectedBundle.vcproj: Added
LayoutTestControllerWin.cpp.
2010-09-20 Martin Robinson <mrobinson@igalia.com>
Reviewed by Daniel Bates.
[GTK] fast/forms/listbox-selection.html fails
https://bugs.webkit.org/show_bug.cgi?id=45942
Use the gdkModifersFromJSValue helper to parse all appropriate modifier
strings in keyDownCallback.
* DumpRenderTree/gtk/EventSender.cpp:
(keyDownCallback): Use the gdkModifersFromJSValue instead of duplicating the
modifier parsing logic.
2010-09-19 Sam Weinig <sam@webkit.org>
Reviewed by Anders Carlsson.
WebKit2 decidePolicyForNavigationAction should include mouse button information
<rdar://problem/8413165>
https://bugs.webkit.org/show_bug.cgi?id=46060
* MiniBrowser/mac/BrowserWindowController.m:
(decidePolicyForNavigationAction):
(decidePolicyForNewWindowAction):
2010-09-19 Daniel Bates <dbates@rim.com>
Reviewed by Martin Robinson.
Add unit tests for diffs that delete or modify a change log entry
or describe changes that are far apart
https://bugs.webkit.org/show_bug.cgi?id=46046
Add additional unit tests to test change log diffs that contain
deletions or changes that are far apart from each other in the
ChangeLog file.
* Scripts/webkitperl/VCSUtils_unittest/fixChangeLogPatch.pl:
- Added the following unit tests:
"fixChangeLogPatch: [no change] In-place change."
"fixChangeLogPatch: [no change] Remove first entry."
"fixChangeLogPatch: [no change] Remove entry in the middle."
"fixChangeLogPatch: [no change] Far apart changes (i.e. more than one chunk)."
2010-09-19 Sam Weinig <sam@webkit.org>
Reviewed by Anders Carlsson.
Replace WKBundleRangeRef with WKBundleRangeHandleRef.
https://bugs.webkit.org/show_bug.cgi?id=46054
The new one acts like WKBundleNodeHandleRef and allows for getting a
wrapper in a specific world for the handle.
* WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
(WTR::propertyValue):
(WTR::propertyValueDouble):
(WTR::propertyValueInt):
(WTR::numericWindowPropertyValue):
(WTR::toStr):
(WTR::InjectedBundlePage::shouldBeginEditing):
(WTR::InjectedBundlePage::shouldEndEditing):
(WTR::InjectedBundlePage::shouldInsertNode):
(WTR::InjectedBundlePage::shouldInsertText):
(WTR::InjectedBundlePage::shouldDeleteRange):
(WTR::InjectedBundlePage::shouldChangeSelectedRange):
(WTR::InjectedBundlePage::shouldApplyStyle):
* WebKitTestRunner/InjectedBundle/InjectedBundlePage.h:
2010-09-19 Kent Tamura <tkent@chromium.org>
Unreviewed. Fix WebGL test regressions by r67809.
* DumpRenderTree/chromium/WebPreferences.cpp:
(WebPreferences::applyTo): Call setExperimentalWebGLEnabled().
2010-09-19 Kent Tamura <tkent@chromium.org>
Reviewed by Adam Barth.
[DRT/Chromium] Fix a WebSettings handling bug
https://bugs.webkit.org/show_bug.cgi?id=45945
Before this change, Chromium DRT reset WebSettings for every new
window. It was wrong.
If new window is not the first one, we have to apply the same
settings as the first window. So, we introduce WebPreference to
store the current settings, and apply it to new windows. It's same
as test_shell's behavior.
* DumpRenderTree/DumpRenderTree.gypi:
* DumpRenderTree/chromium/LayoutTestController.cpp:
Use WebPreferences instead of WebSettings.
(LayoutTestController::setUserStyleSheetEnabled):
(LayoutTestController::setUserStyleSheetLocation):
(LayoutTestController::setAuthorAndUserStylesEnabled):
(LayoutTestController::setPopupBlockingEnabled):
(LayoutTestController::disableImageLoading):
(LayoutTestController::setJavaScriptCanAccessClipboard):
(LayoutTestController::setXSSAuditorEnabled):
(LayoutTestController::setAllowUniversalAccessFromFileURLs):
(LayoutTestController::setAllowFileAccessFromFileURLs):
(LayoutTestController::overridePreference):
(LayoutTestController::setEditingBehavior):
* DumpRenderTree/chromium/TestShell.cpp:
(TestShell::resetWebSettings): Use WebPreferences.
(TestShell::runFileTest): ditto.
(TestShell::createNewWindow): Apply existing WebPreferences to a new WebView.
* DumpRenderTree/chromium/TestShell.h:
(TestShell::preferences):
(TestShell::applyPreferences):
* DumpRenderTree/chromium/WebPreferences.cpp: Added.
* DumpRenderTree/chromium/WebPreferences.h: Added.
2010-09-18 Prasad Tammana <prasadt@chromium.org>
Reviewed by David Levin.
update-webkit --chromium spitting out a spurious error
https://bugs.webkit.org/show_bug.cgi?id=45868
* Scripts/update-webkit-chromium: Use commandExists() function to check for existence of gclient.
2010-09-18 Sam Weinig <sam@webkit.org>
Reviewed by Dan Bernstein.
