⚠ Archived content — this site is no longer maintained.   Current WebKit documentation is at docs.webkit.org.

Timeline



Dec 26, 2013:

6:55 PM Changeset in webkit [161096] by weinig@apple.com
  • 10 edits in trunk/Source/WebCore

Convert some of WebCore/dom over to range-for loops
​https://bugs.webkit.org/show_bug.cgi?id=126250

Reviewed by Andreas Kling.

  • dom/ContainerNode.cpp:
  • dom/Document.cpp:
  • dom/MutationObserver.cpp:
  • dom/MutationObserverInterestGroup.cpp:
  • dom/MutationObserverRegistration.cpp:
  • dom/Node.cpp:
  • dom/NodeRareData.h:
  • dom/ScriptExecutionContext.cpp:
  • dom/WebKitNamedFlow.cpp:
4:21 PM Changeset in webkit [161095] by Martin Robinson
  • 4 edits in trunk/Source

Small build fix for the GTK+ CMake port

Source/WebKit/gtk:

  • tests/CMakeLists.txt: Output the WebKit1 GObject API unit tests to the proper directory

and fix the path to the generated API source.

Source/WebKit2:

  • UIProcess/API/gtk/tests/CMakeLists.txt: Make the way the output directory is specified similar

to the WebKit1 version of these tests, so that it is more difficult to inadvertently specify the
wrong directory.

3:01 PM Changeset in webkit [161094] by thiago.lacerda@openbossa.org
  • 2 edits in trunk/Source/WebCore

[Nix] Cleanup Source/WebCore/PlatformNix.cmake
​https://bugs.webkit.org/show_bug.cgi?id=126226

Reviewed by Csaba Osztrogonác.

No new tests needed.

  • PlatformNix.cmake:
1:24 PM Changeset in webkit [161093] by Martin Robinson
  • 3 edits in trunk

[GTK] [CMake] Add a production build type
​https://bugs.webkit.org/show_bug.cgi?id=126179

Reviewed by Daniel Bates.

  • Source/PlatformGTK.cmake: When the API tests are enabled compile the WebKit GObject

API unit tests. The inclusion of these build files was inadvertently removed in an earlier
patch.

  • Source/cmake/OptionsGTK.cmake: When in production mode turn off all tools and tests and

use the autotools linker script.

11:20 AM Changeset in webkit [161092] by ap@apple.com
  • 5 edits in trunk/Tools

Some links at build.webkit.org/dashboard don't show status messages, and aren't keyboard accessible
​https://bugs.webkit.org/show_bug.cgi?id=126234

Reviewed by Timothy Hatcher.

Changed elements that are semantically links from <div> to <a>.

  • BuildSlaveSupport/build.webkit.org-config/public_html/dashboard/Scripts/StatusLineView.js:

(StatusLineView): Create <a> or <div> elements conditionally on whether we have a URL.
Don't try simulate a link with CSS/JS, it's not needed any more.
(StatusLineView.prototype._clicked): Removed.

  • BuildSlaveSupport/build.webkit.org-config/public_html/dashboard/Styles/BuildbotQueueView.css:

Removed text-decoration rules, default ones in Main.css now work for queue elements.

  • BuildSlaveSupport/build.webkit.org-config/public_html/dashboard/Styles/Main.css:

Simplified cursor rules.
Added a universal rule for links to only show underline on hover, as this is what
we want almost everywhere.

  • BuildSlaveSupport/build.webkit.org-config/public_html/dashboard/Styles/StatusLineView.css:

(.status-line .bubble): Added text-decoration: none, as this is the one place where
we don't want it even on hover.
(.status-line .label): Force display: block for consistent layout between <a>
and <div>.

10:19 AM Changeset in webkit [161091] by Joseph Pecoraro
  • 2 edits in trunk/Source/WebCore

Unreviewed attempt at Windows build fix.

I think Window's "AllInOne.cpp" is causing a using namespace JSC
to cause naming conflicts between Inspector::TypeBuilder::Debugger::types
and JSC::types. So removing the ambiguity.

  • inspector/InjectedScript.cpp:

(WebCore::InjectedScript::getProperties):
(WebCore::InjectedScript::wrapCallFrames):

10:01 AM Changeset in webkit [161090] by ap@apple.com
  • 10 edits
    1 add in trunk/Tools

Please clarify "pending" reporting at build.webkit.org/dashboard
​https://bugs.webkit.org/show_bug.cgi?id=122191

Reviewed by Timothy Hatcher.

Reporting the count of pending runs didn't make a lot of sense - first, runs are
coalesced and sometimes even out of order, and second, buildbot's notion of pending
run was confusingly different from dashboard's.

