Make it easier to use NoLockingNecessary

[WebKit-https.git] / Source / WTF / ChangeLog

2016-06-02  Filip Pizlo  <fpizlo@apple.com>

        Make it easier to use NoLockingNecessary
        https://bugs.webkit.org/show_bug.cgi?id=158306

        Reviewed by Keith Miller.
        
        An idiom that we borrowed from LLVM is that if a function requires a lock to be held, we
        have it take a const Locker& as its first argument. This may not communicate which lock is
        to be held, but it does help us to remember that some lock must be held. So far, it's been
        relatively easy to then figure out which lock. We've had bugs where we forgot to hold a
        lock but I don't remember the last time we had a bug where we grabbed the wrong lock.
        
        But sometimes, we know at the point where we call such a method that we actually don't
        need to hold any lock. This usually happens during object construction. If we're
        constructing some object then we usually know that we have not escaped it yet, so we don't
        need to waste time acquiring its lock. We could solve this by having a separate set of
        methods that don't do or require locking. This would be cumbersome, since usually for
        every variant that takes const Locker&, there is already one that doesn't, and that one
        will grab the lock for you. So this means having a third variant, that also doesn't take a
        const Locker&, but does no locking. That's pretty weird.
        
        So, we introduced NoLockingNecessary for situations like this. The idiom went like so:
        
            Locker<Whatever> locker(Locker<Whatever>::NoLockingNecessary)
            stuff->foo(locker);
        
        Usually though, there would be some distance between where the locker is defined and where
        it's used, so when you just look at stuff->foo(locker) in isolation you don't know if this
        is a real locker or a NoLockingNecessary cast. Also, requiring two lines for this just
        adds code.
        
        This change makes this easier. Now you can just do:
        
            stuff->foo(NoLockingNecessary).
        
        This is because NoLockingNecessary has been pulled out into the WTF namespace (and is
        usinged from the global namespace) and the Locker<> constructor that takes
        NoLockingNecessaryTag is now implicit.
        
        The only possible downside of this change is that people might use this idiom more
        frequently now that it's easier to use. I don't think that's a bad thing. I'm now
        convinced that this is not a bad idiom. When I was fixing an unrelated bug, I almost went
        the way of adding more locking to some core JSC data structures, and in the process, I
        needed to use NoLockingNecessary. It's clear that this is a general-purpose idiom and we
        should not impose artificial constraints on its use.

        * wtf/Locker.h:
        (WTF::Locker::Locker):
        (WTF::Locker::~Locker):

2016-06-01  Brady Eidson  <beidson@apple.com>

        Get rid of StringCapture.
        https://bugs.webkit.org/show_bug.cgi?id=158285

        Reviewed by Chris Dumez.

        * wtf/text/WTFString.h:
        (WTF::StringCapture::StringCapture): Deleted.
        (WTF::StringCapture::string): Deleted.
        (WTF::StringCapture::releaseString): Deleted.
        (WTF::StringCapture::operator=): Deleted.

2016-06-01  Benjamin Poulain  <bpoulain@apple.com>

        [JSC] Some setters for components of Date do not timeClip() their result
        https://bugs.webkit.org/show_bug.cgi?id=158278

        Unreviewed.

        * wtf/DateMath.cpp:
        (WTF::equivalentYearForDST): Deleted.
        The assertion is bogus.
        As the comments above explains, the function is completely wrong for years
        outside [1900-2100].
        The tests passing large values for years are failing (year <= maxYear).
        The weird NaN test is a mystery. The old changelog does not explain it.

2016-05-31  Commit Queue  <commit-queue@webkit.org>

        Unreviewed, rolling out r201363 and r201456.
        https://bugs.webkit.org/show_bug.cgi?id=158240

        "40% regression on date-format-xparb" (Requested by
        keith_miller on #webkit).

        Reverted changesets:

        "LLInt should be able to cache prototype loads for values in
        GetById"
        https://bugs.webkit.org/show_bug.cgi?id=158032
        http://trac.webkit.org/changeset/201363

        "get_by_id should support caching unset properties in the
        LLInt"
        https://bugs.webkit.org/show_bug.cgi?id=158136
        http://trac.webkit.org/changeset/201456

2016-05-31  Brady Eidson  <beidson@apple.com>

        Make createCrossThreadTask() functions return on the stack instead of the heap.
        https://bugs.webkit.org/show_bug.cgi?id=158215

        Reviewed by Darin Adler.

        * WTF.xcodeproj/project.pbxproj:

        * wtf/CrossThreadCopier.cpp:

        * wtf/CrossThreadQueue.h: Added. A lightweight of MessageQueue that deals directly
          in objects instead of in std::unique_ptrs.
        (WTF::CrossThreadQueue::isKilled):
        (WTF::CrossThreadQueue<DataType>::append):
        (WTF::CrossThreadQueue<DataType>::waitForMessage):
        (WTF::CrossThreadQueue<DataType>::tryGetMessage):

        * wtf/CrossThreadTask.h:
        (WTF::createCrossThreadTask):
        (WTF::CrossThreadTask::CrossThreadTask): Deleted.

2016-05-30  Brady Eidson  <beidson@apple.com>

        Move CrossThreadCopier/CrossThreadTask to WTF.
        https://bugs.webkit.org/show_bug.cgi?id=158207

        Reviewed by Alex Christensen.

        * WTF.xcodeproj/project.pbxproj:
        * wtf/CMakeLists.txt:

        * wtf/CrossThreadCopier.cpp: Renamed from Source/WebCore/platform/CrossThreadCopier.cpp.
        * wtf/CrossThreadCopier.h: Renamed from Source/WebCore/platform/CrossThreadCopier.h.
        (WTF::CrossThreadCopierPassThrough::copy):
        
        * wtf/CrossThreadTask.h: Renamed from Source/WebCore/platform/CrossThreadTask.h.
        (WTF::CrossThreadTask::CrossThreadTask):
        (WTF::CrossThreadTask::performTask):
        (WTF::createCrossThreadTask):

2016-05-28  Chris Dumez  <cdumez@apple.com>

        Templatize NoncopyableFunction class similarly to std::function
        https://bugs.webkit.org/show_bug.cgi?id=158185

        Reviewed by Darin Adler.

        Templatize NoncopyableFunction class similarly to std::function, so
        that it can be used as a std::function replacement in more places.

        Previously, NoncopyableFunction could only support "void()" lambdas.

        * wtf/FunctionDispatcher.h:
        * wtf/MainThread.cpp:
        (WTF::functionQueue):
        (WTF::dispatchFunctionsFromMainThread):
        (WTF::callOnMainThread):
        * wtf/MainThread.h:
        * wtf/NoncopyableFunction.h:
        * wtf/RunLoop.cpp:
        (WTF::RunLoop::performWork):
        (WTF::RunLoop::dispatch):
        * wtf/RunLoop.h:
        * wtf/WorkQueue.h:
        * wtf/cocoa/WorkQueueCocoa.cpp:
        (WTF::WorkQueue::dispatch):
        (WTF::WorkQueue::dispatchAfter):
        * wtf/efl/DispatchQueueWorkItemEfl.h:
        (WorkItem::WorkItem):
        (TimerWorkItem::create):
        (TimerWorkItem::TimerWorkItem):
        * wtf/efl/WorkQueueEfl.cpp:
        (WTF::WorkQueue::dispatch):
        (WTF::WorkQueue::dispatchAfter):
        * wtf/generic/RunLoopGeneric.cpp:
        (WTF::RunLoop::TimerBase::ScheduledTask::create):
        (WTF::RunLoop::TimerBase::ScheduledTask::ScheduledTask):
        (WTF::RunLoop::dispatchAfter):
        * wtf/generic/WorkQueueGeneric.cpp:
        (WorkQueue::dispatch):
        (WorkQueue::dispatchAfter):
        * wtf/glib/RunLoopGLib.cpp:
        (WTF::DispatchAfterContext::DispatchAfterContext):
        (WTF::RunLoop::dispatchAfter):
        * wtf/win/WorkItemWin.cpp:
        (WTF::WorkItemWin::WorkItemWin):
        (WTF::WorkItemWin::create):
        (WTF::HandleWorkItem::HandleWorkItem):
        (WTF::HandleWorkItem::createByAdoptingHandle):
        * wtf/win/WorkItemWin.h:
        (WTF::WorkItemWin::function):
        * wtf/win/WorkQueueWin.cpp:
        (WTF::WorkQueue::dispatch):
        (WTF::WorkQueue::dispatchAfter):

2016-05-28  Dan Bernstein  <mitz@apple.com>

        Build fix for projects that include MainThread.h without including FastMalloc.h.

        * wtf/NoncopyableFunction.h: Include FastMalloc.h from here.

2016-05-27  Chris Dumez  <cdumez@apple.com>

        callOnMainThread() should not copy captured lambda variables
        https://bugs.webkit.org/show_bug.cgi?id=158166

        Reviewed by Brady Eidson.

        callOnMainThread() should not copy captured lambda variables. This
        function is usually called cross-thread with a lambda and copying
        the lambda (and its captured variables) can lead to thread-safety
        issues.

        This patch updates callOnMainThread() to take a NoncopyableFunction&&
        in parameter instead of a std::function. The call sites of
        callOnMainThread() have also been updated to use C++14's lambda
        capture with initializer.

        * WTF.xcodeproj/project.pbxproj:

        * wtf/FunctionDispatcher.h:
        * wtf/NoncopyableFunction.h:
        - Moved NoncopyableFunction from FunctionDispatcher.h to
          NoncopyableFunction.h.
        - Add a new operator=(nullptr_t) operator to NoncopyableFunction to
          match std::function, as one of the call sites needed it.

        * wtf/MainThread.cpp:
        (WTF::functionQueue):
        (WTF::dispatchFunctionsFromMainThread):
        (WTF::callOnMainThread):
        * wtf/MainThread.h:

2016-05-27  Yusuke Suzuki  <utatane.tea@gmail.com>

        Unreviewed, build fix for JSCOnly port.
        https://bugs.webkit.org/show_bug.cgi?id=158111

        Use NoncopyableFunction instead of std::function<>.

        * wtf/generic/RunLoopGeneric.cpp:
        (WTF::RunLoop::TimerBase::ScheduledTask::create):
        (WTF::RunLoop::TimerBase::ScheduledTask::ScheduledTask):

2016-05-27  Chris Dumez  <cdumez@apple.com>

        WorkQueue::dispatch() / RunLoop::dispatch() should not copy captured lambda variables
        https://bugs.webkit.org/show_bug.cgi?id=158111

        Reviewed by Darin Adler.

        WorkQueue::dispatch() / RunLoop::dispatch() should not copy captured lambda variables.
        These are often used cross-thread and copying the captured lambda variables can be
        dangerous (e.g. we do not want to copy a String after calling isolatedCopy() upon
        capture).

        This patch introduces a new NoncopyableFunction type that behaves similarly to
        std::function but guarantees that the passed-in lambda (and its captured variables)
        cannot be copied. This new NoncopyableFunction type is now used for
        WorkQueue / RunLoop's dispatch() / dispatchAfter() which are commonly used
        cross-thread. This should now allow us to call WorkQueue::dispatch() with a lambda
        that captures a String like so:
        [str = str.isolatedCopy()]() { }

        Also note that even though this is not leveraged in this patch, NoncopyableFunction
        would allow us to capture move-only types such as std::unique_ptr as so:
        [p = WTFMove(p)]() { }
        This does not work if we convert the lambda into an std::function because
        std::function requires the lambda to be copyable, NoncopyableFunction does not.

        * wtf/FunctionDispatcher.h:
        (WTF::CallableWrapperBase::~CallableWrapperBase):
        (WTF::NoncopyableFunction::NoncopyableFunction):
        (WTF::NoncopyableFunction::operator()):
        (WTF::NoncopyableFunction::operator bool):
        (WTF::NoncopyableFunction::operator=):
        * wtf/RunLoop.cpp:
        (WTF::RunLoop::performWork):
        (WTF::RunLoop::dispatch):
        * wtf/RunLoop.h:
        * wtf/WorkQueue.h:
        * wtf/cocoa/WorkQueueCocoa.cpp:
        (WTF::WorkQueue::dispatch):
        (WTF::WorkQueue::dispatchAfter):
        * wtf/efl/DispatchQueueWorkItemEfl.h:
        (WorkItem::WorkItem):
        (TimerWorkItem::create):
        (TimerWorkItem::TimerWorkItem):
        * wtf/efl/WorkQueueEfl.cpp:
        (WTF::WorkQueue::dispatch):
        (WTF::WorkQueue::dispatchAfter):
        * wtf/generic/RunLoopGeneric.cpp:
        (WTF::RunLoop::dispatchAfter):
        * wtf/generic/WorkQueueGeneric.cpp:
        (WorkQueue::dispatch):
        (WorkQueue::dispatchAfter):
        * wtf/glib/RunLoopGLib.cpp:
        (WTF::DispatchAfterContext::DispatchAfterContext):
        (WTF::RunLoop::dispatchAfter):
        * wtf/win/WorkItemWin.cpp:
        (WTF::WorkItemWin::WorkItemWin):
        (WTF::WorkItemWin::create):
        (WTF::HandleWorkItem::HandleWorkItem):
        (WTF::HandleWorkItem::createByAdoptingHandle):
        * wtf/win/WorkItemWin.h:
        (WTF::WorkItemWin::function):
        * wtf/win/WorkQueueWin.cpp:
        (WTF::WorkQueue::dispatch):
        (WTF::WorkQueue::timerCallback):
        (WTF::WorkQueue::dispatchAfter):

2016-05-26  Filip Pizlo  <fpizlo@apple.com>

        ScopedLambda should have a lifetime story that makes sense to the compiler
        https://bugs.webkit.org/show_bug.cgi?id=158118

        Reviewed by Mark Lam.

        Prior to this change, there were two lifetime bugs in ScopedLambda:
        
        - scopedLambda(Functor&&) would bind Functor to const lambda&, so the resulting ScopedLambdaFunctor
          would hold a reference to the original lambda. This would have surprising behavior; for example
          it meant that this code was wrong:
          
          auto l = scopedLambda<things>([&] ...);
          
          The solution is to have explicit copy/move versions of scopedLambda() rather than rely on perfect
          forwarding.
        
        - ScopedLambdaFunctor did not override its copy or move operations, so if the compiler did not RVO
          scopedLambda(), it would return a ScopedLambdaFunctor whose m_arg points to a dead temporary
          ScopedLambdaFunctor instance. The solution is to have explicit copy/move constructors and
          operators, which preserve the invariant that ScopedLambda::m_arg points to this.
        
        One nice side-effect of all of these constructors and operators being explicit is that we can rely
        on WTFMove's excellent assertions, which helped catch the first issue.
        
        This reverts ParkingLot to use ScopedLambda again.

        * wtf/ParkingLot.cpp:
        (WTF::ParkingLot::parkConditionallyImpl):
        (WTF::ParkingLot::unparkOne):
        (WTF::ParkingLot::unparkOneImpl):
        * wtf/ParkingLot.h:
        (WTF::ParkingLot::parkConditionally):
        (WTF::ParkingLot::unparkOne):
        * wtf/ScopedLambda.h:
        (WTF::scopedLambda):

2016-05-25  Anders Carlsson  <andersca@apple.com>

        Get rid of WTF/Functional.h
        https://bugs.webkit.org/show_bug.cgi?id=158081

        Reviewed by Chris Dumez.

        This is no longer used, and removing it will free up the name for a new Functional.h implementation.