Improve/unify the PageLoadClient interfaces
https://bugs.webkit.org/show_bug.cgi?id=46043
* MiniBrowser/mac/BrowserWindowController.m:
(didStartProvisionalLoadForFrame):
(didReceiveServerRedirectForProvisionalLoadForFrame):
(didFailProvisionalLoadWithErrorForFrame):
(didCommitLoadForFrame):
(didFinishDocumentLoadForFrame):
(didFinishLoadForFrame):
(didFailLoadWithErrorForFrame):
(didReceiveTitleForFrame):
(didFirstLayoutForFrame):
(didFirstVisuallyNonEmptyLayoutForFrame):
* WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
(WTR::InjectedBundlePage::InjectedBundlePage):
(WTR::InjectedBundlePage::didStartProvisionalLoadForFrame):
(WTR::InjectedBundlePage::didReceiveServerRedirectForProvisionalLoadForFrame):
(WTR::InjectedBundlePage::didFailProvisionalLoadWithErrorForFrame):
(WTR::InjectedBundlePage::didCommitLoadForFrame):
(WTR::InjectedBundlePage::didFinishLoadForFrame):
(WTR::InjectedBundlePage::didFinishDocumentLoadForFrame):
(WTR::InjectedBundlePage::didFailLoadWithErrorForFrame):
(WTR::InjectedBundlePage::didReceiveTitleForFrame):
* WebKitTestRunner/InjectedBundle/InjectedBundlePage.h:
* WebKitTestRunner/TestController.cpp:
(WTR::TestController::didReceiveMessageFromInjectedBundle):
(WTR::TestController::didFinishLoadForFrame):
* WebKitTestRunner/TestController.h:
2010-09-17 Eric Seidel <eric@webkit.org>
Unreviewed. Adding a shell script I use to run the
commit-queue. Now that multiple machines run the cq,
it makes sense to share the script between them.
* EWSTools/start-commit-queue.sh: Added.
2010-09-17 Brian Weinstein <bweinstein@apple.com>
Reviewed by Jon Honeycutt.
The colors on the new build.webkit.org (after upgrading the master to 0.8.1) are
slightly harder to read, so this reverts the previous background colors for various
states, making it easier to read.
* BuildSlaveSupport/build.webkit.org-config/public_html/default.css:
(.success):
(.failure):
(.warnings):
(.exception):
(.start,.running,td.building):
(.offline,td.offline):
2010-09-17 Sheriff Bot <webkit.review.bot@gmail.com>
Unreviewed, rolling out r67692.
http://trac.webkit.org/changeset/67692
https://bugs.webkit.org/show_bug.cgi?id=46011
broke layout test dashboard (Requested by johnny_g on
#webkit).
* Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
* Scripts/webkitpy/layout_tests/layout_package/json_results_generator_unittest.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2010-09-17 Martin Robinson <mrobinson@igalia.com>
Reviewed by Oliver Hunt.
Style bot complains about cairo forward declaration naming
https://bugs.webkit.org/show_bug.cgi?id=45867
Exclude Cairo forward declarations from indentifiers with underscores checks.
* Scripts/webkitpy/style/checkers/cpp.py: Add exclusion for Cairo forward-declarations.
* Scripts/webkitpy/style/checkers/cpp_unittest.py: Add some tests for this.
2010-09-16 Kinuko Yasuda <kinuko@chromium.org>
Reviewed by Ojan Vafai.
Include detailed test modifiers (e.g. FLAKY) in results.json for failing non-layout tests
https://bugs.webkit.org/show_bug.cgi?id=45408
This change also tries to remove duplicated upload_results_json methods
in run_webkit_tests.py and json_results_generator.py.
* Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
* Scripts/webkitpy/layout_tests/layout_package/json_results_generator_unittest.py:
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2010-09-16 Antonio Gomes <tonikitoo@webkit.org>
Reviewed by Andreas Kling.
[Qt] When switching views (WebViewGraphicsBased <--> WebViewTraditional), QWebPage signals and QActions have to be re-set.
Recently r67554 changed the way different views use the WebPage class: it was previously being shared between
different views, but now for each view switch, a new WebPage class is constructed and set. Signals and QAction's
were not being set to the new WebPage though. Patch fix that, by re constructing the toolbar, and then re-hooking
all page specific stuff to the UI.
* QtTestBrowser/launcherwindow.cpp:
(LauncherWindow::initializeView):
* QtTestBrowser/mainwindow.cpp:
(MainWindow::MainWindow):
(MainWindow::buildUI):
(MainWindow::setPage):
* QtTestBrowser/mainwindow.h:
2010-09-12 Antonio Gomes <tonikitoo@webkit.org>
Reviewed by Andreas Kling.
Make all accesses to m_item (GraphicsWebView) done by WebViewGraphicsBased class happen though graphicsWebView getter method.
This patch is mostly a preparation for another commit, where the ownership of the m_item/GraphicsWebView
object will move to the qgraphicsscene instead of the qgraphicsview.
* QtTestBrowser/webview.cpp:
(WebViewGraphicsBased::setPage):
(WebViewGraphicsBased::setResizesToContents):
(WebViewGraphicsBased::resizeEvent):
(WebViewGraphicsBased::animatedFlip):
* QtTestBrowser/webview.h:
(WebViewGraphicsBased::setItemCacheMode):
(WebViewGraphicsBased::itemCacheMode):
2010-09-12 Antonio Gomes <tonikitoo@webkit.org>
Reviewed by Andreas Kling.