Let's display how many SVN revisions are pending. This can be somewhat misleading
too, because some revisions (like those for other platforms) don't trigger builds,
but it's better than what we had.

This patch also lays the groundwork for displaying detailed information about
pending revisions.

  • BuildSlaveSupport/build.webkit.org-config/public_html/dashboard/Scripts/BuildbotBuilderQueueView.js:

(BuildbotBuilderQueueView.prototype.update.appendBuilderQueueStatus): Instead of
building pending status line directly, call newly added base class method.

  • BuildSlaveSupport/build.webkit.org-config/public_html/dashboard/Scripts/BuildbotTesterQueueView.js:

(BuildbotTesterQueueView.prototype.update.appendBuilderQueueStatus): Ditto.

  • BuildSlaveSupport/build.webkit.org-config/public_html/dashboard/Scripts/BuildbotQueueView.js:

(BuildbotQueueView): Listen for events when new SVN revisions are landed, and update the view.
(BuildbotQueueView.prototype._appendPendingRevisionCount): Add a line for pending
SVN revisions. It uses a new style, StatusLineView.Status.NoBubble, because this
information is secondary, and doesn't need as much attention.
(BuildbotQueueView.prototype.revisionLinksForIteration): Build revisionURL through
Trac, not through Buildbot.
(BuildbotQueueView.prototype._newCommitsRecorded): Schedule an update, just like when
iterations are updated.

  • BuildSlaveSupport/build.webkit.org-config/public_html/dashboard/Scripts/Initialization.js:

Initialize a global webkitTrac object.

  • BuildSlaveSupport/build.webkit.org-config/public_html/dashboard/Scripts/StatusLineView.js:
  • BuildSlaveSupport/build.webkit.org-config/public_html/dashboard/Styles/StatusLineView.css:

Added a new style for messages without a bubble.

  • BuildSlaveSupport/build.webkit.org-config/public_html/dashboard/Scripts/Trac.js:

Added a model class for Trac, which keeps track of SVN timeline, and notifies
listeners of changes. It uses an RSS interface to Trac, because there is no JSON one.

  • BuildSlaveSupport/build.webkit.org-config/public_html/dashboard/Scripts/Utilities.js:

(loadXML): Added a function to load XML asynchronouly, just like existing JSON.load.

  • BuildSlaveSupport/build.webkit.org-config/public_html/dashboard/Scripts/WebKitBuildbot.js:

Removed tracRevisionURL() function. Now that we have a Trac object, it just makes
more sense to build trac URLs through it.

  • BuildSlaveSupport/build.webkit.org-config/public_html/dashboard/index.html:

Load Trac.js.

10:00 AM Changeset in webkit [161089] by akling@apple.com
  • 4 edits in trunk/Source/WebCore

Make Text::createTextRenderer() take a const RenderStyle&.
<​https://webkit.org/b/126136>

Nuke a FIXME about constifying a RenderStyle& local.

Reviewed by Anders Carlsson.

9:58 AM Changeset in webkit [161088] by akling@apple.com
  • 4 edits in trunk/Source/WebCore

Move more inlines from RenderObject to RenderElement.
<​https://webkit.org/b/126134>

Lift some inline functions that use style() from RenderObject over
to RenderElement, making them branchless.

Reviewed by Anders Carlsson.

7:04 AM Changeset in webkit [161087] by thiago.lacerda@openbossa.org
  • 3 edits in trunk/Source/WebKit2

[WK2][CoordinatedGraphics] Removing duplicate scale information from WebVIew.cpp
​https://bugs.webkit.org/show_bug.cgi?id=126243

Reviewed by Benjamin Poulain.

The scale factor of the WebView must reflect the same information present in
WebPageProxy.
Previously, in WebView.cpp, there was a m_contentScaleFactor member, which adds
information duplication and needs to be synchronizing with pageScaleFactor in
WebPageProxy. We can avoid this by just making WebView access and set WebPageProxy's
pageScaleFactor directly.

  • UIProcess/CoordinatedGraphics/WebView.cpp:

(WebKit::WebView::WebView):
(WebKit::WebView::setContentScaleFactor):
(WebKit::WebView::transformToScene):
(WebKit::WebView::visibleContentsSize):

  • UIProcess/CoordinatedGraphics/WebView.h:

(WebKit::WebView::contentScaleFactor):

1:24 AM Changeset in webkit [161086] by ChangSeok Oh
  • 2 edits in trunk/Source/WebCore

Unreviewed build fix after r159526.
isBoxValue is used at out of ENABLE_CSS_SHAPES gaurd. It causes a compile failure.
So I moved isBoxValue to out side of ENABLE_CSS_SHAPES.