        * WTF.xcodeproj/project.pbxproj:
        * wtf/Functional.h: Removed.
        (WTF::RefAndDeref::ref): Deleted.
        (WTF::RefAndDeref::deref): Deleted.
        (WTF::ParamStorageTraits::wrap): Deleted.
        (WTF::ParamStorageTraits::unwrap): Deleted.
        (WTF::ParamStorageTraits<PassRefPtr<T>>::wrap): Deleted.
        (WTF::ParamStorageTraits<PassRefPtr<T>>::unwrap): Deleted.
        (WTF::ParamStorageTraits<RefPtr<T>>::wrap): Deleted.
        (WTF::ParamStorageTraits<RefPtr<T>>::unwrap): Deleted.
        (WTF::ParamStorageTraits<RetainPtr<T>>::wrap): Deleted.
        (WTF::ParamStorageTraits<RetainPtr<T>>::unwrap): Deleted.
        (WTF::FunctionImplBase::~FunctionImplBase): Deleted.
        (WTF::FunctionBase::isNull): Deleted.
        (WTF::FunctionBase::FunctionBase): Deleted.
        (WTF::FunctionBase::impl): Deleted.
        (WTF::bind): Deleted.
        * wtf/mac/DeprecatedSymbolsUsedBySafari.mm:

2016-05-24  Keith Miller  <keith_miller@apple.com>

        LLInt should be able to cache prototype loads for values in GetById
        https://bugs.webkit.org/show_bug.cgi?id=158032

        Reviewed by Filip Pizlo.

        Add move constructors/initializers to Bags.

        * wtf/Bag.h:
        (WTF::Bag::Bag):
        (WTF::Bag::operator=):

2016-05-24  Chris Dumez  <cdumez@apple.com>

        Use auto for some of our lambda function parameters
        https://bugs.webkit.org/show_bug.cgi?id=158001

        Reviewed by Darin Adler.

        Use auto for some of our lambda function parameters now that we build with c++14.

        * wtf/BubbleSort.h:
        (WTF::bubbleSort):

2016-05-23  Chris Dumez  <cdumez@apple.com>

        Speed up move of vectors of POD types that have an inline buffer
        https://bugs.webkit.org/show_bug.cgi?id=158003

        Reviewed by Benjamin Poulain.

        When moving a vector of POD types that have an inline buffer, we would
        call std::swap() on the inline buffers. This unnecessarily slow because:
        1. It does not consider the vector size, and therefore may end up doing
           more work than necessary when the inline buffer is not full.
        2. In the "move" case, the destination buffer is completely empty so
           we don't really want a swap. We merely want the move the content of
           the source's inline buffer into the destination's one.

        Instead of calling std::swap(), we now call swapInlineBuffers() which
        was already used for non-POD types. swapInlineBuffers() will do just
        what we want in the "move" case because swapBound is going to be 0.
        As a result, we will only move the content of the source buffer into
        the destination one. Also swapInlineBuffers() is aware of the source
        vector's size so it will only move what's strictly needed.

        This seems to be a 2% progression on Dromaeo DOM attributes test.

        * wtf/Vector.h:
        (WTF::VectorBuffer::swapInlineBuffer):

2016-05-22  Dan Bernstein  <mitz@apple.com>

        Added NSEventMaskMouseMoved to AppKitCompatibilityDeclarations.h.

        Rubber-stamped by Anders Carlsson.

        * wtf/mac/AppKitCompatibilityDeclarations.h:

2016-05-22  Dan Bernstein  <mitz@apple.com>

        Another attempt to fix the GTK build after my previous changes.

        * wtf/StdLibExtras.h:

2016-05-22  Dan Bernstein  <mitz@apple.com>

        Tried to fix the GTK build after r201257.

        * wtf/StdLibExtras.h:

2016-05-22  Dan Bernstein  <mitz@apple.com>

        Additional fixes for non-C++14 Apple projects that use WTF.

        * wtf/StdLibExtras.h:
        (std::exchange):

2016-05-22  Dan Bernstein  <mitz@apple.com>

        Additional fixes for non-C++14 Apple projects that use WTF.

        * wtf/StdLibExtras.h:
        (std::make_unique):
        (std::move):

2016-05-22  Dan Bernstein  <mitz@apple.com>

        Correct the previous build fix attempt.

        * wtf/StdLibExtras.h:

2016-05-22  Dan Bernstein  <mitz@apple.com>

        Build fix for non-C++14 Apple projects that use WTF.

        Rubber-stamped by Anders.

        * wtf/StdLibExtras.h:

2016-05-22  Brady Eidson  <beidson@apple.com>

        Move to C++14.
        https://bugs.webkit.org/show_bug.cgi?id=157948

        Reviewed by Michael Catanzaro.

        * Configurations/Base.xcconfig:
        
        Delete a lot of "stl additions until we can adopt C++14" code:
        * wtf/StdLibExtras.h:
        (std::make_unique): Deleted.
        (std::index_sequence::size): Deleted.
        (std::exchange): Deleted.
        (std::literals::chrono_literals::operator _s): Deleted.
        (std::literals::chrono_literals::operator _ms): Deleted.

2016-05-20  Rawinder Singh  <rawinder.singh-webkit@cisra.canon.com.au>

        Implement operator== for WeakPtr
        https://bugs.webkit.org/show_bug.cgi?id=157883

        Reviewed by Chris Dumez.

        Implement operator== and operator!= for WeakPtr and update code to use the operators.

        * wtf/WeakPtr.h:
        (WTF::operator==):
        (WTF::operator!=):

2016-05-19  Chris Dumez  <cdumez@apple.com>

        Improve compile-time assertions in is<>() / downcast<>()
        https://bugs.webkit.org/show_bug.cgi?id=157817

        Reviewed by Darin Adler.

        * wtf/Ref.h:
        (WTF::is):
        Add is<>() overloads taking a Ref<>() so that is<>() keeps working when
        passing a Ref<>(), despite the new static assertions on the input type.
        Some call sites were already leveraging this as it was working by
        implicitly converting the Ref<T> into a T&.

        * wtf/TypeCasts.h:
        (WTF::is):
        (WTF::downcast):
        Make static assertions stricter in is<>() / downcast<>() to catch more
        cases where those are either unnecessary or incorrect.

2016-05-19  Filip Pizlo  <fpizlo@apple.com>

        Unreviewed, fix all of the builds. I failed a second time.

        * wtf/BackwardsGraph.h:
        (WTF::BackwardsGraph::rootName):

2016-05-19  Filip Pizlo  <fpizlo@apple.com>

        Unreviewed, fix all of the builds. I had made an additional change that I did not mean to
        commit. This fixes it.

        * wtf/BackwardsGraph.h:
        (WTF::BackwardsGraph::rootName):
        (WTF::BackwardsGraph::Set::dump):
        (WTF::BackwardsGraph::dump):

2016-05-18  Filip Pizlo  <fpizlo@apple.com>

        DFG::LICMPhase shouldn't hoist type checks unless it knows that the check will succeed at the loop pre-header
        https://bugs.webkit.org/show_bug.cgi?id=144527

        Reviewed by Saam Barati.
        
        This adds an adaptor for graphs called BackwardsGraph. The WTF graph framework is based on
        passing around a Graph template argument that follows the protocol shared by DFG::CFG,
        B3::CFG, and Air::CFG. These graphs always have a single root node but may have many leaf
        nodes. This new BackwardsGraph adaptor reverses the graph by creating a synthetic return
        node that it uses as the root in the inverted graph. This currently may resort to some
        heuristics in programs that have an infinite loop, but other than that, it'll work well in
        the general case.
        
        This allows us to say Dominators<BackwardsGraph<some graph type>> as a way of computing
        backwards dominators, which then allows us to easily answer control flow equivalence
        queries.

        * CMakeLists.txt:
        * WTF.xcodeproj/project.pbxproj:
        * wtf/BackwardsGraph.h: Added.
        (WTF::BackwardsGraph::Node::Node):
        (WTF::BackwardsGraph::Node::root):
        (WTF::BackwardsGraph::Node::operator==):
        (WTF::BackwardsGraph::Node::operator!=):
        (WTF::BackwardsGraph::Node::operator bool):
        (WTF::BackwardsGraph::Node::isRoot):
        (WTF::BackwardsGraph::Node::node):
        (WTF::BackwardsGraph::Set::Set):
        (WTF::BackwardsGraph::Set::add):
        (WTF::BackwardsGraph::Set::remove):
        (WTF::BackwardsGraph::Set::contains):
        (WTF::BackwardsGraph::Set::dump):
        (WTF::BackwardsGraph::Map::Map):
        (WTF::BackwardsGraph::Map::clear):
        (WTF::BackwardsGraph::Map::size):
        (WTF::BackwardsGraph::Map::operator[]):
        (WTF::BackwardsGraph::BackwardsGraph):
        (WTF::BackwardsGraph::root):
        (WTF::BackwardsGraph::newMap):
        (WTF::BackwardsGraph::successors):
        (WTF::BackwardsGraph::predecessors):
        (WTF::BackwardsGraph::index):
        (WTF::BackwardsGraph::node):
        (WTF::BackwardsGraph::numNodes):
        (WTF::BackwardsGraph::dump):
        * wtf/Dominators.h:
        (WTF::Dominators::Dominators):
        (WTF::Dominators::dump):
        (WTF::Dominators::LengauerTarjan::computeDepthFirstPreNumbering):
        * wtf/GraphNodeWorklist.h:
        (WTF::GraphNodeWith::GraphNodeWith):
        (WTF::GraphNodeWith::operator bool):
        * wtf/StdLibExtras.h:
        (WTF::callStatelessLambda):
        (WTF::checkAndSet):

2016-05-18  Saam barati  <sbarati@apple.com>

        StringBuilder::appendQuotedJSONString doesn't properly protect against the math it's doing. Make the math fit the assertion.
        https://bugs.webkit.org/show_bug.cgi?id=157868

        Reviewed by Benjamin Poulain.

        appendQuotedJSONString was rounding up to the next power of two when resizing
        its buffer. Lets call the allocation size X. If X > 2^31, then
        roundUpToPowerOfTwo(X) == 0. This patch fixes this by making the
        assertion reflect what the code is doing. We now allocate to a size
        of X = std::max(maximumCapacityRequired , roundUpToPowerOfTwo(maximumCapacityRequired))

        * wtf/text/StringBuilder.cpp:
        (WTF::StringBuilder::appendQuotedJSONString):

2016-05-17  Joseph Pecoraro  <pecoraro@apple.com>

        REGRESSION(r192855): Math.random() always produces the same first 7 decimal points the first two invocations
        https://bugs.webkit.org/show_bug.cgi?id=157805
        <rdar://problem/26327851>

        Reviewed by Geoffrey Garen.

        * wtf/WeakRandom.h:
        (WTF::WeakRandom::setSeed):
        Advance once to randomize the 32bit seed across the 128bit state
        and avoid re-using 64bits of state in the second advance.

2016-05-17  Dean Jackson  <dino@apple.com>

        Remove ES6_GENERATORS flag
        https://bugs.webkit.org/show_bug.cgi?id=157815
        <rdar://problem/26332894>

        Reviewed by Geoffrey Garen.

        This flag isn't needed. Generators are enabled everywhere and
        part of a stable specification.

        * wtf/FeatureDefines.h:

2016-05-17  Filip Pizlo  <fpizlo@apple.com>

        WTF should know about Language
        https://bugs.webkit.org/show_bug.cgi?id=157756

        Reviewed by Geoffrey Garen.

        This contains two changes:

        - Move everything that WebCore's logic for getting the platform user preferred language
          depended on into WTF. This means CFBundleSPI.h and BlockObjCExceptions.h|cpp.

        - Move WebCore::platformUserPreferredLanguages() to WTF::platformUserPreferredLanguages().
          This is needed by https://bugs.webkit.org/show_bug.cgi?id=157755, which will make JSC
          use this to detect the platform user preferred language when running standalone.

        Moving the dependencies accounts for a huge chunk of this change, since we have to rewire
        all of the references to those headers in all of WebKit.

        Moving platformUserPreferredLanguages() is mostly easy except for the weird callback.
        That function would call languageDidChange(), which needs to stay in WebCore. So, this
        gives WebCore the ability to register a languageDidChange callback. Other than this new
        logic, the code being added to WTF is just being lifted out of WebCore.

        * WTF.xcodeproj/project.pbxproj:
        * wtf/BlockObjCExceptions.h: Added.
        * wtf/BlockObjCExceptions.mm: Added.
        (ReportBlockedObjCException):
        * wtf/PlatformEfl.cmake:
        * wtf/PlatformGTK.cmake:
        * wtf/PlatformJSCOnly.cmake:
        * wtf/PlatformMac.cmake:
        * wtf/PlatformUserPreferredLanguages.h: Added.
        * wtf/PlatformUserPreferredLanguagesMac.mm: Added.
        (WTF::setPlatformUserPreferredLanguagesChangedCallback):
        (WTF::preferredLanguages):
        (+[WTFLanguageChangeObserver languagePreferencesDidChange:]):
        (WTF::httpStyleLanguageCode):
        (WTF::isValidICUCountryCode):
        (WTF::platformUserPreferredLanguages):
        * wtf/PlatformUserPreferredLanguagesUnix.cpp: Added.
        (WTF::setPlatformUserPreferredLanguagesChangedCallback):
        (WTF::platformLanguage):
        (WTF::platformUserPreferredLanguages):
        * wtf/PlatformUserPreferredLanguagesWin.cpp: Added.
        (WTF::setPlatformUserPreferredLanguagesChangedCallback):
        (WTF::localeInfo):
        (WTF::platformLanguage):
        (WTF::platformUserPreferredLanguages):
        * wtf/PlatformWin.cmake:
        * wtf/spi/cf: Added.
        * wtf/spi/cf/CFBundleSPI.h: Added.

2016-05-17  Youenn Fablet  <youenn.fablet@crf.canon.fr>

        DOMPromise should only restrict the resolution type
        https://bugs.webkit.org/show_bug.cgi?id=157307

        Reviewed by Darin Adler.

        * wtf/Ref.h: Adding static constexpr to ease detection of Ref for templates.
        * wtf/RefPtr.h: Ditto.

2016-05-16  Michael Saboff  <msaboff@apple.com>

        ARMV7K: Crash at JavaScriptCore: WTF::ScopedLambdaFunctor<bool
        https://bugs.webkit.org/show_bug.cgi?id=157781

        Reviewed by Filip Pizlo.

        Replaced use of ScopedLambda in locking code with std::function much as it was
        before change set 199760 to work around what appears to be a clang compiler issue.

        * wtf/ParkingLot.cpp:
        (WTF::ParkingLot::parkConditionallyImpl):
        (WTF::ParkingLot::unparkOne):
        (WTF::ParkingLot::unparkAll):
        (WTF::ParkingLot::forEach):
        (WTF::ParkingLot::unparkOneImpl): Deleted.
        (WTF::ParkingLot::forEachImpl): Deleted.
        * wtf/ParkingLot.h:
        (WTF::ParkingLot::parkConditionally):
        (WTF::ParkingLot::unparkOne): Deleted.
        (WTF::ParkingLot::forEach): Deleted.

2016-05-15  Chris Dumez  <cdumez@apple.com>

        Use more references in JS wrappers related code
        https://bugs.webkit.org/show_bug.cgi?id=157721

        Reviewed by Darin Adler.

        Add new static_reference_cast() overload that takes a Ref<U>&& in
        in order to downcast Ref<> types without causing ref-counting
        churn.

        * wtf/Ref.h:
        (WTF::static_reference_cast):

2016-05-13  Chris Dumez  <cdumez@apple.com>

        Unreviewed, rolling out r200837.