Moved setYRotation method body out of the class (WebViewGraphicsBased) definition.
We usually separate methods whose body are larger than a few lines from the class definition.
* QtTestBrowser/webview.h:
(WebViewGraphicsBased::yRotation):
(WebViewGraphicsBased::setYRotation):
2010-09-12 Antonio Gomes <tonikitoo@webkit.org>
Reviewed by Andreas Kling.
Make initializeView method of LauncherWindow private.
No one else needs to be able to call it but LauncherWindow.
Also moving the declation of isGraphicsBased method down in the same file, in order
to better group related methods.
* QtTestBrowser/launcherwindow.h:
2010-09-12 Antonio Gomes <tonikitoo@webkit.org>
Reviewed by Andreas Kling.
Make the getter 'page' method of MainWindow const.
* QtTestBrowser/mainwindow.cpp:
(MainWindow::page):
* QtTestBrowser/mainwindow.h:
2010-09-12 Antonio Gomes <tonikitoo@webkit.org>
Reviewed by Andreas Kling.
Remove useless parameter from MainWindow class' constructor.
The parameter defaults to an empty URL and is not being used by any caller. Apart from that it is also
not referred in the constructor body and does not make much sense.
* QtTestBrowser/mainwindow.cpp:
(MainWindow::MainWindow):
* QtTestBrowser/mainwindow.h:
2010-09-15 Tony Chang <tony@chromium.org>
Reviewed by Kent Tamura.
[Chromium] Needs DRT queueLoadHTMLString and setDeferMainResourceLoad-implementations
https://bugs.webkit.org/show_bug.cgi?id=42151
* DumpRenderTree/chromium/LayoutTestController.cpp:
(LayoutTestController::LayoutTestController):
(WorkItemLoadHTMLString::WorkItemLoadHTMLString):
(WorkItemLoadHTMLString::run):
(LayoutTestController::queueLoadHTMLString):
(LayoutTestController::reset):
(LayoutTestController::setDeferMainResourceDataLoad):
* DumpRenderTree/chromium/LayoutTestController.h:
(LayoutTestController::deferMainResourceDataLoad):
* DumpRenderTree/chromium/WebViewHost.cpp:
(WebViewHost::didCreateDataSource):
2010-09-16 Adrienne Walker <enne@google.com>
Reviewed by Kenneth Russell.
Add script to synchronize WebKit and Khronos WebGL tests
https://bugs.webkit.org/show_bug.cgi?id=42336
* Scripts/update-webgl-conformance-tests: Added.
* Scripts/webkitpy/layout_tests/update_webgl_conformance_tests.py: Added.
* Scripts/webkitpy/layout_tests/update_webgl_conformance_tests_unittest.py: Added.
2010-09-16 Sam Weinig <sam@webkit.org>
Reviewed by Anders Carlsson.
Add support for sending synchronous messages from the InjectedBundle to the WKContext
<rdar://problem/8365320>
https://bugs.webkit.org/show_bug.cgi?id=44785
* MiniBrowser/mac/AppDelegate.m:
(didReceiveSynchronousMessageFromInjectedBundle):
(-[BrowserAppDelegate init]):
* MiniBrowser/mac/WebBundle/WebBundleMain.m:
(didCreatePage):
* WebKitTestRunner/TestController.cpp:
(WTR::TestController::initialize):
2010-09-16 Leonid Ebril <leonid.ebril@nokia.com>
Reviewed by Andreas Kling.
[Qt] QtTestBrowser: Disable creation of a new window for screenshot on Symbian platform.
https://bugs.webkit.org/show_bug.cgi?id=45885
Avoid creation of an additional window for screenshot to prevent overlapping with original window.
* QtTestBrowser/launcherwindow.cpp:
(LauncherWindow::screenshot):
2010-09-16 Anders Carlsson <andersca@apple.com>
Fix clang++ build.
* MiniBrowser/mac/BrowserWindowController.m:
(-[BrowserWindowController zoomIn:]):
(-[BrowserWindowController zoomOut:]):
(-[BrowserWindowController resetZoom:]):
Remove trailing semicolons.
2010-09-16 Sheriff Bot <webkit.review.bot@gmail.com>
Unreviewed, rolling out r67628.
http://trac.webkit.org/changeset/67628
https://bugs.webkit.org/show_bug.cgi?id=45904
broke the build (Requested by eric_carlson on #webkit).
* DumpRenderTree/gtk/EventSender.cpp:
(contextClickCallback):
2010-09-16 Balazs Kelemen <kbalazs@webkit.org>
Unreviewed.
Adding myself as a comitter.
* Scripts/webkitpy/common/config/committers.py:
2010-09-15 Philippe Normand <pnormand@igalia.com>
Reviewed by Eric Carlson.
[GTK] enhanced context menu for media elements
https://bugs.webkit.org/show_bug.cgi?id=45021
EventSender::contextClick() now returns an array of js
objects. Each object has a title property and a click() method.
* DumpRenderTree/gtk/EventSender.cpp:
(getMenuItemTitleCallback):
(setMenuItemTitleCallback):
(menuItemClickCallback):
(getMenuItemClass):
(contextClickCallback):
2010-09-14 Philippe Normand <pnormand@igalia.com>
Reviewed by Eric Carlson and Martin Robinson.