  • css/CSSParser.cpp:

Dec 25, 2013:

11:32 PM Changeset in webkit [161085] by gyuyoung.kim@samsung.com
  • 5 edits in trunk/Source/WebCore

Set m_pos as private in InlineIterator, and use getter and setter functions.
​https://bugs.webkit.org/show_bug.cgi?id=125614

Reviewed by Alexey Proskuryakov.

InlineIterator has been exported m_pos as public directly though it is member variable.
This patch set it as private, and add getter/setter functions for it.

No new tests, no behavior changes.

  • rendering/InlineIterator.h:

(WebCore::InlineIterator::setOffset):
(WebCore::operator==):
(WebCore::operator!=):
(WebCore::InlineBidiResolver::appendRun):

  • rendering/RenderBlockLineLayout.cpp:

(WebCore::RenderBlockFlow::appendRunsForObject):
(WebCore::constructBidiRunsForLine):
(WebCore::RenderBlockFlow::layoutRunsAndFloatsInRange):
(WebCore::RenderBlockFlow::matchedEndLine):
(WebCore::LineBreaker::nextSegmentBreak):

  • rendering/line/BreakingContextInlineHeaders.h:

(WebCore::BreakingContext::handleBR):
(WebCore::BreakingContext::handleFloat):
(WebCore::iteratorIsBeyondEndOfRenderCombineText):
(WebCore::ensureCharacterGetsLineBox):
(WebCore::BreakingContext::handleText):
(WebCore::checkMidpoints):
(WebCore::BreakingContext::handleEndOfLine):

  • rendering/line/TrailingObjects.cpp:

(WebCore::TrailingObjects::updateMidpointsForTrailingBoxes):

11:17 PM Changeset in webkit [161084] by mark.lam@apple.com
  • 3 edits in branches/jsCStack/Source/JavaScriptCore

jsc-layout-tests.yaml/js/script-tests/function-apply-aliased.js.layout-no-cjit is failing.
​https://bugs.webkit.org/show_bug.cgi?id=126174.

Not yet reviewed.

When we do a stack check in a function prologue, the activation object
in the frame hasn't been set yet. The test failures came from the stack
unwinding code trying to tear off the frame to a non-existant activation
object. Since we haven't entered the function yet and the frame is
technically not fully "pushed" yet, we can throw i.e. start the unwinding
from the caller frame instead. This fixes the issue.

  • llint/LLIntSlowPaths.cpp:

(JSC::LLInt::LLINT_SLOW_PATH_DECL):

  • When we have a StackOverflowError, return the caller's CallFrame in the second value in the SlowPathReturnType.
  • llint/LowLevelInterpreter.asm:
  • Check if the second value of the SlowPathReturnType from the stack check is 0. If not 0, set the cfr to the returned CallFrame* before we start handling the StackOverflowError and unwinding the stack.
9:12 PM Changeset in webkit [161083] by commit-queue@webkit.org
  • 3 edits
    1 delete in trunk/Source/WebCore

[EFL] Delete file.edc and file_*.png.
​https://bugs.webkit.org/show_bug.cgi?id=125134

Patch by Kim Byung Jun <​bj1987.kim@samsung.com> on 2013-12-25
Reviewed by Gyuyoung Kim.

File_theme uses button form.

  • platform/efl/DefaultTheme/CMakeLists.txt:
  • platform/efl/DefaultTheme/default.edc:
  • platform/efl/DefaultTheme/widget/file/file.edc: Removed.
  • platform/efl/DefaultTheme/widget/file/file_focus.png: Removed.
  • platform/efl/DefaultTheme/widget/file/file_hover.png: Removed.
  • platform/efl/DefaultTheme/widget/file/file_normal.png: Removed.
  • platform/efl/DefaultTheme/widget/file/file_press.png: Removed.
7:06 PM Changeset in webkit [161082] by commit-queue@webkit.org
  • 2 edits in trunk/Tools

[EFL] Activate keyboard homepage shortcut on efl minibrowser.
​https://bugs.webkit.org/show_bug.cgi?id=124636

Patch by Jongwoo Choi <​jw0330.choi@samsung.com> on 2013-12-25
Reviewed by Gyuyoung Kim.

  • MiniBrowser/efl/main.c: Added.

(on_key_down): Add a homepage shortcut using Alt + Home key.