        Seems to have regressed Speedometer and JetStream on iOS

        Reverted changeset:

        "DOMPromise should only restrict the resolution type"
        https://bugs.webkit.org/show_bug.cgi?id=157307
        http://trac.webkit.org/changeset/200837

2016-05-13  Beth Dakin  <bdakin@apple.com>

        Add dyldSPI.h
        https://bugs.webkit.org/show_bug.cgi?id=157401
        -and corresponding-
        rdar://problem/26253396

        Reviewed by Darin Adler.

        New header for dyld spi.
        * WTF.xcodeproj/project.pbxproj:
        * wtf/spi/darwin/dyldSPI.h: Added.

2016-05-13  Youenn Fablet  <youenn.fablet@crf.canon.fr>

        DOMPromise should only restrict the resolution type
        https://bugs.webkit.org/show_bug.cgi?id=157307

        Reviewed by Darin Adler.

        * wtf/Ref.h: Adding static constexpr to ease detection of Ref for templates.
        * wtf/RefPtr.h: Ditto.

2016-05-12  Daniel Bates  <dabates@apple.com>

        Use SecTask SPI to retrieve code signing identifier for user directory suffix
        https://bugs.webkit.org/show_bug.cgi?id=157570

        Reviewed by Darin Adler.
        <rdar://problem/25706517>

        Forward declare SPI SecTaskCopySigningIdentifier().

        * wtf/spi/cocoa/SecuritySPI.h:

2016-05-12  Csaba Osztrogonác  <ossy@webkit.org>

        Remove ENABLE(ES6_ARROWFUNCTION_SYNTAX) guards
        https://bugs.webkit.org/show_bug.cgi?id=157564

        Reviewed by Darin Adler.

        * wtf/FeatureDefines.h:

2016-05-10  Filip Pizlo  <fpizlo@apple.com>

        Internal JSC profiler should have a timestamped log of events for each code block
        https://bugs.webkit.org/show_bug.cgi?id=157538

        Reviewed by Benjamin Poulain.

        * wtf/PrintStream.h:
        (WTF::PrintStream::print):

2016-05-10  Chris Dumez  <cdumez@apple.com>

        Get rid of a lot of calls to RefPtr::release()
        https://bugs.webkit.org/show_bug.cgi?id=157505

        Reviewed by Alex Christensen.

        Get rid of a lot of calls to RefPtr::release() and use WTFMove() instead.

        * wtf/text/AtomicStringImpl.cpp:
        (WTF::HashAndUTF8CharactersTranslator::translate):
        * wtf/text/CString.cpp:
        (WTF::CString::copyBufferIfNeeded):
        * wtf/text/StringBuilder.cpp:
        (WTF::StringBuilder::allocateBuffer):
        (WTF::StringBuilder::allocateBufferUpConvert):
        (WTF::StringBuilder::shrinkToFit):
        * wtf/text/WTFString.cpp:
        (WTF::String::append):
        (WTF::String::insert):
        (WTF::String::removeInternal):
        * wtf/text/WTFString.h:
        (WTF::String::String):
        (WTF::String::operator=):
        (WTF::String::releaseImpl):
        (WTF::String::isNull): Deleted.

2016-05-08  Chris Dumez  <cdumez@apple.com>

        [COCOA] Disable HAVE_DTRACE at build time
        https://bugs.webkit.org/show_bug.cgi?id=157433
        <rdar://problem/26148841>

        Reviewed by Mark Lam.

        Disable HAVE_DTRACE on COCOA since it is mostly unused and has a performance
        impact, at least on iOS.

        * wtf/Platform.h:

2016-05-08  Chris Dumez  <cdumez@apple.com>

        [Bindings] Simplify [RequiresExistingAtomicString] IDL extended attribute handling
        https://bugs.webkit.org/show_bug.cgi?id=157465

        Reviewed by Darin Adler.

        Add an AtomicString(RefPtr<AtomicStringImpl>&&) constructor that is now
        used by the bindings when [RequiresExistingAtomicString] is used. This
        is more efficient than using AtomicString(AtomicStringImpl*) as the
        caller has a RefPtr<AtomicStringImpl> and is consistent with the
        pre-existing String(RefPtr<AtomicStringImpl>&&) constructor.

        * wtf/text/AtomicString.h:
        (WTF::AtomicString::AtomicString):

2016-05-08  Darin Adler  <darin@apple.com>

        * wtf/text/WTFString.h: Remove pragma once. Not working for some reason.

2016-05-08  Darin Adler  <darin@apple.com>

        Follow-up to that last patch (missed review comments).

        * wtf/text/WTFString.h: Use nullptr instead of 0 as needed. Also use pragma once.

2016-05-08  Darin Adler  <darin@apple.com>

        Change EventSource constructor to take an IDL dictionary instead of a WebCore::Dictionary
        https://bugs.webkit.org/show_bug.cgi?id=157459

        Reviewed by Chris Dumez.

        * wtf/text/WTFString.h: Export a symbol now used in WebCore.

2016-05-02  Sam Weinig  <sam@webkit.org>

        On platforms that support it, use a SecTrustRef as the basis of CertificateInfo instead of a chain of SecCertificateRefs.
        https://bugs.webkit.org/show_bug.cgi?id=157220

        Reviewed by Darin Adler.

        * wtf/Platform.h:
        Add support for HAVE(SEC_TRUST_SERIALIZATION).

        * wtf/spi/cocoa/SecuritySPI.h:
        Add SPI access to SecTrustSerialize and SecTrustDeserialize.

2016-05-04  Filip Pizlo  <fpizlo@apple.com>

        Add HLE locks and synchronic TTAS locks to the ToyLocks benchmark suite
        https://bugs.webkit.org/show_bug.cgi?id=157367

        Rubber stamped by Benjamin Poulain.
        
        Turns out they are both a lot slower than WTF::Lock. The HLE lock is written according to Intel's
        documentation. The synchronic lock follows the algorithm of the "ttas_lock" in the synchronic
        test.hpp.

        * benchmarks/ToyLocks.h:

2016-05-04  Filip Pizlo  <fpizlo@apple.com>

        Add a few more WTF locking benchmarks
        https://bugs.webkit.org/show_bug.cgi?id=157365

        Rubber stamped by Benjamin Poulain.
        
        Factors out our toy locks into ToyLocks.h and adds a new test (LockFairnessTest) that measures
        the fairness of locks.
        
        The result: WTF::Lock is pretty fair enough.

        * benchmarks/LockFairnessTest.cpp: Added.
        (main):
        * benchmarks/LockSpeedTest.cpp:
        (main):
        * benchmarks/ToyLocks.h: Added.

2016-05-04  Filip Pizlo  <fpizlo@apple.com>

        Speed up JSGlobalObject initialization by making some properties lazy
        https://bugs.webkit.org/show_bug.cgi?id=157045

        Reviewed by Keith Miller.
        
        This WTF change is at the heart of a large JSC change. In JSC I found myself wanting to
        do this a lot:
        
            static void callback(Foo& foo) { ... }
        
            foo.setCallback(callback);
        
        But that's not very nice to write if many different setCallback() calls are inside of the
        same very large function: you'll have to have a lot of static function definitions in
        one part of the file, and then a bunch of setCallback() calls in another part. It's hard
        to reason about what's going on with such code.
        
        So what if you wrote this instead:
        
            foo.setCallback([] (Foo& foo) { ... });
        
        Much nicer! There is a standard way to do this: lambdas that are stateless are
        convertible to function pointers. This change also offers another approach that is a bit
        more general.
        
        These additions to WTF help you do it:
        
        isStatelessLambda<Func>(): tells you if Func is a stateless lambda. This uses is_empty to
        test if the lambda is stateless. This turns out to be a stronger property than
        convertibility to function pointers. For example, a function pointer is convertible to a
        function pointer, but it is definitely stateful: you cannot successfully call it if you
        only has its type. On the other hand, a stateless lambda is really stateless in the sense
        that you only need its type to call it.
        
        callStatelessLambda<ResultType, Func>(Arguments&&...): calls the given stateless lambda.
        
        JSC uses these to build up some sophisticated lazy-initialization APIs. The use of
        statelessness allows JSC to combine a lambda with other logic into a single function
        pointer.

        * wtf/StdLibExtras.h:
        (WTF::isStatelessLambda):
        (WTF::callStatelessLambda):

2016-05-04  Per Arne Vollan  <peavo@outlook.com>

        [Win] Use NeverDestroyed template.
        https://bugs.webkit.org/show_bug.cgi?id=157345

        Reviewed by Darin Adler.

        * wtf/ThreadingWin.cpp:
        (WTF::threadMapMutex):
        (WTF::threadMap):

2016-05-04  Chris Dumez  <cdumez@apple.com>

        Unreviewed, rolling out r200383 and r200406.

        Seems to have caused crashes on iOS / ARMv7s

        Reverted changesets:

        "Speed up JSGlobalObject initialization by making some
        properties lazy"
        https://bugs.webkit.org/show_bug.cgi?id=157045
        http://trac.webkit.org/changeset/200383

        "REGRESSION(r200383): Setting lazily initialized properties
        across frame boundaries crashes"
        https://bugs.webkit.org/show_bug.cgi?id=157333
        http://trac.webkit.org/changeset/200406

2016-05-03  Filip Pizlo  <fpizlo@apple.com>

        Speed up JSGlobalObject initialization by making some properties lazy
        https://bugs.webkit.org/show_bug.cgi?id=157045

        Reviewed by Keith Miller.
        
        This WTF change is at the heart of a large JSC change. In JSC I found myself wanting to
        do this a lot:
        
            static void callback(Foo& foo) { ... }
        
            foo.setCallback(callback);
        
        But that's not very nice to write if many different setCallback() calls are inside of the
        same very large function: you'll have to have a lot of static function definitions in
        one part of the file, and then a bunch of setCallback() calls in another part. It's hard
        to reason about what's going on with such code.
        
        So what if you wrote this instead:
        
            foo.setCallback([] (Foo& foo) { ... });
        
        Much nicer! There is a standard way to do this: lambdas that are stateless are
        convertible to function pointers. This change also offers another approach that is a bit
        more general.
        
        These additions to WTF help you do it:
        
        isStatelessLambda<Func>(): tells you if Func is a stateless lambda. This uses is_empty to
        test if the lambda is stateless. This turns out to be a stronger property than
        convertibility to function pointers. For example, a function pointer is convertible to a
        function pointer, but it is definitely stateful: you cannot successfully call it if you
        only has its type. On the other hand, a stateless lambda is really stateless in the sense
        that you only need its type to call it.
        
        callStatelessLambda<ResultType, Func>(Arguments&&...): calls the given stateless lambda.
        
        JSC uses these to build up some sophisticated lazy-initialization APIs. The use of
        statelessness allows JSC to combine a lambda with other logic into a single function
        pointer.

        * wtf/StdLibExtras.h:
        (WTF::isStatelessLambda):
        (WTF::callStatelessLambda):

2016-05-03  Per Arne Vollan  <peavo@outlook.com>

        [Win] Remove Windows XP Compatibility Requirements
        https://bugs.webkit.org/show_bug.cgi?id=152899

        Reviewed by Brent Fulgham.

        Windows XP is not supported anymore, we can remove workarounds.

        * wtf/Assertions.cpp:

2016-05-03  Per Arne Vollan  <peavo@outlook.com>

        [Win] Warning fix.
        https://bugs.webkit.org/show_bug.cgi?id=157309

        Reviewed by Darin Adler.

        MSVC gives a warning when converting from signed to unsigned.

        * wtf/SmallPtrSet.h:
        (WTF::SmallPtrSet::begin):

2016-05-02  Brady Eidson  <beidson@apple.com>

        Add the ability to accumulate logs for specific logging channels to help diagnose test timeouts.
        https://bugs.webkit.org/show_bug.cgi?id=157274

        Reviewed by Alex Christensen.

        This patch changes WTFLog to add the logging string to the logging accumulator if the logging channel says so.
        It also adds support for using this new accumulation mechanism.

        * WTF.xcodeproj/project.pbxproj:

        * wtf/Assertions.cpp:
        (WTF::resetAccumulatedLogs):
        (WTF::getAndResetAccumulatedLogs):
        * wtf/Assertions.h:

        * wtf/LoggingAccumulator.h: Added. Expose functions to get accumulated logs and to reset the accumulation.

        * wtf/text/WTFString.cpp:
        (WTF::String::format): Update to handle %@ on CF (Well, really ObjC) platforms.

2016-05-02  Per Arne Vollan  <peavo@outlook.com>

        [Win] Enable IndexedDB.
        https://bugs.webkit.org/show_bug.cgi?id=157192

        Reviewed by Brent Fulgham.

        Fix compile errors when format specifiers like PRIu64 is used, by defining
        __STDC_FORMAT_MACROS before inttypes.h is included.

        * wtf/Platform.h: 

2016-04-27  Brady Eidson  <beidson@apple.com>

        Modern IDB: Implement native IDBFactory.getAllDatabaseNames for WebInspector.
        https://bugs.webkit.org/show_bug.cgi?id=157072

        Reviewed by Alex Christensen.

        Moved these Hex Digit utilities from WebCore URL code (???), 
        and add a checked version of getting the hex digit value.
        
        * wtf/HexNumber.h:
        (WTF::isHexDigit):
        (WTF::uncheckedHexDigit):
        (WTF::hexDigitValue):
        (WTF::uncheckedHexDigitValue):

2016-04-25  Ryosuke Niwa  <rniwa@webkit.org>

        Remove the build flag for template elements
        https://bugs.webkit.org/show_bug.cgi?id=157022

        Reviewed by Daniel Bates.

        * wtf/FeatureDefines.h:

2016-04-25  Fujii Hironori  <Hironori.Fujii@sony.com>

        Heap corruption is detected when destructing JSGlobalObject
        https://bugs.webkit.org/show_bug.cgi?id=156831

        Reviewed by Mark Lam.

        WebKit uses CRT static library on Windows.  Each copy of the CRT
        library has its own heap manager, allocating memory in one CRT
        library and passing the pointer across a DLL boundary to be freed
        by a different copy of the CRT library is a potential cause for
        heap corruption.

          Potential Errors Passing CRT Objects Across DLL Boundaries
          <https://msdn.microsoft.com/en-us/library/ms235460(v=vs.140).aspx>

        JSGlobalObject::createRareDataIfNeeded is inlined but
        JSGlobalObject::~JSGlobalObject is not.  Then, the heap of
        allocating JSGlobalObjectRareData is WebKit.dll, but deallocating
        JavaScriptCore.dll.  Adding WTF_MAKE_FAST_ALLOCATED to
        JSGlobalObjectRareData ensures heap consistency of it.  WTF::Lock
        also needs WTF_MAKE_FAST_ALLOCATED because it is allocated from
        the inlined constructor of JSGlobalObjectRareData.

        * wtf/Lock.h: Add WTF_MAKE_FAST_ALLOCATED.

2016-04-21  Saam barati  <sbarati@apple.com>

        Lets do less locking of symbol tables in the BytecodeGenerator where we don't have race conditions
        https://bugs.webkit.org/show_bug.cgi?id=156821

        Reviewed by Filip Pizlo.

        This patch introduces a new constructor for Locker which implies no
        locking is necessary. You instantiate such a locker like so:
        `Locker<Lock> locker(Locker<Lock>::NoLockingNecessary);`

        This is useful to for very specific places when it is not yet
        required to engage in a specified locking protocol. As an example,
        we use this in JSC when we allocate a particular object that
        engages in a locking protocol with the concurrent compiler thread,
        but before a concurrent compiler thread that could have access
        to the object exists.

        * wtf/Locker.h:
        (WTF::Locker::Locker):
        (WTF::Locker::~Locker):

2016-04-19  Michael Saboff  <msaboff@apple.com>

        iTunes crashing JavaScriptCore.dll
        https://bugs.webkit.org/show_bug.cgi?id=156647

        Reviewed by Filip Pizlo.