[GTK] eventSender.contextClick() should return the contents of the context menu
https://bugs.webkit.org/show_bug.cgi?id=39102
Make the eventSender use the new private WebKitGtk+ API to
retrieve the context-menu item titles and store them in an array.
* DumpRenderTree/gtk/EventSender.cpp:
(contextClickCallback):
2010-09-16 Eric Uhrhane <ericu@chromium.org>
Reviewed by Jian Li.
Unify FILE_SYSTEM and FILE_WRITER enables under the name FILE_SYSTEM.
https://bugs.webkit.org/show_bug.cgi?id=45798
* Scripts/build-webkit:
2010-09-16 Robert Hogan <robert@webkit.org>
Reviewed by Antonio Gomes.
[Qt] Support globalhistory tests
https://bugs.webkit.org/show_bug.cgi?id=45774
* DumpRenderTree/qt/DumpRenderTreeQt.cpp:
(WebCore::isGlobalHistoryTest):
(WebCore::DumpRenderTree::open):
* DumpRenderTree/qt/LayoutTestControllerQt.cpp:
(LayoutTestController::reset):
(LayoutTestController::dumpHistoryCallbacks):
(LayoutTestController::removeAllVisitedLinks):
* DumpRenderTree/qt/LayoutTestControllerQt.h:
2010-09-10 Tony Chang <tony@chromium.org>
Reviewed by Eric Seidel.
deduplicate-tests should be runnable from any WebKit directory
https://bugs.webkit.org/show_bug.cgi?id=44709
* Scripts/webkitpy/layout_tests/deduplicate_tests.py:
* Scripts/webkitpy/layout_tests/deduplicate_tests_unittest.py:
2010-09-15 Tony Chang <tony@chromium.org>
Reviewed by Kent Tamura.
[chromium] sort testing methods in DRT
https://bugs.webkit.org/show_bug.cgi?id=45850
* DumpRenderTree/chromium/EventSender.cpp:
(EventSender::EventSender):
* DumpRenderTree/chromium/LayoutTestController.cpp:
(LayoutTestController::LayoutTestController):
* DumpRenderTree/chromium/TextInputController.cpp:
(TextInputController::TextInputController):
2010-09-15 Mihai Parparita <mihaip@chromium.org>
Reviewed by Adam Barth.
CQ status shows items out of order
https://bugs.webkit.org/show_bug.cgi?id=45846
Only group consecutive status items with the same patch ID into the same
group, so that overall item ordering is preserved.
Also don't indent status items that only have a single item in their
group.
* QueueStatusServer/handlers/queuestatus.py:
* QueueStatusServer/templates/includes/singlequeuestatus.html: Added.
* QueueStatusServer/templates/queuestatus.html:
2010-09-15 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
Add queue start/stop messages
https://bugs.webkit.org/show_bug.cgi?id=45853
I ended up needing to clean up a bunch of our unit testing
in order to test this new code path nicely.
There are also a few PEP8 changes needed to pass check-webkit-style.
* Scripts/webkitpy/tool/bot/queueengine.py:
* Scripts/webkitpy/tool/bot/queueengine_unittest.py:
* Scripts/webkitpy/tool/commands/earlywarningsystem_unittest.py:
* Scripts/webkitpy/tool/commands/queues.py:
* Scripts/webkitpy/tool/commands/queues_unittest.py:
* Scripts/webkitpy/tool/commands/queuestest.py:
* Scripts/webkitpy/tool/commands/sheriffbot_unittest.py:
2010-09-15 Simon Fraser <simon.fraser@apple.com>
https://bugs.webkit.org/show_bug.cgi?id=45849
WKURLCreateWithCFURL crashes with null url
Fix MiniBrowser crash when the url is not parseable.
* MiniBrowser/mac/BrowserWindowController.m:
(-[BrowserWindowController fetch:]):
2010-09-15 Patrick Gansterer <paroga@paroga.com>
Unreviewed.
Adding myself to the committers list.
* Scripts/webkitpy/common/config/committers.py:
2010-09-14 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
Make it possible to run more than one commit-queue instance
https://bugs.webkit.org/show_bug.cgi?id=45786
Mostly we need to make sure the two (or more) instances get
different patches to work on. To do this, I re-worked
the code responsible for getting the next work item to
round trip through the status server. The status server only
vends patches from the work items list, only if those patches
have not had status reported for them in the last hour.
This is another step towards making all queues go through the
status server, thus making it possible to run more than one
instance of various EWS bots (as requested in bug 44292).
The webkitpy changes are already covered by existing unit tests.
The QueueStatusSever sadly has no testing infrastructure yet. :(
* QueueStatusServer/handlers/nextpatch.py: Added.
* QueueStatusServer/index.yaml:
* QueueStatusServer/main.py:
* Scripts/webkitpy/tool/commands/queues.py:
2010-09-15 Mihai Parparita <mihaip@chromium.org>
Reviewed by Tony Chang.
NRWT fails with UnicodeDecodeError on editing/selection/mixed-editability-10.html
https://bugs.webkit.org/show_bug.cgi?id=45791
Force filenames to be raw bytes before running difflib.unified_diff.
* Scripts/webkitpy/layout_tests/port/base.py:
* Scripts/webkitpy/layout_tests/port/base_unittest.py:
2010-09-15 Simon Hausmann <simon.hausmann@nokia.com>
Reviewed by Tor Arne Vestbø.