7:03 PM Changeset in webkit [161081] by thiago.lacerda@openbossa.org
  • 2 edits in trunk/Source/WebCore

[Nix] Adding createDragImageIconForCachedImageFilename method to DragImageNix
​https://bugs.webkit.org/show_bug.cgi?id=126230

Reviewed by Daniel Bates.

Also returning nullptr in other functions that were returning 0 as a pointer.

  • platform/nix/DragImageNix.cpp:

(WebCore::createDragImageFromImage):
(WebCore::createDragImageIconForCachedImage):
(WebCore::createDragImageIconForCachedImageFilename):

6:57 PM Changeset in webkit [161080] by thiago.lacerda@openbossa.org
  • 2 edits in trunk/Source/WebCore

[Nix] Fixing DragData::asFragment signature in DragDataNix.cpp
​https://bugs.webkit.org/show_bug.cgi?id=126229

Reviewed by Daniel Bates.

  • platform/nix/DragDataNix.cpp:

(WebCore::DragData::asFragment):

6:51 PM Changeset in webkit [161079] by thiago.lacerda@openbossa.org
  • 2 edits in trunk/Source/WTF

[Nix] Building with DRAG_SUPPORT enabled
​https://bugs.webkit.org/show_bug.cgi?id=126233

Reviewed by Daniel Bates.

  • wtf/nix/FeatureDefinesNix.h:
4:33 PM Changeset in webkit [161078] by ap@apple.com
  • 3 edits in trunk/Tools

Revision selection behavior at build.webkit.org/dashboard is annoying
​https://bugs.webkit.org/show_bug.cgi?id=126235

Reviewed by Tim Horton.

  • BuildSlaveSupport/build.webkit.org-config/public_html/dashboard/Scripts/BuildbotQueueView.js:

(BuildbotQueueView.prototype.revisionLinksForIteration): Make the whole message
selectable, not just the numbers in it.

  • BuildSlaveSupport/build.webkit.org-config/public_html/dashboard/Styles/Main.css:

(.selectable): Make selectable enable default behavior, not user-select:all.

4:29 PM Changeset in webkit [161077] by commit-queue@webkit.org
  • 29 edits in trunk

Unreviewed, rolling out r161033 and r161074.
​http://trac.webkit.org/changeset/161033
​http://trac.webkit.org/changeset/161074
​https://bugs.webkit.org/show_bug.cgi?id=126240

Oliver says that a rollout would be better (Requested by ap on
#webkit).

Source/JavaScriptCore:

  • API/JSObjectRef.cpp:

(JSObjectSetProperty):

  • dfg/DFGOperations.cpp:

(JSC::DFG::operationPutByValInternal):

  • interpreter/Interpreter.cpp:

(JSC::Interpreter::execute):

  • jit/JITOperations.cpp:
  • llint/LLIntSlowPaths.cpp:

(JSC::LLInt::LLINT_SLOW_PATH_DECL):

  • runtime/Arguments.cpp:

(JSC::Arguments::putByIndex):

  • runtime/ArrayPrototype.cpp:

(JSC::putProperty):
(JSC::arrayProtoFuncPush):

  • runtime/JSCJSValue.cpp:

(JSC::JSValue::putToPrimitiveByIndex):

  • runtime/JSCell.cpp:

(JSC::JSCell::putByIndex):

  • runtime/JSFunction.cpp:

(JSC::JSFunction::put):

  • runtime/JSGenericTypedArrayViewInlines.h:

(JSC::JSGenericTypedArrayView<Adaptor>::putByIndex):

  • runtime/JSONObject.cpp:

(JSC::Walker::walk):

  • runtime/JSObject.cpp:

(JSC::JSObject::putByIndex):
(JSC::JSObject::putDirectNonIndexAccessor):
(JSC::JSObject::deleteProperty):

  • runtime/JSObject.h:

(JSC::JSObject::putDirect):

  • runtime/Lookup.h:

(JSC::putEntry):
(JSC::lookupPut):

  • runtime/PutPropertySlot.h:

(JSC::PutPropertySlot::PutPropertySlot):
(JSC::PutPropertySlot::setNewProperty):
(JSC::PutPropertySlot::isCacheable):

Source/WebCore:

  • bindings/js/JSDOMWindowCustom.cpp:

(WebCore::JSDOMWindow::put):

  • bindings/objc/WebScriptObject.mm:

(-[WebScriptObject setValue:forKey:]):

  • bindings/scripts/CodeGeneratorJS.pm:

(GenerateImplementation):

  • bindings/scripts/test/JS/JSTestInterface.cpp:

(WebCore::JSTestInterface::putByIndex):

  • bridge/NP_jsobject.cpp:

(_NPN_SetProperty):

Source/WebKit/mac:

  • Plugins/Hosted/NetscapePluginInstanceProxy.mm:

(WebKit::NetscapePluginInstanceProxy::setProperty):

Source/WebKit2:

  • WebProcess/Plugins/Netscape/NPJSObject.cpp:

(WebKit::NPJSObject::setProperty):

LayoutTests:

4:29 PM Changeset in webkit [161076] by beidson@apple.com
  • 19 edits in trunk/Source

DatabaseProcess: Implement version changing
​https://bugs.webkit.org/show_bug.cgi?id=126099

Reviewed by Sam Weinig.