        If a thread was created without using the WTF thread apis and that thread uses
        a JavaScript VM and that thread exits with the VM still around, JSC won't know
        that the thread has exited.  Currently, we use ThreadSpecificThreadExit() to
        clean up any thread specific keys.  Cleaning up these keys is how JSC is
        notified of a thread exit.  We only call ThreadSpecificThreadExit() from
        wtfThreadEntryPoint() when the thread entry point function returns.
        This mechanism was put in place for Windows because we layer the WTF::ThreadSpecific
        functionality on top of TLS (Thread Local Storage), but TLS doesn't have
        a thread exiting callback the way that pthread_create_key does.

        The fix is to change from using TLS to using FLS (Fiber Local Storage).  Although
        Windows allows multiple fibers per thread, WebKit is not designed to work with a
        multiple fibers per thread.  When there is only one fiber per thread, FLS works just
        like TLS, but it has the destroy callback.

        I restructured the Windows version of WTF::ThreadSpecific to be almost the same
        as the pthread version.  Added THREAD_SPECIFIC_CALL to set the correct
        calling convenction for Windows 32 bit.

        * wtf/ThreadSpecific.h:
        (WTF::threadSpecificKeyCreate):
        (WTF::threadSpecificKeyDelete):
        (WTF::threadSpecificSet):
        (WTF::threadSpecificGet):
        (WTF::ThreadSpecific<T>::ThreadSpecific):
        (WTF::ThreadSpecific<T>::~ThreadSpecific):
        (WTF::ThreadSpecific<T>::get):
        (WTF::ThreadSpecific<T>::set):
        (WTF::ThreadSpecific<T>::destroy):
        Restructured to use FLS.  Renamed TLS* to FLS*.

        * wtf/ThreadSpecificWin.cpp:
        (WTF::flsKeyCount):
        (WTF::flsKeys):
        Renamed from tlsKey*() to flsKey*().

        (WTF::destructorsList): Deleted.
        (WTF::destructorsMutex): Deleted.
        (WTF::PlatformThreadSpecificKey::PlatformThreadSpecificKey): Deleted.
        (WTF::PlatformThreadSpecificKey::~PlatformThreadSpecificKey): Deleted.
        (WTF::PlatformThreadSpecificKey::setValue): Deleted.
        (WTF::PlatformThreadSpecificKey::value): Deleted.
        (WTF::PlatformThreadSpecificKey::callDestructor): Deleted.
        (WTF::tlsKeyCount): Deleted.
        (WTF::tlsKeys): Deleted.
        (WTF::threadSpecificKeyCreate): Deleted.
        (WTF::threadSpecificKeyDelete): Deleted.
        (WTF::threadSpecificSet): Deleted.
        (WTF::threadSpecificGet): Deleted.
        (WTF::ThreadSpecificThreadExit): Deleted.

        * wtf/ThreadingWin.cpp:
        (WTF::wtfThreadEntryPoint): Eliminated call to ThreadSpecificThreadExit.

2016-04-19  Filip Pizlo  <fpizlo@apple.com>

        Add more locking algorithms to the LockSpeedTest.

        Rubber stamped by Saam Barati.

        * benchmarks/LockSpeedTest.cpp:
        (main):

2016-04-19  Filip Pizlo  <fpizlo@apple.com>

        Clean up the ParkingLot uparking API a bit
        https://bugs.webkit.org/show_bug.cgi?id=156746

        Reviewed by Saam Barati and Geoffrey Garen.
        
        Previously, unparkOne() would either return a boolean to tell you if there are any more threads on
        the queue or it would pass your callback a pair of booleans - one to tell if a thread was unparked
        and another to tell if there are any more threads. This was an annoying inconsistency. What if you
        wanted to know if unparkOne() unparked a thread but you don't care to use callbacks?

        This fixes unparkOne() to use a struct called UnparkResult for both of its variants. This makes the
        code a bit cleaner.

        * wtf/Atomics.h: Add some more atomic ops.
        (WTF::Atomic::exchangeAndAdd):
        (WTF::Atomic::exchange):
        * wtf/Condition.h: Change calls to unparkOne().
        (WTF::ConditionBase::notifyOne):
        * wtf/Lock.cpp: Change calls to unparkOne().
        (WTF::LockBase::unlockSlow):
        * wtf/ParkingLot.cpp:
        (WTF::ParkingLot::parkConditionally):
        (WTF::ParkingLot::unparkOne):
        * wtf/ParkingLot.h: Switch to using ScopedLambda and introduce UnparkResult.

2016-04-19  Commit Queue  <commit-queue@webkit.org>

        Unreviewed, rolling out r199726.
        https://bugs.webkit.org/show_bug.cgi?id=156748

        WebKit tests crash on Windows 32 (Requested by msaboff on
        #webkit).

        Reverted changeset:

        "iTunes crashing JavaScriptCore.dll"
        https://bugs.webkit.org/show_bug.cgi?id=156647
        http://trac.webkit.org/changeset/199726

2016-04-19  Michael Saboff  <msaboff@apple.com>

        iTunes crashing JavaScriptCore.dll
        https://bugs.webkit.org/show_bug.cgi?id=156647

        Reviewed by Saam Barati.

        If a thread was created without using the WTF thread apis and that thread uses
        a JavaScript VM and that thread exits with the VM still around, JSC won't know
        that the thread has exited.  Currently, we use ThreadSpecificThreadExit() to
        clean up any thread specific keys.  Cleaning up these keys is how JSC is
        notified of a thread exit.  We only call ThreadSpecificThreadExit() from
        wtfThreadEntryPoint() when the thread entry point function returns.
        This mechanism was put in place for Windows because we layer the WTF::ThreadSpecific
        functionality on top of TLS (Thread Local Storage), but TLS doesn't have
        a thread exiting callback the way that pthread_create_key.

        The fix is to change from using TLS to using FLS (Fiber Local Storage).  Although
        Windows allows multiple fibers per thread, WebKit is not designed to work with a
        multiple fibers per thread.  When there is only one fiber per thread, FLS works just
        like TLS, but it has the destroy callback.

        I restructured the Windows version of WTF::ThreadSpecific to be almost the same
        as the pthread version.

        * wtf/ThreadSpecific.h:
        (WTF::threadSpecificKeyCreate):
        (WTF::threadSpecificKeyDelete):
        (WTF::threadSpecificSet):
        (WTF::threadSpecificGet):
        (WTF::ThreadSpecific<T>::ThreadSpecific):
        (WTF::ThreadSpecific<T>::~ThreadSpecific):
        (WTF::ThreadSpecific<T>::get):
        (WTF::ThreadSpecific<T>::set):
        (WTF::ThreadSpecific<T>::destroy):
        Restructured to use FLS.  Renamed TLS* to FLS*.

        * wtf/ThreadSpecificWin.cpp:
        (WTF::flsKeyCount):
        (WTF::flsKeys):
        Renamed from tlsKey*() to flsKey*().

        (WTF::destructorsList): Deleted.
        (WTF::destructorsMutex): Deleted.
        (WTF::PlatformThreadSpecificKey::PlatformThreadSpecificKey): Deleted.
        (WTF::PlatformThreadSpecificKey::~PlatformThreadSpecificKey): Deleted.
        (WTF::PlatformThreadSpecificKey::setValue): Deleted.
        (WTF::PlatformThreadSpecificKey::value): Deleted.
        (WTF::PlatformThreadSpecificKey::callDestructor): Deleted.
        (WTF::tlsKeyCount): Deleted.
        (WTF::tlsKeys): Deleted.
        (WTF::threadSpecificKeyCreate): Deleted.
        (WTF::threadSpecificKeyDelete): Deleted.
        (WTF::threadSpecificSet): Deleted.
        (WTF::threadSpecificGet): Deleted.
        (WTF::ThreadSpecificThreadExit): Deleted.

        * wtf/ThreadingWin.cpp:
        (WTF::wtfThreadEntryPoint): Eliminated call to ThreadSpecificThreadExit.

2016-04-19  Yusuke Suzuki  <utatane.tea@gmail.com>

        [GTK] Use Generic WorkQueue instead of WorkQueueGLib
        https://bugs.webkit.org/show_bug.cgi?id=156721

        Reviewed by Benjamin Poulain.

        WorkQueueGLib includes platform-dependent part only for WorkQueue::dispatchAfter.
        However, its code is related to RunLoopGLib rather than WorkQueueGLib.
        In this patch, we move the code from WorkQueueGLib to RunLoopGLib, drop WorkQueueGLib
        and use WorkQueueGeneric.

        * wtf/PlatformGTK.cmake:
        * wtf/RunLoop.h:
        * wtf/WorkQueue.h:
        * wtf/glib/RunLoopGLib.cpp:
        (WTF::DispatchAfterContext::DispatchAfterContext):
        (WTF::DispatchAfterContext::dispatch):
        (WTF::RunLoop::dispatchAfter):
        * wtf/glib/WorkQueueGLib.cpp: Removed.
        (WTF::WorkQueue::platformInitialize): Deleted.
        (WTF::WorkQueue::platformInvalidate): Deleted.
        (WTF::WorkQueue::dispatch): Deleted.
        (WTF::DispatchAfterContext::DispatchAfterContext): Deleted.
        (WTF::DispatchAfterContext::~DispatchAfterContext): Deleted.
        (WTF::DispatchAfterContext::dispatch): Deleted.
        (WTF::WorkQueue::dispatchAfter): Deleted.

2016-04-18  Yusuke Suzuki  <utatane.tea@gmail.com>

        [JSCOnly] Implement RunLoop and remove glib dependency
        https://bugs.webkit.org/show_bug.cgi?id=155706

        Reviewed by Michael Catanzaro.

        Add missing RunLoop and WorkQueue platform code for JSCOnly port.
        The implementation does not use platform specific things. Instead, we
        implement them on WTF framework; using condition variables to construct
        the run loop and timers.

        Since the fallback is implemented, it is not necessary for JSCOnly port
        to depend on GLib's RunLoop abstraction. So this patch removes its
        dependency. As a result, now, JSCOnly port removes dependencies except for the system ICU.

        We clean up event loop ifdefs by introducing USE_XXX_EVENT_LOOP flags.
        USE(XXX_EVENT_LOOP) is exclusive to each other. So from now, we do not need to take care
        of the order of if-defs for the event loops. (For example, USE(GLIB) should have lead before
        OS(DARWIN) in WorkQueue.h for WebKitGTK on Darwin.)

        EVENT_LOOP determination is done in Platform.h. This follows the style of WTF PLATFORM.

        * wtf/Platform.h:
        * wtf/PlatformJSCOnly.cmake:
        * wtf/RunLoop.h:
        * wtf/WorkQueue.h:
        * wtf/generic/MainThreadGeneric.cpp: Renamed from Source/WTF/wtf/none/MainThreadNone.cpp.
        (WTF::initializeMainThreadPlatform):
        (WTF::scheduleDispatchFunctionsOnMainThread):
        * wtf/generic/RunLoopGeneric.cpp: Added.
        (WTF::RunLoop::TimerBase::ScheduledTask::create):
        (WTF::RunLoop::TimerBase::ScheduledTask::ScheduledTask):
        (WTF::RunLoop::TimerBase::ScheduledTask::fired):
        (WTF::RunLoop::TimerBase::ScheduledTask::scheduledTimePoint):
        (WTF::RunLoop::TimerBase::ScheduledTask::updateReadyTime):
        (WTF::RunLoop::TimerBase::ScheduledTask::EarliestSchedule::operator()):
        (WTF::RunLoop::TimerBase::ScheduledTask::isActive):
        (WTF::RunLoop::TimerBase::ScheduledTask::deactivate):
        (WTF::RunLoop::RunLoop):
        (WTF::RunLoop::~RunLoop):
        (WTF::RunLoop::populateTasks):
        (WTF::RunLoop::runImpl):
        (WTF::RunLoop::run):
        (WTF::RunLoop::iterate):
        (WTF::RunLoop::stop):
        (WTF::RunLoop::wakeUp):
        (WTF::RunLoop::schedule):
        (WTF::RunLoop::scheduleAndWakeUp):
        (WTF::RunLoop::dispatchAfter):
        (WTF::RunLoop::TimerBase::TimerBase):
        (WTF::RunLoop::TimerBase::~TimerBase):
        (WTF::RunLoop::TimerBase::start):
        (WTF::RunLoop::TimerBase::stop):
        (WTF::RunLoop::TimerBase::isActive):
        * wtf/generic/WorkQueueGeneric.cpp: Added.
        (WorkQueue::platformInitialize):
        (WorkQueue::platformInvalidate):
        (WorkQueue::dispatch):
        (WorkQueue::dispatchAfter):
        * wtf/none/MainThreadNone.cpp:
        (WTF::initializeMainThreadPlatform): Deleted.
        (WTF::scheduleDispatchFunctionsOnMainThread): Deleted.
        * wtf/none/RunLoopNone.cpp:
        (WTF::RunLoop::RunLoop): Deleted.
        (WTF::RunLoop::~RunLoop): Deleted.
        (WTF::RunLoop::run): Deleted.
        (WTF::RunLoop::stop): Deleted.
        (WTF::RunLoop::wakeUp): Deleted.
        (WTF::RunLoop::TimerBase::TimerBase): Deleted.
        (WTF::RunLoop::TimerBase::~TimerBase): Deleted.
        (WTF::RunLoop::TimerBase::start): Deleted.
        (WTF::RunLoop::TimerBase::stop): Deleted.
        (WTF::RunLoop::TimerBase::isActive): Deleted.
        * wtf/none/WorkQueueNone.cpp:
        (WorkQueue::platformInitialize): Deleted.
        (WorkQueue::platformInvalidate): Deleted.
        (WorkQueue::dispatch): Deleted.
        (WorkQueue::dispatchAfter): Deleted.

2016-04-18  Commit Queue  <commit-queue@webkit.org>

        Unreviewed, rolling out r199621.
        https://bugs.webkit.org/show_bug.cgi?id=156708

        made windows tests crash (Requested by alexchristensen on
        #webkit).

        Reverted changeset:

        "iTunes crashing JavaScriptCore.dll"
        https://bugs.webkit.org/show_bug.cgi?id=156647
        http://trac.webkit.org/changeset/199621

2016-04-15  Michael Saboff  <msaboff@apple.com>

        iTunes crashing JavaScriptCore.dll
        https://bugs.webkit.org/show_bug.cgi?id=156647

        Reviewed by Geoffrey Garen.

        If a thread was created without using the WTF thread apis and that thread uses
        a JavaScript VM and that thread exits with the VM still around, JSC won't know
        that the thread has exited.  Currently, we use ThreadSpecificThreadExit() to
        clean up any thread specific keys.  Cleaning up these keys is how JSC is
        notified of a thread exit.  We only call ThreadSpecificThreadExit() from
        wtfThreadEntryPoint() when the thread entry point function returns.
        This mechanism was put in place for Windos because we layer the WTF::ThreadSpecific
        functionality on top of TLS (Thread Local Storage), but TLS doesn't have
        a thread exiting callback the way that pthread_create_key.

        The fix is to change from using TLS to using FLS (Fiber Local Storage).  Although
        Windows allows multiple fibers per thread, WebKit is not designed to work with a
        multiple fibers per thread.  When ther is only one fiber per thread, FLS works just
        like TLS, but it has the destroy callback.

        I restructured the Windows version of WTF::ThreadSpecific to be almost the same
        as the pthread version.
        
        * wtf/ThreadSpecific.h:
        (WTF::threadSpecificKeyCreate):
        (WTF::threadSpecificKeyDelete):
        (WTF::threadSpecificSet):
        (WTF::threadSpecificGet):
        (WTF::ThreadSpecific<T>::ThreadSpecific):
        (WTF::ThreadSpecific<T>::~ThreadSpecific):
        (WTF::ThreadSpecific<T>::get):
        (WTF::ThreadSpecific<T>::set):
        (WTF::ThreadSpecific<T>::destroy):
        Restructured to use FLS.  Renamed TLS* to FLS*.