[Qt] QtTestBrowser crashes when enabling QGraphicsView mode after first loading page without it enabled
https://bugs.webkit.org/show_bug.cgi?id=35247
The main issue when changing the views is that the plugins and any other components that
depend on view specific attributes such as the native window id are not teared down.
Even if we had a tear-down procedure, we'd have to re-initialize the plugin after
switching to the new view (QGraphicsWebView). This is a rather uncommon situation, so
we decided to work around this in QtTestBrowser by making toggling between QWebView
and QGraphicsWebView also re-create the page.
* QtTestBrowser/launcherwindow.cpp:
(LauncherWindow::initializeView):
* QtTestBrowser/mainwindow.h:
2010-09-14 Mihai Parparita <mihaip@chromium.org>
Reviewed by Adam Barth.
Send webkit accessibility notifications to Chromium
https://bugs.webkit.org/show_bug.cgi?id=45156
Move printf inside shouldDumpAccessibilityNotifications check.
* DumpRenderTree/chromium/WebViewHost.cpp:
(WebViewHost::postAccessibilityNotification):
2010-09-14 Tony Chang <tony@chromium.org>
Reviewed by Kent Tamura.
[chromium] fix http/tests/security/local-user-CSS-from-remote.html
https://bugs.webkit.org/show_bug.cgi?id=45788
* DumpRenderTree/chromium/LayoutTestController.cpp:
(LayoutTestController::setUserStyleSheetLocation): use a base64 url like in Chrome
2010-09-14 Tony Chang <tony@chromium.org>
Reviewed by Ojan Vafai.
[chromium] Remove WebKit::areLayoutTestImagesOpaque since it's no longer needed
https://bugs.webkit.org/show_bug.cgi?id=45768
* DumpRenderTree/chromium/TestShell.cpp:
(TestShell::dumpImage): new baselines are checked in and Linux will match Windows going forward.
2010-09-14 Kevin Ollivier <kevino@theolliviers.com>
[wx] Build fix, add newly added directories to the directory list.
* wx/build/settings.py:
2010-09-14 Sam Weinig <sam@webkit.org>
Reviewed by Anders Carlsson.
Remove WKBundleNodeRef, replacing uses with WKBundleNodeHandleRef.
https://bugs.webkit.org/show_bug.cgi?id=45785
* WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
(WTR::propertyValue):
(WTR::numericWindowPropertyValue):
(WTR::dumpPath):
(WTR::toStr):
(WTR::operator<<):
(WTR::InjectedBundlePage::InjectedBundlePage):
(WTR::InjectedBundlePage::shouldInsertNode):
(WTR::InjectedBundlePage::shouldBeginEditing):
(WTR::InjectedBundlePage::shouldEndEditing):
(WTR::InjectedBundlePage::shouldInsertText):
(WTR::InjectedBundlePage::shouldDeleteRange):
(WTR::InjectedBundlePage::shouldChangeSelectedRange):
(WTR::InjectedBundlePage::shouldApplyStyle):
* WebKitTestRunner/InjectedBundle/InjectedBundlePage.h:
2010-09-14 Adam Barth <abarth@webkit.org>
Fix two typos in commit-queue. The first is harmless. The second
causes the queue to reject patches when the build is broken. :(
* Scripts/webkitpy/tool/commands/queues.py:
* Scripts/webkitpy/tool/commands/queues_unittest.py:
2010-09-14 Chris Guillory <chris.guillory@google.com>
Reviewed by Chris Fleizach.
Send all accessibility notifications to Chromium.
https://bugs.webkit.org/show_bug.cgi?id=45156
* DumpRenderTree/chromium/WebViewHost.cpp:
(WebViewHost::postAccessibilityNotification):
* DumpRenderTree/chromium/WebViewHost.h:
2010-09-14 Adam Barth <abarth@webkit.org>
Reviewed by Eric Seidel.
commit-queue is slow during the day
https://bugs.webkit.org/show_bug.cgi?id=45780
Thanks to the new logging, we've noticed that checkout-is-out-of-date
errors in the first pass of landing don't retry the land. Instead,
they're treated as failures and cause the commit-queue to do two more
builds before really trying to land the patch. Worse, in the second
build, we can get bitten by a flaky test.
This patch takes a slightly different approach to the commit-queue's
main control logic. We now use a separate subprocess for building and
testing and for landing. This means we should very rarely see the
checkout-is-out-of-date error, and when we do see it, we should retry
more quickly. If my understanding is correct, this should be a big
speed win for the commit-queue.
* Scripts/webkitpy/tool/commands/download.py:
* Scripts/webkitpy/tool/commands/queues.py:
* Scripts/webkitpy/tool/commands/queues_unittest.py:
2010-09-14 Tony Chang <tony@chromium.org>
Reviewed by Dimitri Glazkov.
[chromium] implement layoutTestController.setDomainRelaxationForbiddenForURLScheme
https://bugs.webkit.org/show_bug.cgi?id=45762
* DumpRenderTree/chromium/LayoutTestController.cpp:
(LayoutTestController::LayoutTestController):
(LayoutTestController::setDomainRelaxationForbiddenForURLScheme):
* DumpRenderTree/chromium/LayoutTestController.h:
2010-09-14 Csaba Osztrogonác <ossy@webkit.org>
Adding myself as a reviewer.
* Scripts/webkitpy/common/config/committers.py:
2010-09-14 Zoltan Horvath <zoltan@webkit.org>
Reviewed by Kenneth Rohde Christiansen.