Source/WebCore:

No new tests (No change in WebCore behavior).

  • Modules/indexeddb/IDBTransactionBackend.cpp:

(WebCore::IDBTransactionBackend::commit): Update some logging.

  • Modules/indexeddb/IDBTransactionBackendOperations.cpp:

(WebCore::IDBDatabaseBackend::VersionChangeOperation::perform): Move some things that used to be in

IDBServerConnectionLevelDB::changeDatabaseVersion to this cross-platform location.

  • Modules/indexeddb/IDBTransactionBackendOperations.h:

(WebCore::IDBDatabaseBackend::VersionChangeOperation::transaction):

  • Modules/indexeddb/leveldb/IDBServerConnectionLevelDB.cpp:

(WebCore::IDBServerConnectionLevelDB::changeDatabaseVersion): Move the actual updating of the database backend

metadata to the VersionChangeOperation.

Source/WebKit2:

  • DatabaseProcess/IndexedDB/DatabaseProcessIDBConnection.cpp:

(WebKit::DatabaseProcessIDBConnection::getOrEstablishIDBDatabaseMetadata): Add logging.
(WebKit::DatabaseProcessIDBConnection::openTransaction): Ditto.
(WebKit::DatabaseProcessIDBConnection::beginTransaction): Ditto.
(WebKit::DatabaseProcessIDBConnection::commitTransaction): Ditto.
(WebKit::DatabaseProcessIDBConnection::resetTransaction): Ditto.
(WebKit::DatabaseProcessIDBConnection::rollbackTransaction): Ditto.
(WebKit::DatabaseProcessIDBConnection::changeDatabaseVersion): Implement this!

  • DatabaseProcess/IndexedDB/DatabaseProcessIDBConnection.h:
  • DatabaseProcess/IndexedDB/DatabaseProcessIDBConnection.messages.in:
  • DatabaseProcess/IndexedDB/UniqueIDBDatabase.cpp:

(WebKit::UniqueIDBDatabase::shutdown): Clean up all requests, and also cleanup the backing store

on the database queue.

(WebKit::UniqueIDBDatabase::shutdownBackingStore): Shutdown the backing store.
(WebKit::UniqueIDBDatabase::didShutdownBackingStore):
(WebKit::UniqueIDBDatabase::changeDatabaseVersion): Pass along the new version to the database queue.
(WebKit::UniqueIDBDatabase::didChangeDatabaseVersion):
(WebKit::UniqueIDBDatabase::changeDatabaseVersionInBackingStore): Set the new version in the backing store.
(WebKit::UniqueIDBDatabase::postMainThreadTask): Add a ref() to keep the database alive incase it is being

shutdown on the main thread.

(WebKit::UniqueIDBDatabase::performNextMainThreadTask): Balance the ref() from postMainThreadTask, and

accept an empty mainThreadTask queue.

  • DatabaseProcess/IndexedDB/UniqueIDBDatabase.h:
  • DatabaseProcess/IndexedDB/UniqueIDBDatabaseBackingStore.h: Add changeDatabaseVersion.

Flesh out SQLiteIDBTransaction to actually operation on a SQLiteTransaction in a real SQLiteDatabase:

  • DatabaseProcess/IndexedDB/sqlite/SQLiteIDBTransaction.cpp:

(WebKit::SQLiteIDBTransaction::SQLiteIDBTransaction):
(WebKit::SQLiteIDBTransaction::~SQLiteIDBTransaction):
(WebKit::SQLiteIDBTransaction::begin):
(WebKit::SQLiteIDBTransaction::commit):
(WebKit::SQLiteIDBTransaction::reset):
(WebKit::SQLiteIDBTransaction::rollback):
(WebKit::SQLiteIDBTransaction::inProgress):

  • DatabaseProcess/IndexedDB/sqlite/SQLiteIDBTransaction.h:

(WebKit::SQLiteIDBTransaction::create):
(WebKit::SQLiteIDBTransaction::mode):

Rename m_metadataDB to m_sqliteDB,

  • DatabaseProcess/IndexedDB/sqlite/UniqueIDBDatabaseBackingStoreSQLite.cpp:

(WebKit::UniqueIDBDatabaseBackingStoreSQLite::createAndPopulateInitialMetadata):
(WebKit::UniqueIDBDatabaseBackingStoreSQLite::extractExistingMetadata):
(WebKit::UniqueIDBDatabaseBackingStoreSQLite::openSQLiteDatabaseAtPath): Disable threading checks on the DB

as work queues can change threads.