        * wtf/ThreadSpecificWin.cpp:
        (WTF::flsKeyCount):
        (WTF::flsKeys):
        Renamed from tlsKey*() to flsKey*().

        (WTF::destructorsList): Deleted.
        (WTF::destructorsMutex): Deleted.
        (WTF::PlatformThreadSpecificKey::PlatformThreadSpecificKey): Deleted.
        (WTF::PlatformThreadSpecificKey::~PlatformThreadSpecificKey): Deleted.
        (WTF::PlatformThreadSpecificKey::setValue): Deleted.
        (WTF::PlatformThreadSpecificKey::value): Deleted.
        (WTF::PlatformThreadSpecificKey::callDestructor): Deleted.
        (WTF::tlsKeyCount): Deleted.
        (WTF::tlsKeys): Deleted.
        (WTF::threadSpecificKeyCreate): Deleted.
        (WTF::threadSpecificKeyDelete): Deleted.
        (WTF::threadSpecificSet): Deleted.
        (WTF::threadSpecificGet): Deleted.
        (WTF::ThreadSpecificThreadExit): Deleted.

        * wtf/ThreadingWin.cpp:
        (WTF::wtfThreadEntryPoint): Eliminated call to ThreadSpecificThreadExit.

2016-04-12  Filip Pizlo  <fpizlo@apple.com>

        PolymorphicAccess should buffer AccessCases before regenerating
        https://bugs.webkit.org/show_bug.cgi?id=156457

        Reviewed by Benjamin Poulain.

        * wtf/TinyPtrSet.h:
        (WTF::TinyPtrSet::add): Add a helpful comment because I had forgotten what the bool return meant.

2016-04-12  Tomas Popela  <tpopela@redhat.com>

        S390X and PPC64 architectures detection is wrong
        https://bugs.webkit.org/show_bug.cgi?id=156337

        Reviewed by Carlos Garcia Campos.

        After the http://trac.webkit.org/changeset/198919 was committed
        it showed that the PPC64 detection is wrong as the CPU(PPC) path was
        activated even for PPC64. The thing is that GCC defines __ppc__
        even on PPC64 and not just on PPC(32). The same applies for S390X.

        * wtf/Platform.h:

2016-04-05  Oliver Hunt  <oliver@apple.com>

        Remove compile time define for SEPARATED_HEAP
        https://bugs.webkit.org/show_bug.cgi?id=155508

        Reviewed by Mark Lam.

        * wtf/FeatureDefines.h:
        * wtf/Platform.h:

2016-04-11  Fujii Hironori  <Hironori.Fujii@jp.sony.com>

        [CMake] Make FOLDER property INHERITED
        https://bugs.webkit.org/show_bug.cgi?id=156460

        Reviewed by Brent Fulgham.

        * CMakeLists.txt:
        Set FOLDER directory property

2016-04-08  Alex Christensen  <achristensen@webkit.org>

        Progress towards running CMake WebKit2 on Mac
        https://bugs.webkit.org/show_bug.cgi?id=156426

        Reviewed by Tim Horton.

        * wtf/PlatformMac.cmake:

2016-04-08  Yusuke Suzuki  <utatane.tea@gmail.com>

        [JSC] Enable Concurrent JIT by default
        https://bugs.webkit.org/show_bug.cgi?id=156341

        Reviewed by Filip Pizlo.

        We enable Concurrent JIT by default when DFG JIT and JSVALUE64 are enabled.
        This change offers Concurrent JIT to the JSCOnly port.

        * wtf/Platform.h:

2016-04-07  Joseph Pecoraro  <pecoraro@apple.com>

        Remove ENABLE(ENABLE_ES6_CLASS_SYNTAX) guards
        https://bugs.webkit.org/show_bug.cgi?id=156384

        Reviewed by Ryosuke Niwa.

        * wtf/FeatureDefines.h:

2016-04-06  Filip Pizlo  <fpizlo@apple.com>

        JSC should have a simple way of gathering IC statistics
        https://bugs.webkit.org/show_bug.cgi?id=156317

        Reviewed by Benjamin Poulain.
        
        Make it easier to do relative sleeping on a condition. Previously you could do this using
        std::chrono. I now believe that std::chrono is just a bad decision, and I always want to
        use doubles instead. This makes it easier to do the right thing and use doubles.

        * wtf/Condition.h:
        (WTF::ConditionBase::waitUntilMonotonicClockSeconds):
        (WTF::ConditionBase::waitForSeconds):

2016-04-06  Simon Fraser  <simon.fraser@apple.com>

        Fix Windows build by converting clampToInteger() into a template that only
        takes integral types.

        * wtf/MathExtras.h:
        (clampToInteger):

2016-04-06  Simon Fraser  <simon.fraser@apple.com>

        Page tiles are missing when graphics acceleration is unavailable
        https://bugs.webkit.org/show_bug.cgi?id=156325

        Reviewed by Tim Horton.

        Add clampToInteger(size_t).

        * wtf/MathExtras.h:
        (clampToInteger):

2016-04-05  Simon Fraser  <simon.fraser@apple.com>

        Implement operator== and operator!= for Optional<>
        https://bugs.webkit.org/show_bug.cgi?id=156266

        Reviewed by Anders Carlsson.
        
        Implement non-member operator== and operator!= variants for Optional<>.

        * wtf/Optional.h:
        (WTF::operator==):
        (WTF::operator!=):

2016-03-19  Filip Pizlo  <fpizlo@apple.com>

        DFG and FTL should constant-fold RegExpExec
        https://bugs.webkit.org/show_bug.cgi?id=155270

        Reviewed by Saam Barati.

        Make executeInsertions() return the amount by which the vector increased in size. This is a
        convenient feature that I use in DFG::InsertionSet.

        * wtf/Insertion.h:
        (WTF::executeInsertions):

2016-04-05  Antoine Quint  <graouts@apple.com>

        [WebGL2] Allow enabling WebGL2 with a runtime flag
        https://bugs.webkit.org/show_bug.cgi?id=156166
        <rdar://problem/25526929>

        Removed the manual overrides of ENABLE_WEBGL2.

        Reviewed by Dean Jackson.

        * wtf/FeatureDefines.h:

2016-03-31  Konstantin Tokarev  <annulen@yandex.ru>

        Removed leftovers of ENABLE(INSPECTOR) and ENABLE(VIEWPORT).
        https://bugs.webkit.org/show_bug.cgi?id=156064

        Reviewed by Csaba Osztrogonác.

        This ifdef guards were removed in r178820 and r147714, respectively.

        * wtf/FeatureDefines.h:

2016-03-31  Brent Fulgham  <bfulgham@apple.com>

        [WK2] Support download attribute feature
        https://bugs.webkit.org/show_bug.cgi?id=102914
        <rdar://problem/13177492>

        Reviewed by Darin Adler.

        * wtf/FeatureDefines.h: Turn the ENABLE_DOWNLOAD_ATTRIBUTE flag
        on to see what happens.

2016-03-30  Brian Burg  <bburg@apple.com>

        Web Automation: Add Automation.performKeyboardInteractions
        https://bugs.webkit.org/show_bug.cgi?id=155990
        <rdar://problem/25426408>

        Reviewed by Timothy Hatcher.

        Add a missing NSEventType declaration.

        * wtf/mac/AppKitCompatibilityDeclarations.h:

2016-03-29  Benjamin Poulain  <bpoulain@apple.com>

        [WTF] Removing a smart pointer from HashTable issues two stores to the same location
        https://bugs.webkit.org/show_bug.cgi?id=155676

        Reviewed by Darin Adler.

        While working on the hot loop of r198376, I noticed something
        weird...
        Every time we removed a smart pointer from the hash table,
        the code generated was something like:
            Load([bucket]) -> Tmp
            Store(0 -> [bucket])
            JumpIfZero(Tmp, ->End)
            Call fastFree()
            Store(-1 -> [bucket])
            -> End:

        The useless store before the branch is annoying, especially on ARM.

        Here is what happens:
            1) The destructor of the smart pointer swaps its internal value with nullptr.
            2) Since the smart pointer is not a local in-register value, that nullptr
               is stored in memory because it could be observable from fastFree().
            3) The destructor destroy the value if not zero (or deref for RefPtr).
               The "if-not-zero" may or may not be eliminated depending on what
               is between getting the iterator and the call to remove().
            4) fastFree() is called.
            5) The deleted value is set in the bucket.

        This patch adds custom deletion for those cases to avoid the useless
        store. The useless null check is still eliminated when we are lucky.

        I went this path instead of changing the destructor of RefPtr for two reasons:
        -I need this to work in unique_ptr for JSC.
        -Nulling the memory may have security advantages in the cases where we do not immediately
         write over that memory again.

        This patch removes 13kb out of x86_64 WebCore.

        * wtf/HashTable.h:
        (WTF::HashTable::deleteBucket):
        (WTF::KeyTraits>::removeIf):
        * wtf/HashTraits.h:
        (WTF::HashTraits<RefPtr<P>>::customDeleteBucket):
        (WTF::hashTraitsDeleteBucket):
        (WTF::KeyValuePairHashTraits::customDeleteBucket):
        * wtf/text/AtomicStringHash.h:
        (WTF::HashTraits<WTF::AtomicString>::isEmptyValue):
        (WTF::HashTraits<WTF::AtomicString>::customDeleteBucket):
        * wtf/text/StringHash.h:
        (WTF::HashTraits<String>::customDeleteBucket):

2016-03-28  Brian Burg  <bburg@apple.com>

        Web Automation: implement Automation.performMouseInteraction
        https://bugs.webkit.org/show_bug.cgi?id=155606
        <rdar://problem/25227576>

        Reviewed by Timothy Hatcher.

        * wtf/mac/AppKitCompatibilityDeclarations.h: Add missing some NSEventTypes.

2016-03-29  Yusuke Suzuki  <utatane.tea@gmail.com>

        REGRESSION(r192914): 10% regression on Sunspider's date-format-tofte
        https://bugs.webkit.org/show_bug.cgi?id=155559

        Reviewed by Saam Barati.

        Add HashTable::inlineLookup and HashMap::fastGet.

        * wtf/HashMap.h:
        * wtf/HashTable.h:

2016-03-25  Alex Christensen  <achristensen@webkit.org>

        Add a compile time flag for using QTKit
        https://bugs.webkit.org/show_bug.cgi?id=155868

        Reviewed by Dan Bates.

        * wtf/Platform.h:

2016-03-24  Alex Christensen  <achristensen@webkit.org>

        Fix iOS9 performance regression after r197572
        https://bugs.webkit.org/show_bug.cgi?id=155845
        <rdar://problem/25144924>

        Reviewed by Chris Dumez.

        * wtf/Platform.h:
        Use CFURLConnection instead of NSURLConnection on internal iOS9 builds.

2016-03-23  Saam Barati  <sbarati@apple.com>

        SmallPtrSet leaks memory in its move assignment operator when !this->isSmall()
        https://bugs.webkit.org/show_bug.cgi?id=155701

        Reviewed by Darin Adler.

        * wtf/SmallPtrSet.h:
        (WTF::SmallPtrSet::SmallPtrSet):
        (WTF::SmallPtrSet::operator=):

2016-03-22  Per Arne Vollan  <peavo@outlook.com>

        [Win] [64-bit] Remove MSVC 2013 FMA3 Bug Workaround
        https://bugs.webkit.org/show_bug.cgi?id=141499

        Reviewed by Brent Fulgham.

        As we have moved on to VS2015, this workaround is no longer needed.

        * wtf/PlatformWin.cmake:
        * wtf/win/WTFDLL.cpp: Removed.

2016-03-20  Dan Bernstein  <mitz@apple.com>

        [Mac] Determine TARGET_MAC_OS_X_VERSION_MAJOR from MACOSX_DEPLOYMENT_TARGET rather than from MAC_OS_X_VERSION_MAJOR
        https://bugs.webkit.org/show_bug.cgi?id=155707
        <rdar://problem/24980691>

        Reviewed by Darin Adler.

        * Configurations/Base.xcconfig: Set TARGET_MAC_OS_X_VERSION_MAJOR based on the last
          component of MACOSX_DEPLOYMENT_TARGET.
        * Configurations/DebugRelease.xcconfig: For engineering builds, preserve the behavior of
          TARGET_MAC_OS_X_VERSION_MAJOR being the host’s OS version.

2016-03-20  Dan Bernstein  <mitz@apple.com>

        Update build settings

        Rubber-stamped by Andy Estes.

        * Configurations/DebugRelease.xcconfig:

2016-03-17  Benjamin Poulain  <bpoulain@apple.com>

        [JSC] Make CSE's ImpureData faster when dealing with large blocks
        https://bugs.webkit.org/show_bug.cgi?id=155594

        Reviewed by Filip Pizlo.

        * wtf/HashSet.h:
        (WTF::V>::removeIf):

2016-03-17  Saam barati  <sbarati@apple.com>

        Implement SmallPtrSet and integrate it into the Parser
        https://bugs.webkit.org/show_bug.cgi?id=155552

        Reviewed by Filip Pizlo.

        This patch implements the SmallPtrSet data struture.
        Inspired by the implementation in llvm:
        http://llvm.org/docs/doxygen/html/SmallPtrSet_8h_source.html

        The data structure uses an inline array for storage up until
        a fixed limit (8 entries in our implementation). If that storage
        fills up, we fall back to a simple hash table implementation.
        Crucially, this implementation doesn't support the remove
        operation. This is on purpose. The hash table will only ever
        grow.

        Also, the implementation allows for it to be memcopied around.
        I.e, we can put SmallPtrSet inside a Vector and allow that
        Vector to use memcpy as its move operation (of course this
        is only valid if the SmallPtrSet in the old memory doesn't have 
        its destructor called unless it is set back to its initial state.)

        For now, SmallPtrSet only supports pointer types that are trivially
        destructible. It's probably not too difficult to extend this to
        smart pointers, but it's not part of this original implementation.

        I've also implemented a pure forwarding varargs constructAndAppend
        method on Vector. This allows you to do:
        Vector<T> v;
        v.constructAndAppend(a1, a2, ...)
        as long as T has a constructor that accepts arguments (a1, a2, ...).

        * WTF.xcodeproj/project.pbxproj:
        * wtf/CMakeLists.txt:
        * wtf/SmallPtrSet.h: Added.
        (WTF::SmallPtrSet::SmallPtrSet):
        (WTF::SmallPtrSet::operator=):
        (WTF::SmallPtrSet::~SmallPtrSet):
        (WTF::SmallPtrSet::add):
        (WTF::SmallPtrSet::contains):
        (WTF::SmallPtrSet::iterator::operator++):
        (WTF::SmallPtrSet::iterator::operator*):
        (WTF::SmallPtrSet::iterator::operator==):
        (WTF::SmallPtrSet::iterator::operator!=):
        (WTF::SmallPtrSet::begin):
        (WTF::SmallPtrSet::end):
        (WTF::SmallPtrSet::size):
        (WTF::SmallPtrSet::emptyValue):
        (WTF::SmallPtrSet::isValidEntry):
        (WTF::SmallPtrSet::isSmall):
        (WTF::SmallPtrSet::initialize):
        (WTF::SmallPtrSet::grow):
        (WTF::SmallPtrSet::bucket):
        * wtf/Vector.h:
        (WTF::Vector::append):
        (WTF::Vector::uncheckedAppend):
        (WTF::minCapacity>::append):
        (WTF::minCapacity>::constructAndAppend):
        (WTF::minCapacity>::appendSlowCase):
        (WTF::minCapacity>::constructAndAppendSlowCase):

2016-03-16  Filip Pizlo  <fpizlo@apple.com>

        Replace all of the various non-working and non-compiling sampling profiler hacks with a single super hack
        https://bugs.webkit.org/show_bug.cgi?id=155561

        Reviewed by Saam Barati.