[Qt] Build-webkit shows a warning when WebKitBuild directory is not exist
https://bugs.webkit.org/show_bug.cgi?id=45736
Build-webkit has showed a warning when WebKitBuild didn't exist, add
an additional condition to avoid it.
* Scripts/build-webkit:
2010-09-13 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
commit-queue should log when it's retrying due to checkout out of date error
https://bugs.webkit.org/show_bug.cgi?id=45725
This will allow us to better track how often we see these errors
and how long it takes the queue to recover from them.
In order to pass the tests, I had to correct an error
from a previous change. The UnitTestPort should not have
ever been using the base port method as that will read
a file from disk. If the unit tests haven't passed the contents
it should raise an IOError as though the file doesn't exist.
* Scripts/webkitpy/layout_tests/port/base_unittest.py:
* Scripts/webkitpy/tool/commands/queues.py:
* Scripts/webkitpy/tool/commands/queues_unittest.py:
2010-09-13 Mihai Parparita <mihaip@chromium.org>
Reviewed by Adam Barth.
Upload results when cq patch landing fails
https://bugs.webkit.org/show_bug.cgi?id=45683
Include script errors when CommitQueue._land fails.
* Scripts/webkitpy/tool/commands/queues.py:
2010-09-13 Chris Fleizach <cfleizach@apple.com>
Reviewed by David Kilzer.
AX: accessibilityIsIgnored is returning nil when return value expects a BOOL
https://bugs.webkit.org/show_bug.cgi?id=45548
* DumpRenderTree/AccessibilityUIElement.cpp:
(isIgnoredCallback):
(AccessibilityUIElement::getJSClass):
* DumpRenderTree/AccessibilityUIElement.h:
* DumpRenderTree/gtk/AccessibilityUIElementGtk.cpp:
(AccessibilityUIElement::isIgnored):
* DumpRenderTree/mac/AccessibilityUIElementMac.mm:
(AccessibilityUIElement::isIgnored):
* DumpRenderTree/win/AccessibilityUIElementWin.cpp:
(AccessibilityUIElement::isIgnored):
2010-09-13 Csaba Osztrogonác <ossy@webkit.org>
Reviewed by Andreas Kling.
jscPath() is incorrect in Windows' cmd.exe shell
https://bugs.webkit.org/show_bug.cgi?id=45651
* Scripts/webkitdirs.pm: Add ".exe" suffix to $jscName on Windows.
2010-09-13 Chris Fleizach <cfleizach@apple.com>
Reviewed by Darin Adler.
REGRESSION (r64816-r64889): Crash in WebCore::AccessibilityRenderObject
https://bugs.webkit.org/show_bug.cgi?id=43807
* DumpRenderTree/AccessibilityUIElement.cpp:
(textMarkerForPointCallback):
(AccessibilityUIElement::textMarkerForPoint):
(AccessibilityUIElement::getJSClass):
* DumpRenderTree/AccessibilityUIElement.h:
* DumpRenderTree/mac/AccessibilityUIElementMac.mm:
(AccessibilityUIElement::textMarkerForPoint):
2010-09-13 Andreas Kling <andreas.kling@nokia.com>
Reviewed by Kenneth Rohde Christiansen.
[Qt] build-webkit: Add --v8 option to build against V8
Note that V8 is picked up from QtScript, so this currently requires building
against the experimental qt-script-v8 branch.
* Scripts/build-webkit:
2010-09-13 Hans Wennborg <hans@chromium.org>
Reviewed by Jeremy Orlow.
Add WebDeviceOrientationClientMock::create() and use it from DRT.
https://bugs.webkit.org/show_bug.cgi?id=45650
Do not use the WebDeviceOrientationClientMock constructor directly,
but use the create() member function.
This should fix the linker errors for the multi-DLL Chromium DRT build.
* DumpRenderTree/chromium/LayoutTestController.cpp:
(LayoutTestController::setMockDeviceOrientation):
(LayoutTestController::deviceOrientationClient):
2010-08-27 Kenneth Rohde Christiansen <kenneth.christiansen@openbossa.org>
Reviewed by Antti Koivisto.
Add support for testing the viewport meta tag algorithm,
based on the following draft spec:
http://people.opera.com/rune/TR/ED-css-viewport-20100806/
Add common handling of viewport meta tag based on new Opera spec
https://bugs.webkit.org/show_bug.cgi?id=44201
* DumpRenderTree/qt/LayoutTestControllerQt.cpp:
(LayoutTestController::dumpConfigurationForViewport):
* DumpRenderTree/qt/LayoutTestControllerQt.h:
2010-09-12 Kent Tamura <tkent@chromium.org>
Reviewed by Tony Chang.
[DRT/Chromium] Remove dependency to base/string16.h and gfx/codec/png_codec.h
https://bugs.webkit.org/show_bug.cgi?id=45517
Use webkit_support_gfx.h for PNG encoding/decoding instead of png_codec.h
* DumpRenderTree/chromium/ImageDiff.cpp:
(Image::craeteFromStdin):
(Image::createFromFilename):
(diffImages):
* DumpRenderTree/chromium/TestShell.cpp:
(TestShell::dumpImage):
2010-09-10 MORITA Hajime <morrita@google.com>
Reviewed by Tony Chang.