(WebKit::UniqueIDBDatabaseBackingStoreSQLite::getOrEstablishMetadata):
(WebKit::UniqueIDBDatabaseBackingStoreSQLite::establishTransaction):
(WebKit::UniqueIDBDatabaseBackingStoreSQLite::beginTransaction):
(WebKit::UniqueIDBDatabaseBackingStoreSQLite::changeDatabaseVersion): After validating the transaction, actually

update the version number on disk.

  • DatabaseProcess/IndexedDB/sqlite/UniqueIDBDatabaseBackingStoreSQLite.h:
  • WebProcess/Databases/IndexedDB/WebIDBServerConnection.cpp:

(WebKit::WebIDBServerConnection::~WebIDBServerConnection): Make sure to abort in-flight requests.
(WebKit::WebIDBServerConnection::changeDatabaseVersion):
(WebKit::WebIDBServerConnection::didChangeDatabaseVersion):

  • WebProcess/Databases/IndexedDB/WebIDBServerConnection.h:
  • WebProcess/Databases/IndexedDB/WebIDBServerConnection.messages.in:
4:09 PM Changeset in webkit [161075] by ap@apple.com
  • 2 edits in trunk/LayoutTests

[Mac] [WK2] http/tests/security/cross-frame-access-put.html fails
​https://bugs.webkit.org/show_bug.cgi?id=116649

  • platform/mac-wk2/TestExpectations: Removing an expectation for this old bug that

was no longer occurring. Failures on this test are now tracked via bug 126238.

3:59 PM Changeset in webkit [161074] by ap@apple.com
  • 2 edits in trunk/LayoutTests

REGRESSION (r161033): http/tests/security/cross-frame-access-put.html fails
​https://bugs.webkit.org/show_bug.cgi?id=126238

3:54 PM Changeset in webkit [161073] by ap@apple.com
  • 2 edits in trunk/LayoutTests

REGRESSION (Lazy tree creation): css3/calc/transitions-dependent.html is frequently failing
​https://bugs.webkit.org/show_bug.cgi?id=126142

3:44 PM Changeset in webkit [161072] by fpizlo@apple.com
  • 11 edits
    3 adds in trunk

DFG PhantomArguments shouldn't rely on a dead Phi graph
​https://bugs.webkit.org/show_bug.cgi?id=126218

Source/JavaScriptCore:

Reviewed by Oliver Hunt.

This change dramatically rationalizes our handling of PhantomArguments (i.e.
speculative elision of arguments object allocation).

It's now the case that if we decide that we can elide arguments allocation, we just
turn the arguments-creating node into a PhantomArguments and mark all locals that
it's stored to as being arguments aliases. Being an arguments alias and being a
PhantomArguments means basically the same thing: in DFG execution you have the empty
value, on OSR exit an arguments object is allocated in your place, and all operations
that use the value now just refer directly to the actual arguments in the call frame
header (or the arguments we know that we passed to the call, in case of inlining).

This means that we no longer have arguments simplification creating a dead Phi graph
that then has to be interpreted by the OSR exit logic. That sort of never made any
sense.

This means that PhantomArguments now has a clear story in SSA: basically SSA just
gets rid of the "locals" but everything else is the same.

Finally, this means that we can more easily get rid of forward exiting. As I was
working on the code to get rid of forward exiting, I realized that I'd have to
carefully preserve the special meanings of MovHint and SetLocal in the case of
PhantomArguments. It was really bizarre: even the semantics of MovHint were tied to
our specific treatment of PhantomArguments. After this change this is no longer the
case.

One of the really cool things about this change is that arguments reification now
just becomes a special kind of FlushFormat. This further unifies things: it means
that a MovHint(PhantomArguments) and a SetLocal(PhantomArguments) both have the same
meaning, since both of them dictate that the way we recover the local on exit is by
reifying arguments. Previously, the SetLocal(PhantomArguments) case needed some
special handling to accomplish this.

A downside of this approach is that we will now emit code to store the empty value
into aliased arguments variables, and we will even emit code to load that empty value
as well. As far as I can tell this doesn't cost anything, since PhantomArguments are
most profitable in cases where it allows us to simplify control flow and kill the
arguments locals entirely. Of course, this isn't an issue in SSA form since SSA form
also eliminates the locals.