        This patch replaces all of our various ad hoc profiling hacks with a single ad hoc profiling hack.
        That needs to be able to sleep a thread, so I added a portable way to do it.

        This also removes a bunch of ENABLE flags for all of the old non-working hacks.

        * wtf/CurrentTime.cpp:
        (WTF::currentCPUTime):
        (WTF::sleep):
        * wtf/CurrentTime.h:
        (WTF::sleepMS):
        * wtf/Platform.h:

2016-03-17  Chris Dumez  <cdumez@apple.com>

        Set the WebContent process's main thread QoS to USER-INTERACTIVE
        https://bugs.webkit.org/show_bug.cgi?id=155595
        <rdar://problem/22534965>

        Reviewed by Antti Koivisto.

        Add a relativePriority parameter to setCurrentThreadIsUser*() so that
        we can do more fine-grained prioritization of threads that have the
        same QoS.

        * wtf/Threading.cpp:
        (WTF::setCurrentThreadIsUserInteractive):
        (WTF::setCurrentThreadIsUserInitiated):
        (WTF::createThread): Deleted.
        * wtf/Threading.h:

2016-03-17  Filip Pizlo  <fpizlo@apple.com>

        Silence leaks in ParkingLot
        https://bugs.webkit.org/show_bug.cgi?id=155510

        Reviewed by Alexey Proskuryakov.

        ParkingLot has a concurrent hashtable that it reallocates on demand. It will not reallocate
        it in steady state. The hashtable is sized to accommodate the high watermark of the number
        of active threads - so long as the program doesn't just keep starting an unbounded number
        of threads that are all active, the hashtable will stop resizing. Each resize operation is
        designed to stay out of the way of the data-access-parallel normal path, in which two
        threads operating on different lock addresses don't have to synchronize. To do this, it
        simply drops the old hashtable without deleting it, so that threads that were still using
        it don't crash. They will realize that they have the wrong hashtable before doing anything
        bad, but we don't have a way of proving when all of those threads are no longer going to
        read from the old hashtables. So, we just leak them.

        This is a bounded leak, since the hashtable resizes exponentially. Thus the total memory
        utilization of all hashtables, including the leaked ones, converges to a linear function of
        the current hashtable's size (it's 2 * size of current hashtable).

        But this leak is a problem for leaks tools, which will always report this leak. This is not
        useful. It's better to silence the leak. That's what this patch does by ensuring that all
        hashtables, including leaked ones, end up in a global vector. This is perf-neutral.

        This requires making a StaticWordLock variant of WordLock. That's probably the biggest part
        of this change.

        * wtf/ParkingLot.cpp:
        * wtf/WordLock.cpp:
        (WTF::WordLockBase::lockSlow):
        (WTF::WordLockBase::unlockSlow):
        (WTF::WordLock::lockSlow): Deleted.
        (WTF::WordLock::unlockSlow): Deleted.
        * wtf/WordLock.h:
        (WTF::WordLockBase::lock):
        (WTF::WordLockBase::isLocked):
        (WTF::WordLock::WordLock):
        (WTF::WordLock::lock): Deleted.
        (WTF::WordLock::isLocked): Deleted.

2016-03-16  Chris Dumez  <cdumez@apple.com>

        Unreviewed, rolling out r198235, r198240, r198241, and
        r198252.

        Causing crashes on ARM

        Reverted changesets:

        "Remove compile time define for SEPARATED_HEAP"
        https://bugs.webkit.org/show_bug.cgi?id=155508
        http://trac.webkit.org/changeset/198235

        "Gardening: build fix after r198235."
        http://trac.webkit.org/changeset/198240

        "Build fix."
        http://trac.webkit.org/changeset/198241

        "Rename performJITMemcpy to something more inline with our
        normal webkit function names"
        https://bugs.webkit.org/show_bug.cgi?id=155525
        http://trac.webkit.org/changeset/198252

2016-03-15  Oliver Hunt  <oliver@apple.com>

        Remove compile time define for SEPARATED_HEAP
        https://bugs.webkit.org/show_bug.cgi?id=155508

        Reviewed by Mark Lam.

        Remove the feature define.

        * wtf/FeatureDefines.h:
        * wtf/Platform.h:

2016-03-14  Mark Lam  <mark.lam@apple.com>

        Need to distinguish between Symbol() and Symbol("").
        https://bugs.webkit.org/show_bug.cgi?id=155438

        Reviewed by Saam Barati.

        While toString of both Symbol() and Symbol("") yields "Symbol()", Function.name
        should yield "" and "[]" respectively.  Hence, we need to tell between the two.
        This functionality will be needed later in https://bugs.webkit.org/show_bug.cgi?id=155437.

        We achieve this by creating another singleton instance like the empty StringImpl
        as the null StringImpl.  isNullSymbol() tests if the Stringimpl instance is a
        symbol, and its substring is the null() singleton.

        * wtf/text/StringImpl.cpp:
        (WTF::StringImpl::createSymbol):
        (WTF::StringImpl::createNullSymbol):
        (WTF::StringImpl::containsOnlyWhitespace):
        (WTF::StringImpl::createSymbolEmpty): Deleted.
        * wtf/text/StringImpl.h:
        (WTF::StringImpl::tryCreateUninitialized):
        (WTF::StringImpl::stringKind):
        (WTF::StringImpl::isSymbol):
        (WTF::StringImpl::isAtomic):
        (WTF::StringImpl::isNullSymbol):
        * wtf/text/StringStatics.cpp:
        (WTF::StringImpl::null):

2016-03-13  Joseph Pecoraro  <pecoraro@apple.com>

        Remove ENABLE(ES6_TEMPLATE_LITERAL_SYNTAX) guards
        https://bugs.webkit.org/show_bug.cgi?id=155417

        Reviewed by Yusuke Suzuki.

        * wtf/FeatureDefines.h:

2016-03-13  Konstantin Tokarev  <annulen@yandex.ru>

        Added new port JSCOnly.
        https://bugs.webkit.org/show_bug.cgi?id=154512

        Reviewed by Michael Catanzaro.

        This port allows to build JavaScriptCore engine with minimal
        dependencies.

        * wtf/PlatformJSCOnly.cmake: Added.
        * wtf/none/MainThreadNone.cpp: Added.
        * wtf/none/RunLoopNone.cpp: Added.
        * wtf/none/WorkQueueNone.cpp: Added.

2016-03-13  David Kilzer  <ddkilzer@apple.com>

        REGRESSION (r198079): Windows build broke because of "%PRId64" format specifier

        This fixes the following build failure in WebCore for Windows:
            C:\Source\WebCore\page\DOMTimer.cpp(396): error C2146: syntax error: missing ')' before identifier 'PRId64' [C:\WebKitBuild\Debug\Source\WebCore\WebCore.vcxproj]
            C:\Source\WebCore\page\DOMTimer.cpp(399): error C2146: syntax error: missing ')' before identifier 'PRId64' [C:\WebKitBuild\Debug\Source\WebCore\WebCore.vcxproj]

        * wtf/StdLibExtras.h: Define "PRId64" for Windows here so it may
        be shared.  This should fix DOMTimer.cpp after r198079 since it
        already includes StdLibExtras.h.

2016-03-11  Youenn Fablet  <youenn.fablet@crf.canon.fr>

        WTF should have a similar function as equalLettersIgnoringASCIICase to match beginning of strings
        https://bugs.webkit.org/show_bug.cgi?id=153419

        Reviewed by Darin Adler.

        Introducing startsWithLettersIgnoringASCIICase, to check the beginning of a string.
        Moving some code from WTF::equalLettersIgnoringASCIICaseCommonWithoutLength in
        WTF::hasPrefixWithLettersIgnoringASCIICaseCommon to enable reuse in
        WTF::startsWithLettersIgnoringASCIICaseCommon.

        * wtf/text/StringCommon.h:
        (WTF::hasPrefixWithLettersIgnoringASCIICaseCommon):
        (WTF::equalLettersIgnoringASCIICaseCommonWithoutLength):
        (WTF::startsWithLettersIgnoringASCIICaseCommonWithoutLength):
        (WTF::startsWithLettersIgnoringASCIICaseCommon):
        * wtf/text/StringImpl.h:
        (WTF::startsWithLettersIgnoringASCIICase):
        * wtf/text/WTFString.h:
        (WTF::startsWithLettersIgnoringASCIICase):

2016-03-10  Frederic Wang  <fwang@igalia.com>

        [GTK] Add support for WOFF2
        https://bugs.webkit.org/show_bug.cgi?id=152616

        Reviewed by Carlos Garcia Campos.

        * wtf/FeatureDefines.h: Enable USE_WOFF2 flag on GTK.

2016-03-09  Ryosuke Niwa  <rniwa@webkit.org>

        Add runtime flags for shadow DOM and custom elements
        https://bugs.webkit.org/show_bug.cgi?id=155213

        Reviewed by Dean Jackson.

        Removed the manual overrides of ENABLE_SHADOW_DOM and ENABLE_CUSTOM_ELEMENTS as they were
        making --no-shadow-dom and --no-custom-elements flags on build-webkit useless.

        * wtf/FeatureDefines.h:

2016-03-09  Keith Rollin  <krollin@apple.com>

        Add state dumping facility
        https://bugs.webkit.org/show_bug.cgi?id=154930
        <rdar://problem/24939135>

        Reviewed by Anders Carlsson.

        Add an OS_STATE flag to control the inclusion of process state dumping
        functionality.

        * wtf/Platform.h:

2016-03-09  Oliver Hunt  <oliver@apple.com>

        Fix old iOS

        * wtf/Platform.h:

2016-03-08  Filip Pizlo  <fpizlo@apple.com>

        Remove failing assertion. There are strings that claim to be atomic but that the
        compiler thread can totally deal with, like the empty string.

        Rubber stamped by Mark Lam.

        * wtf/text/StringImpl.h:
        (WTF::StringImpl::ref):
        (WTF::StringImpl::deref):

2016-03-08  Filip Pizlo  <fpizlo@apple.com>

        DFG should be able to constant-fold strings
        https://bugs.webkit.org/show_bug.cgi?id=155200

        Reviewed by Geoffrey Garen.

        Also disable assertions about reference counting strings on the JIT thread. We will do
        that now and it's OK.

        * wtf/text/StringImpl.h:
        (WTF::StringImpl::ref):
        (WTF::StringImpl::deref):

2016-03-08  Anders Carlsson  <andersca@apple.com>

        Fix AppKitCompatibilityDeclarations build.

        * wtf/mac/AppKitCompatibilityDeclarations.h:
        Remove duplicate declarations, conditionally define NSTextAlignment and
        add a NSWindowStyleMask typedef.

2016-03-08  Oliver Hunt  <oliver@apple.com>

        Start moving to separated writable and executable mappings in the JIT
        https://bugs.webkit.org/show_bug.cgi?id=155178

        Reviewed by Fil Pizlo.

        Update feature defines.

        * wtf/FeatureDefines.h:
        * wtf/Platform.h: ARM64 for now.

2016-03-08  Anders Carlsson  <andersca@apple.com>

        Add AppKit compatibility header
        https://bugs.webkit.org/show_bug.cgi?id=155202

        Reviewed by Beth Dakin.

        * WTF.xcodeproj/project.pbxproj:
        * wtf/mac/AppKitCompatibilityDeclarations.h: Added.

2016-03-08  Commit Queue  <commit-queue@webkit.org>

        Unreviewed, rolling out r197793 and r197799.
        https://bugs.webkit.org/show_bug.cgi?id=155195

        something weird happened while landing this and everything
        broke (Requested by olliej on #webkit).

        Reverted changesets:

        "Start moving to separated writable and executable mappings in
        the JIT"
        https://bugs.webkit.org/show_bug.cgi?id=155178
        http://trac.webkit.org/changeset/197793

        "arm64 build fix after r197793."
        http://trac.webkit.org/changeset/197799

2016-03-08  Oliver Hunt  <oliver@apple.com>

        Start moving to separated writable and executable mappings in the JIT
        https://bugs.webkit.org/show_bug.cgi?id=155178

        Reviewed by Filip Pizlo.

        Update feature defines.

        * wtf/FeatureDefines.h:
        * wtf/Platform.h: ARM64 for now.

2016-03-08  Daniel Bates  <dabates@apple.com>

        Support iterating over an OptionSet and checking if it is empty
        https://bugs.webkit.org/show_bug.cgi?id=154941
        <rdar://problem/24964187>

        Reviewed by Darin Adler.

        Implements support for iterating over the enumerators in an OptionSet as well as
        determining if the set is empty.

        Iterating over an OptionSet is in Big Theta(N) where N is the number of items in
        the set. More precisely, it is in Big Theta(log M) where M is the bitmask represented
        by the bitwise OR-ing of all enumerators in the set.

        * wtf/OptionSet.h: Added comment to describe the purpose of this class and its invariant -
        the enumerators must be positive powers of two.
        (WTF::OptionSet::Iterator::operator*): Returns the enumerator pointed to by the iterator.
        (WTF::OptionSet::Iterator::operator++): Advance to the next smallest enumerator in the set.
        (WTF::OptionSet::Iterator::operator==): Returns whether the iterator is equal to the specified iterator.
        (WTF::OptionSet::Iterator::operator!=): Returns whether the iterator is not equal to the specified iterator.
        (WTF::OptionSet::Iterator::Iterator): Added.
        (WTF::OptionSet::fromRaw): Instantiate using specialized private constructor to allow
        instantiation with a raw value of 0.
        (WTF::OptionSet::OptionSet): Specialized constructor that asserts that the specified value
        is a positive power of two. This variant is only compiled when assertions are enabled (i.e. !ASSERT_DISABLED).
        (WTF::OptionSet::isEmpty): Returns whether the set is empty.
        (WTF::OptionSet::begin): Returns an iterator to the enumerator with the smallest value in the set.
        (WTF::OptionSet::end): Returns an iterator that represents the end sentinel of the set.

2016-03-07  Keith Rollin  <krollin@apple.com>

        Enhance logging: Use "always on" macros
        https://bugs.webkit.org/show_bug.cgi?id=154499
        <rdar://problem/24757730>

        Reviewed by Chris Dumez.

        Update LOG_ALWAYS and LOG_ALWAYS_ERROR macros to take an expression to
        be evaluated. If the expression evaluates to True, the associated
        message is logged. This facility is used to prevent logging from being
        performed in private sessions, but it could be extended to prevent
        logging under other circumstances as well.

        * wtf/Assertions.h:

2016-03-04  Alex Christensen  <achristensen@webkit.org>

        Remove vcxproj build system
        https://bugs.webkit.org/show_bug.cgi?id=154388

        Rubber-stamped by Brent Fulgham.