[Chromium] Implement textInputController.hasSpellingMarker() for Chromium
https://bugs.webkit.org/show_bug.cgi?id=45441
Added TextInputController::hasSpellingMarker().
* DumpRenderTree/chromium/TextInputController.cpp:
(TextInputController::TextInputController):
(TextInputController::hasSpellingMarker):
* DumpRenderTree/chromium/TextInputController.h:
2010-09-12 Dirk Pranke <dpranke@chromium.org>
Unreviewed, build fix.
My last change accidentally contained a local change to the
rebaselinig script; webkit-patch land picked it up without telling
me and committed it :(
* Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
2010-09-11 Maciej Stachowiak <mjs@apple.com>
Reviewed by Anders Carlsson.
Attempt to fix Windows MiniBrowser build (untested)
https://bugs.webkit.org/show_bug.cgi?id=45609
* MiniBrowser/win/BrowserView.cpp:
(runJavaScriptConfirm):
(runJavaScriptPrompt):
2010-09-11 Sam Weinig <sam@webkit.org>
Reviewed by Maciej Stachowiak.
Fix zoom related tests when testing WebKit2.
* WebKitTestRunner/InjectedBundle/EventSendingController.cpp:
(WTR::EventSendingController::zoomPageIn): Zoom in and zoom out are not the same.
2010-09-11 Sam Weinig <sam@webkit.org>
Reviewed by Dan Bernstein.
Implement WebKit2 callback equivalent to - [WebUIDelegate webView:setStatusText:]
<rdar://problem/8359252>
https://bugs.webkit.org/show_bug.cgi?id=45605
* MiniBrowser/mac/BrowserWindowController.m:
(setStatusText):
(contentsSizeChanged):
(-[BrowserWindowController awakeFromNib]):
* MiniBrowser/win/BrowserView.cpp:
(runJavaScriptConfirm):
(runJavaScriptPrompt):
(setStatusText):
(contentsSizeChanged):
(BrowserView::create):
* WebKitTestRunner/TestController.cpp:
(WTR::createOtherPage):
(WTR::TestController::initialize):
2010-09-11 Sam Weinig <sam@webkit.org>
Reviewed by Dan Bernstein.
Add callback mechanism for the getting the source of a frame
<rdar://problem/8364681>
https://bugs.webkit.org/show_bug.cgi?id=45604
Add ability to dump the main frame's source to the console.
* MiniBrowser/mac/BrowserWindowController.h:
* MiniBrowser/mac/BrowserWindowController.m:
(-[BrowserWindowController dumpSourceToConsole:]):
* MiniBrowser/mac/MainMenu.xib:
2010-09-11 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
Remove unneeded Empty Queue messages
https://bugs.webkit.org/show_bug.cgi?id=45602
We already have a /gc job to delete these.
recentstatus.py is already smart enough to use the
most recent of the workitems last update or the most recent status,
so removing these should not change the _last_status_date() return
value more than a few milliseconds.
These messages just spam the recent status log.
* Scripts/webkitpy/tool/commands/queues.py:
2010-09-11 Mihai Parparita <mihaip@chromium.org>
Reviewed by Adam Barth.
Fix Python exception when generating synthetic patch IDs
https://bugs.webkit.org/show_bug.cgi?id=45592
Can't concatenate strings and numbers directly.
* QueueStatusServer/handlers/queuestatus.py:
2010-09-11 Mihai Parparita <mihaip@chromium.org>
Reviewed by Adam Barth.
Group statuses in queue status by patch ID
https://bugs.webkit.org/show_bug.cgi?id=45588
Group statuses by patch ID so that the status page is easier
to scan.
* QueueStatusServer/handlers/queuestatus.py:
* QueueStatusServer/stylesheets/dashboard.css:
(.status-group):
(.status-bug):
(.status-group ul):
(.status-group ul li):
(.status-group ul li:hover):
(.status-cell):
(.status-cell:hover):
(.status-cell.pass):
(.status-cell.fail):
(.status-cell.pending):
(.status-cell.error):
* QueueStatusServer/templates/dashboard.html:
* QueueStatusServer/templates/queuestatus.html:
2010-09-11 Mihai Parparita <mihaip@chromium.org>
Reviewed by Tony Chang.
Implement layoutTestController.dumpResourceResponseMIMETypes in Chromium DRT
https://bugs.webkit.org/show_bug.cgi?id=45479
Implement layoutTestController.dumpResourceResponseMIMETypes (modelled after
implementation in ResourceLoadDelegate in the Mac port).
* DumpRenderTree/chromium/LayoutTestController.cpp:
(LayoutTestController::LayoutTestController):
(LayoutTestController::dumpResourceResponseMIMETypes):
(LayoutTestController::reset):
* DumpRenderTree/chromium/LayoutTestController.h:
(LayoutTestController::setShouldDumpResourceResponseMIMETypes):
(LayoutTestController::shouldDumpResourceResponseMIMETypes):
* DumpRenderTree/chromium/TestShell.h:
(TestShell::shouldDumpResourceResponseMIMETypes):
* DumpRenderTree/chromium/WebViewHost.cpp:
(WebViewHost::didReceiveResponse):
2010-09-11 Eric Seidel <eric@webkit.org>
Reviewed by Adam Barth.
queues.webkit.org should expose /remote_api for data upload/download
https://bugs.webkit.org/show_bug.cgi?id=45559
* QueueStatusServer/app.yaml:
- /remote_api is a standard app-engine service which allows use of
the bulk uploader for data upload/download. Exposing this
(for admin access only) allows us to use
2010-09-10 Sam Weinig <sam@webkit.org>
Reviewed by Dan Bernstein.