  • dfg/DFGArgumentsSimplificationPhase.cpp:

(JSC::DFG::ArgumentsSimplificationPhase::run):
(JSC::DFG::ArgumentsSimplificationPhase::detypeArgumentsReferencingPhantomChild):

  • dfg/DFGFlushFormat.cpp:

(WTF::printInternal):

  • dfg/DFGFlushFormat.h:

(JSC::DFG::resultFor):
(JSC::DFG::useKindFor):
(JSC::DFG::dataFormatFor):

  • dfg/DFGSpeculativeJIT.cpp:

(JSC::DFG::SpeculativeJIT::compileCurrentBlock):

  • dfg/DFGSpeculativeJIT32_64.cpp:

(JSC::DFG::SpeculativeJIT::compile):

  • dfg/DFGSpeculativeJIT64.cpp:

(JSC::DFG::SpeculativeJIT::compile):

  • dfg/DFGValueSource.h:

(JSC::DFG::ValueSource::ValueSource):
(JSC::DFG::ValueSource::forFlushFormat):

  • dfg/DFGVariableAccessData.h:

(JSC::DFG::VariableAccessData::flushFormat):

  • ftl/FTLLowerDFGToLLVM.cpp:

(JSC::FTL::LowerDFGToLLVM::buildExitArguments):

LayoutTests:

Reviewed by Oliver Hunt.

Added a test for an obvious case that I don't think we had coverage for in
microbenchmarks. Of course, this case was already covered by more complex tests.

  • js/regress/inline-arguments-aliased-access-expected.txt: Added.
  • js/regress/inline-arguments-aliased-access.html: Added.
  • js/regress/script-tests/inline-arguments-aliased-access.js: Added.

(foo):
(bar):

12:39 PM Changeset in webkit [161071] by Martin Robinson
  • 2 edits in trunk/Source/WebKit2

[GTK] [CMake] Fix CMake style errors in Source/WebKit2/PlatformGTK.cmake
​https://bugs.webkit.org/show_bug.cgi?id=126221

Reviewed by Gustavo Noronha Silva.

  • PlatformGTK.cmake: Fix CMake style errors.
12:38 PM Changeset in webkit [161070] by Martin Robinson
  • 2 edits in trunk

[GTK] [CMake] Properly name the JavaScriptCore library
​https://bugs.webkit.org/show_bug.cgi?id=126220

Reviewed by Gustavo Noronha Silva.

  • Source/cmake/OptionsGTK.cmake: Set the JSC library name properly.
11:24 AM Changeset in webkit [161069] by rniwa@webkit.org
  • 2 edits in trunk/Websites/webkit.org

Add a description about webkit-patch setup-git-clone to checkout.html
​https://bugs.webkit.org/show_bug.cgi?id=126217

Reviewed by Philippe Normand.

Added.

  • building/checkout.html:
11:07 AM Changeset in webkit [161068] by Martin Robinson
  • 20 edits in trunk

[GTK] [CMake] Clean up generated sources directories
​https://bugs.webkit.org/show_bug.cgi?id=126216

Reviewed by Gustavo Noronha Silva.

.:

  • Source/cmake/OptionsGTK.cmake: Pre-define the main derived sources directories

so that they definitions can be shared easily throughout the build system.

Source/WebCore:

  • PlatformGTK.cmake: Use the new directory variables.

Source/WebKit:

  • PlatformGTK.cmake: Use the new derived source variables.

Source/WebKit2:

  • PlatformGTK.cmake: Use the new derived sources variables.
  • UIProcess/API/gtk/tests/CMakeLists.txt: Ditto.

Tools:

  • DumpRenderTree/PlatformGTK.cmake: Use the new derived sources variables.
  • GtkLauncher/CMakeLists.txt: Ditto.
  • MiniBrowser/gtk/CMakeLists.txt: Ditto.
  • TestWebKitAPI/CMakeLists.txt: Now that the EFL and GTK+ forwarding header locations differ,

move the platform-specific includes to the EFL and GTK platform files. Also clean up forwarding
header generation into on target.

  • TestWebKitAPI/PlatformEfl.cmake: Added EFL-specific locations to the include list.
  • TestWebKitAPI/PlatformGTK.cmake: Added GTK-specific locations to the include list.
  • WebKitTestRunner/CMakeLists.txt: Same as for TestWebKitAPI.
  • WebKitTestRunner/PlatformEfl.cmake: Ditto.
  • WebKitTestRunner/PlatformGTK.cmake: Ditto.
  • gtk/generate-gtkdoc: Update to select the appropriate derived sources locations for

the cmake build.