        * WTF.vcxproj/WTF.submit.sln: Removed.
        * WTF.vcxproj/WTF.vcxproj: Removed.
        * WTF.vcxproj/WTF.vcxproj.filters: Removed.
        * WTF.vcxproj/WTFCFLite.props: Removed.
        * WTF.vcxproj/WTFCommon.props: Removed.
        * WTF.vcxproj/WTFCoreFoundation.props: Removed.
        * WTF.vcxproj/WTFDebug.props: Removed.
        * WTF.vcxproj/WTFDebugWinCairo.props: Removed.
        * WTF.vcxproj/WTFGenerated.make: Removed.
        * WTF.vcxproj/WTFGenerated.vcxproj: Removed.
        * WTF.vcxproj/WTFGenerated.vcxproj.filters: Removed.
        * WTF.vcxproj/WTFGeneratedCommon.props: Removed.
        * WTF.vcxproj/WTFGeneratedDebug.props: Removed.
        * WTF.vcxproj/WTFGeneratedProduction.props: Removed.
        * WTF.vcxproj/WTFGeneratedRelease.props: Removed.
        * WTF.vcxproj/WTFPostBuild.cmd: Removed.
        * WTF.vcxproj/WTFPreBuild.cmd: Removed.
        * WTF.vcxproj/WTFProduction.props: Removed.
        * WTF.vcxproj/WTFRelease.props: Removed.
        * WTF.vcxproj/WTFReleaseWinCairo.props: Removed.
        * WTF.vcxproj/build-generated-files.pl: Removed.
        * WTF.vcxproj/copy-files.cmd: Removed.
        * WTF.vcxproj/work-around-vs-dependency-tracking-bugs.py: Removed.

2016-03-04  Alex Christensen  <achristensen@webkit.org>

        Use NSURLSession for loading in WebKit2
        https://bugs.webkit.org/show_bug.cgi?id=154993

        Reviewed by Sam Weinig.

        * wtf/Platform.h:

2016-03-03  Daniel Bates  <dabates@apple.com>

        Add unit tests for WTF::OptionSet
        https://bugs.webkit.org/show_bug.cgi?id=154925
        <rdar://problem/24964211>

        Reviewed by Darin Adler.

        * wtf/CMakeLists.txt: Add header OptionSet.h to the list of WTF headers.
        * wtf/OptionSet.h: Use in-class initialization to initialize m_storage and declare
        the trivial constexpr constructor as default.
        (WTF::OptionSet::OptionSet): For convenience add a constructor that takes a std::initializer_list.
        This code was written by Anders Carlsson.

2016-03-03  Andy Estes  <aestes@apple.com>

        Adopt CFNetwork storage partitioning SPI
        https://bugs.webkit.org/show_bug.cgi?id=154957
        rdar://problem/23614620

        Reviewed by Darin Adler.

        * wtf/Platform.h: Defined HAVE_CFNETWORK_STORAGE_PARTITIONING.

2016-03-02  Konstantin Tokarev  <annulen@yandex.ru>

        [cmake] Moved PRE/POST_BUILD_COMMAND to WEBKIT_FRAMEWORK.
        https://bugs.webkit.org/show_bug.cgi?id=154651

        Reviewed by Alex Christensen.

        * CMakeLists.txt: Moved shared code to WEBKIT_FRAMEWORK macro.

2016-03-01  Commit Queue  <commit-queue@webkit.org>

        Unreviewed, rolling out r197226 and r197256.
        https://bugs.webkit.org/show_bug.cgi?id=154910

        Caused crashes on Mac 32-bit and on ARM (Requested by ap on
        #webkit).

        Reverted changesets:

        "Remove the on demand executable allocator"
        https://bugs.webkit.org/show_bug.cgi?id=154749
        http://trac.webkit.org/changeset/197226

        "CLoop build fix."
        http://trac.webkit.org/changeset/197256

2016-03-01  Alex Christensen  <achristensen@webkit.org>

        Reduce size of internal windows build output
        https://bugs.webkit.org/show_bug.cgi?id=154763

        Reviewed by Brent Fulgham.

        * WTF.vcxproj/WTF.proj:

2016-03-01  Commit Queue  <commit-queue@webkit.org>

        Unreviewed, rolling out r197056.
        https://bugs.webkit.org/show_bug.cgi?id=154870

        broke win ews (Requested by alexchristensen on #webkit).

        Reverted changeset:

        "[cmake] Moved PRE/POST_BUILD_COMMAND to WEBKIT_FRAMEWORK."
        https://bugs.webkit.org/show_bug.cgi?id=154651
        http://trac.webkit.org/changeset/197056

2016-02-29  Gavin Barraclough  <barraclough@apple.com>

        RefCounter<T>::Event -> RefCounterEvent
        https://bugs.webkit.org/show_bug.cgi?id=154767

        Reviewed by Darin Adler.

        RefCounter<T>::Event is kinda verbose to use, and there is no need for this
        to be specific to a particular typeof RefCounter. Move the enum class up to
        the top level & rename to RefCounterEvent.

        * wtf/RefCounter.h:
        (WTF::RefCounter<T>::Count::ref):
        (WTF::RefCounter<T>::Count::deref):

2016-02-26  Oliver Hunt  <oliver@apple.com>

        Remove the on demand executable allocator
        https://bugs.webkit.org/show_bug.cgi?id=154749

        Reviewed by Geoffrey Garen.

        Remove the DeamndExecutableAllocator compile flags.

        * wtf/Platform.h:

2016-02-26  Keith Rollin  <krollin@apple.com>

        Enhance logging: Add "always on" macros
        https://bugs.webkit.org/show_bug.cgi?id=154498
        <rdar://problem/24757759>

        Reviewed by Chris Dumez.

        Add support for efficient always-on logging (logging that is available
        in both debug and release builds). This is implemented in the form of
        some new macros:

        - LOG_ALWAYS: Always log information-level statements.
        - LOG_ALWAYS_ERROR: Always log error-level statements. These can be
          filtered out of the normal logging so that they can be found more
          easily.

        In cases where there is no efficient underlying facility for it to
        utilize, LOG_ALWAYS_ERROR is redirected to WTFReportError as a backup
        mechanism. LOG_ALWAYS is not given a similar treatment so that we
        don't overwhelm other logging systems that aren't prepared for "always
        on" logging.

        * wtf/Assertions.h:
        * wtf/Platform.h:

2016-02-26  Anders Carlsson  <andersca@apple.com>

        Add WTF::OptionSet and use it for the website data types enum
        https://bugs.webkit.org/show_bug.cgi?id=154733

        Reviewed by Geoffrey Garen.

        Add WTF::OptionSet which makes it easier to use strong enums as flags.

        * WTF.xcodeproj/project.pbxproj:
        * wtf/Forward.h:
        * wtf/OptionSet.h: Copied from Source/WebKit2/Shared/WebsiteData/WebsiteData.h.
        (WTF::OptionSet::fromRaw):
        (WTF::OptionSet::OptionSet):
        (WTF::OptionSet::toRaw):
        (WTF::OptionSet::contains):
        (WTF::OptionSet::operator|=):

2016-02-26  Commit Queue  <commit-queue@webkit.org>

        Unreviewed, rolling out r197168.
        https://bugs.webkit.org/show_bug.cgi?id=154728

        crashing on some devices (Requested by kling on #webkit).

        Reverted changeset:

        "[Darwin] Use vm_kernel_page_size for WTF::pageSize()."
        https://bugs.webkit.org/show_bug.cgi?id=154726
        http://trac.webkit.org/changeset/197168

2016-02-26  Andreas Kling  <akling@apple.com>

        [Darwin] Use vm_kernel_page_size for WTF::pageSize().
        <https://webkit.org/b/154726>

        Reviewed by Antti Koivisto.

        Make sure we use the right VM page size on Darwin systems.
        On devices where the native page size is 4 KB, it's both
        possible and profitable to madvise in smaller chunks.

        * wtf/PageBlock.cpp:
        (WTF::systemPageSize):

2016-02-25  Gavin Barraclough  <barraclough@apple.com>

        RefCounter value changed callback should be called on all changes (not just zero edge).
        https://bugs.webkit.org/show_bug.cgi?id=154699

        Reviewed by Anders Carlsson.

        RefCounter currently only triggers a callback when the count goes from zero
        to non-zero and vice-versa. Change that, to be useful to more clients.

        * wtf/RefCounter.h:
        (WTF::RefCounter::Count::Count):
        (WTF::RefCounter::RefCounter):
            - Removed superfluous WTF_EXPORT_PRIVATE.
        (WTF::RefCounter::value):
            - Changed value() to a size_t.
        (WTF::RefCounter<T>::Count::ref):
        (WTF::RefCounter<T>::Count::deref):
            - Trigger the callback on all increments/decrements.
        (WTF::RefCounter<T>::RefCounter):
            - Changed siganture of callback.

2016-02-25  Gavin Barraclough  <barraclough@apple.com>

        Replace RefCounter::Token implementation with RefPtr
        https://bugs.webkit.org/show_bug.cgi?id=154698

        Reviewed by Anders Carlsson.

        * wtf/RefCounter.h:
        (WTF::RefCounter::RefCounter):
        (WTF::RefCounter::count):
        (WTF::RefCounter::value):
        (WTF::RefCounter<T>::~RefCounter):
        (WTF::RefCounter::Token::Token): Deleted.
        (WTF::RefCounter::Token::operator bool): Deleted.
        (WTF::RefCounter<T>::Token::Token): Deleted.
        (WTF::=): Deleted.
            - Delete.

2016-02-25  Gavin Barraclough  <barraclough@apple.com>

        Should template RefCounter instead of RefCounter::Token
        https://bugs.webkit.org/show_bug.cgi?id=154691

        Speculative windows build fix.

        * wtf/RefCounter.h:
        (WTF::RefCounter::RefCounter):
        (WTF::RefCounter::count):

2016-02-25  Commit Queue  <commit-queue@webkit.org>

        Unreviewed, rolling out r197137.
        https://bugs.webkit.org/show_bug.cgi?id=154700

        broke windows build (Requested by alexchristensen on #webkit).

        Reverted changeset:

        "Replace RefCounter::Token implementation with RefPtr"
        https://bugs.webkit.org/show_bug.cgi?id=154698
        http://trac.webkit.org/changeset/197137

2016-02-25  Gavin Barraclough  <barraclough@apple.com>

        Replace RefCounter::Token implementation with RefPtr
        https://bugs.webkit.org/show_bug.cgi?id=154698

        Reviewed by Anders Carlsson.

        * wtf/RefCounter.h:
        (WTF::RefCounter::RefCounter):
        (WTF::RefCounter::count):
        (WTF::RefCounter::value):
        (WTF::RefCounter<T>::~RefCounter):
        (WTF::RefCounter::Token::Token): Deleted.
        (WTF::RefCounter::Token::operator bool): Deleted.
        (WTF::RefCounter<T>::Token::Token): Deleted.
        (WTF::=): Deleted.
            - Delete.

2016-02-25  Gavin Barraclough  <barraclough@apple.com>

        Should template RefCounter instead of RefCounter::Token
        https://bugs.webkit.org/show_bug.cgi?id=154691

        Reviewed by Anders Carlsson.

        My real goal here is to make the counter accurate. Currently returning a Token from token<>()
        results in ref-count churn. Fixing this either means changing the return value, or improving
        Token (which will probably mean replacing it with RefPtr). Either way would break the current
        type checking. Move type tag to RefCount so this can still be enforced.

        * WTF.vcxproj/WTF.vcxproj:
        * WTF.vcxproj/WTF.vcxproj.filters:
        * WTF.xcodeproj/project.pbxproj:
        * wtf/CMakeLists.txt:
        * wtf/RefCounter.cpp: Removed.
            - Removed RefCounter.cpp.
        * wtf/RefCounter.h:
        (WTF::RefCounter::Token::Token):
        (WTF::RefCounter::Token::operator bool):
        (WTF::RefCounter::RefCounter):
        (WTF::RefCounter::count):
        (WTF::RefCounter::value):
        (WTF::RefCounter<T>::Count::ref):
        (WTF::RefCounter<T>::Count::deref):
        (WTF::RefCounter<T>::RefCounter):
        (WTF::RefCounter<T>::~RefCounter):
        (WTF::RefCounter<T>::Token::Token):
        (WTF::=):
        (WTF::RefCounter::token): Deleted.
        (WTF::RefCounter::Token<T>::Token): Deleted.
            - RefCounter -> RefCounter<T>,
            - Token<T> -> Token,
            - renamed token<>() -> count().

2016-02-25  Sam Weinig  <sam@webkit.org>

        HashMap::ensure() should return an AddResult like all the other add-like functions.
        https://bugs.webkit.org/show_bug.cgi?id=154680

        Reviewed by Anders Carlsson.

        While adopting HashMap::ensure(), I found it was useful in some circumstances to know
        if the value was added or not. While I could discern this information by setting a bool
        in the passed in lambda, it seemed clearer and more idiomatic to just have ensure return
        an AddResult like all the other add-like functions do.

        * wtf/HashMap.h:
        Change return type of HashMap::ensure() to be an AddResult.

2016-02-24  Nikos Andronikos  <nikos.andronikos-webkit@cisra.canon.com.au>

        [web-animations] Add AnimationTimeline, DocumentTimeline and add extensions to Document interface
        https://bugs.webkit.org/show_bug.cgi?id=151688

        Reviewed by Dean Jackson.

        Enables the WEB_ANIMATIONS compiler switch.

        * wtf/FeatureDefines.h:

2016-02-24  Konstantin Tokarev  <annulen@yandex.ru>

        [cmake] Moved PRE/POST_BUILD_COMMAND to WEBKIT_FRAMEWORK.
        https://bugs.webkit.org/show_bug.cgi?id=154651

        Reviewed by Alex Christensen.

        * CMakeLists.txt: Moved shared code to WEBKIT_FRAMEWORK macro.

2016-02-23  Dan Bernstein  <mitz@apple.com>

        [Xcode] Linker errors display mangled names, but no longer should
        https://bugs.webkit.org/show_bug.cgi?id=154632

        Reviewed by Sam Weinig.

        * Configurations/Base.xcconfig: Stop setting LINKER_DISPLAYS_MANGLED_NAMES to YES.

2016-02-23  Gavin Barraclough  <barraclough@apple.com>

        Remove HIDDEN_PAGE_DOM_TIMER_THROTTLING feature define
        https://bugs.webkit.org/show_bug.cgi?id=112323

        Reviewed by Chris Dumez.

        This feature is controlled by a runtime switch, and defaults off.

        * wtf/FeatureDefines.h:

2016-02-22  Konstantin Tokarev  <annulen@yandex.ru>

        [cmake] Moved library setup code to WEBKIT_FRAMEWORK macro.
        https://bugs.webkit.org/show_bug.cgi?id=154450

        Reviewed by Alex Christensen.

        * wtf/CMakeLists.txt:

2016-02-20  Mark Lam  <mark.lam@apple.com>

        Use of inlined asm statements causes problems for -std=c99 builds.
        https://bugs.webkit.org/show_bug.cgi?id=154507

        Reviewed by Dan Bernstein.

        WTF's Assertions.h may inadvertantly get included by other projects that are built
        with -std=c99.  The use of the inlined asm statements with the keyword "asm" is
        not recognized when the -std compiler flag is used.

        https://gcc.gnu.org/onlinedocs/gcc/Extended-Asm.html says "When writing code that
        can be compiled with -ansi and the various -std options, use __asm__ instead of
        asm (see Alternate Keywords)."

        So, to be a good citizen, we can change the use of "asm" in CRASH() to "__asm__"
        so that we don't break the build of such other projects.

        * wtf/Assertions.h:

2016-02-18  Brent Fulgham  <bfulgham@apple.com>

        Extend HashCountedSet with a method to efficiently set the count of an entry
        https://bugs.webkit.org/show_bug.cgi?id=154352

        Reviewed by Geoffrey Garen.

        Tested by new TestWebKitAPI tests.

        Update the HashCountedSet class with a new 'add' method to support efficient initialization of
        the count of a given key. Also provide move and pointer template specializations to expand the
        types of data that can be used as 'keys' in the HashCountedSet to match the underlying HashMap
        implementation.

        * wtf/HashCountedSet.h:
        (WTF::Traits>::add): Added new overload supporting a supplied count.