Add zoom support to WebKit2 API
<rdar://problem/7660657>
https://bugs.webkit.org/show_bug.cgi?id=45585
- Add zoom options to MiniBrowser that mimic Safari's options using
the new WebKit2 zoom APIs.
- Switch uses of float for the zoom APIs to use double.
* MiniBrowser/mac/BrowserWindowController.h:
* MiniBrowser/mac/BrowserWindowController.m:
(-[BrowserWindowController initWithPageNamespace:]):
(-[BrowserWindowController validateMenuItem:]):
(-[BrowserWindowController validateUserInterfaceItem:]):
(-[BrowserWindowController currentZoomFactor]):
(-[BrowserWindowController setCurrentZoomFactor:]):
(-[BrowserWindowController canZoomIn]):
(-[BrowserWindowController zoomIn:]):
(-[BrowserWindowController canZoomOut]):
(-[BrowserWindowController zoomOut:]):
(-[BrowserWindowController canResetZoom]):
(-[BrowserWindowController resetZoom:]):
(-[BrowserWindowController toggleZoomMode:]):
* MiniBrowser/mac/MainMenu.xib:
* WebKitTestRunner/InjectedBundle/EventSendingController.cpp:
(WTR::EventSendingController::textZoomIn):
(WTR::EventSendingController::textZoomOut):
(WTR::EventSendingController::zoomPageIn):
(WTR::EventSendingController::zoomPageOut):
2010-09-10 Robert Hogan <robert@webkit.org>
Reviewed by Andreas Kling.
[Qt] Fix crash in DRT in standalone mode
https://bugs.webkit.org/show_bug.cgi?id=45454
* DumpRenderTree/qt/DumpRenderTreeQt.cpp:
(WebCore::DumpRenderTree::loadNextTestInStandAloneMode):
2010-09-10 Sam Weinig <sam@webkit.org>
Reviewed by Darin Adler.
Remove unnecessary constraint in WebCore of choosing either text zoom or full page zoom.
Precursor to <rdar://problem/7660657>
https://bugs.webkit.org/show_bug.cgi?id=45522
Update for changes to the Bundle API to allow separate control of page and text zoom
levels.
* WebKitTestRunner/InjectedBundle/EventSendingController.cpp:
(WTR::EventSendingController::textZoomIn):
(WTR::EventSendingController::textZoomOut):
(WTR::EventSendingController::zoomPageIn):
(WTR::EventSendingController::zoomPageOut):
* WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
(WTR::InjectedBundlePage::reset):
2010-09-10 Yael Aharon <yael.aharon@nokia.com>
Reviewed by Andreas Kling.
[Qt] Support click event for notifications
https://bugs.webkit.org/show_bug.cgi?id=44836
Add support for simulateDesktopNotificationClick by calling
back to DumpRenderTreeSupportQt.cpp.
* DumpRenderTree/qt/LayoutTestControllerQt.cpp:
(LayoutTestController::simulateDesktopNotificationClick):
* DumpRenderTree/qt/LayoutTestControllerQt.h:
2010-09-10 Mihai Parparita <mihaip@chromium.org>
Reviewed by Eric Seidel.
Update queue status when patches fail their first land attempt
https://bugs.webkit.org/show_bug.cgi?id=45583
Add a few more _update_status calls to make it more obvious if a patch is
being retried because tests failed.
* Scripts/webkitpy/tool/commands/queues.py:
2010-09-10 Balazs Kelemen <kb@inf.u-szeged.hu>
Reviewed by Andreas Kling.
[Qt] MiniBrowser crashes with multiply windows when closing one of them
https://bugs.webkit.org/show_bug.cgi?id=45536
* MiniBrowser/qt/BrowserView.cpp:
(BrowserView::BrowserView): Reference the context instead of adopting it
because it can be shared across different views.
2010-09-10 Tony Chang <tony@chromium.org>
Unreviewed, rolling out r67241.
http://trac.webkit.org/changeset/67241
https://bugs.webkit.org/show_bug.cgi?id=44709
Accidentally committed.
* Scripts/webkitpy/layout_tests/deduplicate_tests.py:
* Scripts/webkitpy/layout_tests/deduplicate_tests_unittest.py:
2010-09-10 Tony Chang <tony@chromium.org>
Reviewed by Ojan Vafai.
deduplicate-tests should be runnable from any WebKit directory
https://bugs.webkit.org/show_bug.cgi?id=44709
* Scripts/webkitpy/layout_tests/deduplicate_tests.py:
* Scripts/webkitpy/layout_tests/deduplicate_tests_unittest.py:
2010-09-10 Tony Chang <tony@chromium.org>
Unreviewed, rolling out r67216.
http://trac.webkit.org/changeset/67216
https://bugs.webkit.org/show_bug.cgi?id=44709
Broke
* Scripts/webkitpy/layout_tests/deduplicate_tests.py:
* Scripts/webkitpy/layout_tests/deduplicate_tests_unittest.py:
2010-09-10 Tony Chang <tony@chromium.org>
Reviewed by Ojan Vafai.
deduplicate-tests should be runnable from any WebKit directory
https://bugs.webkit.org/show_bug.cgi?id=44709