11:03 AM Changeset in webkit [161067] by krit@webkit.org
  • 8 edits in trunk

Support <box> values parsing on 'clip-path' property
​https://bugs.webkit.org/show_bug.cgi?id=126147

Reviewed by Ryosuke Niwa.

Source/WebCore:

Support parsing of the background reference boxes, margin-box and bounding-box.
A box will be a reference box and define the origin for a basic shape.
If no basic shape is specified, the box defines the clipping path itself.
The specification text follows the changes to CSS Shapes now.

​https://dvcs.w3.org/hg/FXTF/raw-file/3f213145303e/css-masking-1/index.html#the-clip-path

Existing parsing test have been extended to test box values as well.

  • css/CSSParser.cpp:

(WebCore::CSSParser::parseValue):
(WebCore::CSSParser::parseClipPath):

  • css/CSSParser.h:
  • css/CSSValueKeywords.in:
  • css/DeprecatedStyleBuilder.cpp:

(WebCore::ApplyPropertyClipPath::applyValue):

LayoutTests:

Add test to check different combinations of the reference boxes
content-box, padding-box, border-box, margin-box and bounding-box
with other clip-path values.

  • fast/masking/parsing-clip-path-shape.html:
10:58 AM Changeset in webkit [161066] by ddkilzer@apple.com
  • 2 edits
    4 adds in trunk/Source/WebCore

[iOS] Upstream WebCore/pdf changes
​http://webkit.org/b/126097

Reviewed by Sam Weinig.

  • WebCore.xcodeproj/project.pbxproj: Added files to project.
  • pdf/ios/PDFDocument.cpp: Added.

(WebCore::PDFDocumentParser::create):
(WebCore::PDFDocumentParser::document):
(WebCore::PDFDocumentParser::PDFDocumentParser):
(WebCore::PDFDocument::createParser):

  • pdf/ios/PDFDocument.h: Added.

(WebCore::PDFDocument::create):
(WebCore::PDFDocument::PDFDocument):

6:48 AM Changeset in webkit [161065] by Philippe Normand
  • 2 edits in trunk/Source/WebKit2

Unreviewed, GTK build fix after r161063

  • GNUmakefile.am: Use locally built .gir dependencies when

building the WebExtensions gir file.

1:56 AM Changeset in webkit [161064] by Carlos Garcia Campos
  • 2 edits in trunk/Source/WebKit2

[GTK] Fix warnings when generating WebExtensions gobject-instrospection files
​https://bugs.webkit.org/show_bug.cgi?id=126225

Reviewed by Philippe Normand.

Add transfer annotation to webkit_web_page_get_dom_document() and
document webkit_web_page_get_main_frame().

  • WebProcess/InjectedBundle/API/gtk/WebKitWebPage.cpp:
1:53 AM Changeset in webkit [161063] by Carlos Garcia Campos
  • 2 edits in trunk/Source/WebKit2

[GTK] Missing introspection information for WebExtensions
​https://bugs.webkit.org/show_bug.cgi?id=122407

Reviewed by Philippe Normand.

Generate WebKit2WebExtension-3.0.gir with the WebExtensions API
and move the DOM bindings API from WebKit2-3.0.gir to
WebKit2WebExtension-3.0.gir.

  • GNUmakefile.am:
1:13 AM Changeset in webkit [161062] by ap@apple.com
  • 9 edits
    1 add in trunk/Tools

Add a JSON interface for getting EWS queue status
​https://bugs.webkit.org/show_bug.cgi?id=126182

Reviewed by Ryosuke Niwa.

The current goal is to add an indicator of how many patches are in the queue, and
the JSON also includes some other information that was easy to add.

  • QueueStatusServer/app.yaml: Updated version for deployment.
  • QueueStatusServer/handlers/queuestatusjson.py: Added. Returns a JSON with some

bot and patch queue information.

  • QueueStatusServer/index.yaml: Added an index for a new query that only returns

bot ids.

  • QueueStatusServer/main.py: Map /queue-status-json/ directory to QueueStatusJSON

handler.

  • QueueStatusServer/config: Added property svn:ignore.
  • QueueStatusServer/filters: Added property svn:ignore.
  • QueueStatusServer/loggers: Added property svn:ignore.
  • QueueStatusServer/handlers: Added property svn:ignore.
  • QueueStatusServer/model: Added property svn:ignore.

Ignore compiled *.pyc files in all directories they appear in when running locally.

Note: See TracTimeline for information about the timeline view.