2016-02-18  Commit Queue  <commit-queue@webkit.org>

        Unreviewed, rolling out r196791.
        https://bugs.webkit.org/show_bug.cgi?id=154438

        broke windows build (Requested by alexchristensen on #webkit).

        Reverted changeset:

        "Extend HashCountedSet with a method to efficiently set the
        count of an entry"
        https://bugs.webkit.org/show_bug.cgi?id=154352
        http://trac.webkit.org/changeset/196791

2016-02-18  Brent Fulgham  <bfulgham@apple.com>

        Extend HashCountedSet with a method to efficiently set the count of an entry
        https://bugs.webkit.org/show_bug.cgi?id=154352

        Reviewed by Geoffrey Garen.

        Tested by new TestWebKitAPI tests.

        Update the HashCountedSet class with a new 'add' method to support efficient initialization of
        the count of a given key. Also provide move and pointer template specializations to expand the
        types of data that can be used as 'keys' in the HashCountedSet to match the underlying HashMap
        implementation.

        * wtf/HashCountedSet.h:
        (WTF::Traits>::add): Added new overload supporting a supplied count.

2016-02-17  Filip Pizlo  <fpizlo@apple.com>

        Remove LLVM dependencies from WebKit
        https://bugs.webkit.org/show_bug.cgi?id=154323

        Reviewed by Antti Koivisto and Benjamin Poulain.

        * wtf/Platform.h:

2016-02-16  Sam Weinig  <sam@webkit.org>

        Add an ensure function on HashMap that takes a key and a function to make the lazy value initialization idiom easier
        https://bugs.webkit.org/show_bug.cgi?id=134857

        Reviewed by Geoffrey Garen.

        The current idiom for adding a value to a HashMap if the key is not already present, to allow for not
        unnecessarily constructing the new value if it won't be used, is:
        
        {
            auto result = map.add(key, Value());
            if (!result.isNewEntry)
                return result.iterator->value;

            result.iterator->value = createNewValue();
            return result.iterator->value;
        }

        or sometimes

        {
            auto& value = map.add(key, Value()).iterator->value;
            if (!value)
                value = createNewValue();
            return value;
        }

        With this patch, you can now use the new function HashMap::ensure(key, functor). This will allow us to 
        change to using the following idiom:

        {
            return map.ensure(key, [] { return createNewValue(); });
        }

        The passed in functor will only be called if the key is not already present in the HashMap.

        * wtf/HashMap.h:
        (WTF::HashMapTranslator::translate):
        (WTF::HashMapEnsureTranslator::hash):
        (WTF::HashMapEnsureTranslator::equal):
        (WTF::HashMapEnsureTranslator::translate):
        (WTF::HashMapTranslatorAdapter::hash):

2016-02-16  Filip Pizlo  <fpizlo@apple.com>

        FTL_USES_B3 should be unconditionally true
        https://bugs.webkit.org/show_bug.cgi?id=154324

        Reviewed by Benjamin Poulain.

        * wtf/Platform.h:

2016-02-13  Mark Lam  <mark.lam@apple.com>

        Eliminate the need for WTFCrashImpl().
        https://bugs.webkit.org/show_bug.cgi?id=154202

        Reviewed by Alexey Proskuryakov.

        * wtf/Assertions.cpp:
        * wtf/Assertions.h:

2016-02-12  Sukolsak Sakshuwong  <sukolsak@gmail.com>

        Update ICU header files to version 52
        https://bugs.webkit.org/show_bug.cgi?id=154160

        Reviewed by Alex Christensen.

        Update ICU header files to version 52 to allow the use of newer APIs.

        * icu/unicode/bytestream.h:
        * icu/unicode/localpointer.h:
        * icu/unicode/platform.h:
        * icu/unicode/ptypes.h:
        * icu/unicode/putil.h:
        * icu/unicode/rep.h:
        (Replaceable::Replaceable):
        * icu/unicode/std_string.h:
        * icu/unicode/strenum.h:
        * icu/unicode/stringpiece.h:
        * icu/unicode/uchar.h:
        * icu/unicode/ucnv.h:
        * icu/unicode/ucol.h:
        * icu/unicode/uconfig.h:
        * icu/unicode/uenum.h:
        * icu/unicode/uiter.h:
        * icu/unicode/uloc.h:
        * icu/unicode/umachine.h:
        * icu/unicode/unistr.h:
        (UnicodeString::UnicodeString):
        (UnicodeString::operator== ):
        (UnicodeString::startsWith):
        (UnicodeString::setTo):
        (UnicodeString::remove):
        (UnicodeString::replace): Deleted.
        (UnicodeString::extract): Deleted.
        (UnicodeString::char32At): Deleted.
        (UnicodeString::getChar32Start): Deleted.
        (UnicodeString::getChar32Limit): Deleted.
        (UnicodeString::getTerminatedBuffer): Deleted.
        (UnicodeString::append): Deleted.
        (UnicodeString::truncate): Deleted.
        * icu/unicode/unorm2.h:
        * icu/unicode/uobject.h:
        * icu/unicode/urename.h:
        * icu/unicode/uscript.h:
        * icu/unicode/uset.h:
        * icu/unicode/ustring.h:
        * icu/unicode/utf.h:
        * icu/unicode/utf16.h:
        * icu/unicode/utf8.h:
        * icu/unicode/utf_old.h:
        * icu/unicode/utypes.h:
        * icu/unicode/uvernum.h:
        * icu/unicode/uversion.h:

2016-02-11  Mark Lam  <mark.lam@apple.com>

        Need WTFCrash workaround for shipping SafariForWebKitDevelopment binaries.
        https://bugs.webkit.org/show_bug.cgi?id=154125

        Reviewed by Joseph Pecoraro.

        Presently shipping SafariForWebKitDevelopment binaries still expect to link to a
        WTFCrash function.  We need to provide this function as a workaround until we can
        update SafariForWebKitDevelopment to use the new inlined version.

        We do this by doing:
        1. Make WTFCrashImpl() the sole function for implementing a crash.
           The CRASH() macro is now defined to be WTFCrashImpl() instead of WTFCrash().
        2. Renamed the legacy WTFCrash() to WTFCrashImpl() for debug or non-Darwin builds.
           For (non-debug && OS(DARWIN)) builds, WTFCrashImpl() will be an inlined
           function with an asm statement that issues a breakpoint trap.
        3. Implement WTFCrash() as a function that calls CRASH().
           This satisfies the need of shipping SafariForWebKitDevelopment binaries.
        4. Change WTFCrashWithSecurityImplication() to call CRASH().
           This ensures that we have a consistent implementation of how we crash.
        5. Changed WTFLogAlwaysAndCrash() to call CRASH() instead of WTFCrash().
           This is just to have consistency in that all code in the WebKit project
           now crashes by calling CRASH(), not WTFCrash().

        * wtf/Assertions.cpp:
        * wtf/Assertions.h:

2016-02-09  Mark Lam  <mark.lam@apple.com>

        Changed WTFCrash to not trash the crash site register state.
        https://bugs.webkit.org/show_bug.cgi?id=153996

        Reviewed by Geoffrey Garen.

        When doing post-mortem crash site analysis using data from crash reports, it is
        immensely valuable to be able to infer the crashing program's state from the
        register values at crash time.  However, for RELEASE_ASSERT failures, we crash
        using WTFCrash(), and WTFCrash() is currently implemented as a function call
        that, in turn, calls a lot of other functions to do crash handling before
        actually crashing.  As a result, the register values captured in the crash
        reports are not likely to still contain the values used by the caller function
        that failed the RELEASE_ASSERT.

        This patch aims to remedy this issue for non-debug builds on OS(DARWIN) ports.
        It does so by changing WTFCrash() into an inlined function that has an inlined
        asm statement to issues the CPU specific breakpoint trap instruction.  As a
        result, for non-debug OS(DARWIN) builds, crashes due to failed RELEASE_ASSERTs
        will now show up in crash reports as crashing due to EXC_BREAKPOINT (SIGTRAP)
        instead of a EXC_BAD_ACCESS (SIGSEGV) on address 0xbbadbeef.

        For debug and non-DARWIN builds, WTFCrash() behavior currently remains unchanged.

        * wtf/Assertions.cpp:
        * wtf/Assertions.h:

2016-02-09  Csaba Osztrogonác  <ossy@webkit.org>

        [GTK][EFL] Fix several build configuration related to SamplingProfiler after r196245
        https://bugs.webkit.org/show_bug.cgi?id=154033

        Reviewed by Michael Catanzaro.

        * wtf/Platform.h:

2016-02-08  Filip Pizlo  <fpizlo@apple.com>

        Make sure that locking code that relies on module boundaries for compiler fences uses NEVER_INLINE
        https://bugs.webkit.org/show_bug.cgi?id=153972

        Reviewed by Andreas Kling.

        When this code was written, we assumed that module boundaries were compiler fences. That might
        not be the case if we ever do LTO.

        * wtf/Lock.cpp:
        (WTF::LockBase::lockSlow):
        (WTF::LockBase::unlockSlow):
        * wtf/ParkingLot.cpp:
        (WTF::ParkingLot::parkConditionally):
        (WTF::ParkingLot::unparkOne):
        (WTF::ParkingLot::unparkAll):
        (WTF::ParkingLot::forEach):
        * wtf/WordLock.cpp:
        (WTF::WordLock::lockSlow):
        (WTF::WordLock::unlockSlow):

2016-02-07  Yusuke Suzuki  <utatane.tea@gmail.com>

        [GTK][EFL] Enable SamplingProfiler
        https://bugs.webkit.org/show_bug.cgi?id=153638

        Reviewed by Michael Catanzaro.

        * wtf/Platform.h:

2016-02-07  Dan Bernstein  <mitz@apple.com>

        [Cocoa] Replace __has_include guards around inclusion of Apple-internal-SDK headers with USE(APPLE_INTERNAL_SDK)
        https://bugs.webkit.org/show_bug.cgi?id=153963

        Reviewed by Sam Weinig.

        * wtf/SystemTracing.h:
        * wtf/WTFThreadData.h:
        * wtf/spi/darwin/CommonCryptoSPI.h:

2016-02-06  Darin Adler  <darin@apple.com>

        Finish auditing call sites of upper() and lower(), eliminate many, and rename the functions
        https://bugs.webkit.org/show_bug.cgi?id=153905

        Reviewed by Sam Weinig.

        * wtf/text/AtomicString.cpp:
        (WTF::AtomicString::lower): Deleted.
        * wtf/text/AtomicString.h: Deleted the lower function.

        * wtf/text/StringImpl.cpp:
        (WTF::StringImpl::convertToLowercaseWithoutLocale): Renamed from lower.
        (WTF::StringImpl::convertToUppercaseWithoutLocale): Renamed from upper.
        (WTF::StringImpl::convertToLowercaseWithLocale): Renamed from lower.
        (WTF::StringImpl::convertToUppercaseWithLocale): Renamed from upper.
        (WTF::StringImpl::foldCase): Added fast cases for ASCII since this is
        now used in some more-performance-critical code.
        * wtf/text/StringImpl.h: Renamed lower and upper.

        * wtf/text/WTFString.cpp:
        (WTF::String::convertToLowercaseWithoutLocale): Renamed from lower.
        (WTF::String::convertToUppercaseWithoutLocale): Renamed from upper.
        (WTF::String::convertToLowercaseWithLocale): Renamed from lower.
        (WTF::String::convertToUppercaseWithLocale): Renamed from upper.
        * wtf/text/WTFString.h: Renamed lower and upper. Removed unneeded comment.

2016-02-03  Darin Adler  <darin@apple.com>

        Convert another batch of String::lower callsites to something better, typically convertToASCIILowercase
        https://bugs.webkit.org/show_bug.cgi?id=153789

        Reviewed by Sam Weinig.

        * wtf/text/StringView.h:
        (WTF::StringView::toInt): Added an overload without the out parameter.

2016-02-03  Michael Catanzaro  <mcatanzaro@igalia.com>

        [GTK][EFL] Switch FTL to B3
        https://bugs.webkit.org/show_bug.cgi?id=153478

        Reviewed by Csaba Osztrogonác.

        Enable B3 by default on all x86.

        * wtf/Platform.h:

2016-02-03  Anders Carlsson  <andersca@apple.com>

        Fix BlockPtr's call operator
        https://bugs.webkit.org/show_bug.cgi?id=153836

        Reviewed by Enrica Casucci.

        * wtf/BlockPtr.h:

2016-02-01  Said Abou-Hallawa  <sabouhallawa@apple.com>

        Cache the Path instead of creating it every time it is required
        https://bugs.webkit.org/show_bug.cgi?id=152939

        Reviewed by Darin Adler.

        If the key type of an LRU cache can't to be strongly tided to a specific
        data type; e.g. FloatRect -> Path, we need to be able to pass the policy
        type to the TinyLRUCache template instead of just specializing it. This
        will make the code more readable and will allow different caches for the
        same key type.

        * wtf/TinyLRUCache.h:
        (WebCore::TinyLRUCache::get):

2016-02-01  Alex Christensen  <achristensen@webkit.org>

        [Win] WTFHeaderDetection.h no longer needed
        https://bugs.webkit.org/show_bug.cgi?id=153753
        rdar://problem/24434627

        Reviewed by Darin Adler.

        * wtf/Platform.h:
        * wtf/PlatformWin.cmake:

2016-01-31  Darin Adler  <darin@apple.com>

        Cut down on calls to String::lower; mostly replace with convertToASCIILowercase
        https://bugs.webkit.org/show_bug.cgi?id=153732

        Reviewed by Dean Jackson.

        * wtf/text/StringView.h: Added toIntStrict. Not thrilled about the name of this
        function, but generally it seems likely to be useful more often than toInt.
        And it's needed for one call site in WebCore that was using String::lower.

2016-01-31  Darin Adler  <darin@apple.com>

        Get rid of most calls to String::upper; mostly replace them with convertToASCIIUppercase
        https://bugs.webkit.org/show_bug.cgi?id=153715

        Reviewed by Gyuyoung Kim.

        * wtf/text/AtomicString.h:
        (WTF::AtomicString::upper): Deleted.

2016-01-31  Dan Bernstein  <mitz@apple.com>

        [Cocoa] Remove unused definition of HAVE_HEADER_DETECTION_H
        https://bugs.webkit.org/show_bug.cgi?id=153729

        Reviewed by Sam Weinig.

        After r141700, HAVE_HEADER_DETECTION_H is no longer used.

        * Configurations/Base.xcconfig:

2016-01-31  Darin Adler  <darin@apple.com>

        Replace CaseFoldingHash with ASCIICaseInsensitiveHash
        https://bugs.webkit.org/show_bug.cgi?id=153639

        Reviewed by Filip Pizlo.

        * wtf/text/StringHash.h: Renamed CaseFoldingHash to ASCIICaseInsensitiveHash.
        (WTF::ASCIICaseInsensitiveHash::foldCase): Use toASCIILower.
        (WTF::ASCIICaseInsensitiveHash::equal): Use equalIgnoringASCIICase.
        Also added some assertions.

        * wtf/text/StringImpl.cpp: Made the latin1CaseFoldTable private to this file,
        since it's no longer needed by CaseFoldingHash. Moved it up before its first use.
        (WTF::equalCompatibilityCaseless): Fixed typo in the name of this function.
        (WTF::equalCompatibiltyCaselessNonNull): Deleted.

        * wtf/text/StringImpl.h: Removed declarations of latin1CaseFoldTable and
        equalCompatibiltyCaselessNonNull

2016-01-30  Commit Queue  <commit-queue@webkit.org>

        Unreviewed, rolling out r195911.
        https://bugs.webkit.org/show_bug.cgi?id=153723

        Caused frequent assertion failures on bots (Requested by ap on
        #webkit).

        Reverted changeset:

        "Replace CaseFoldingHash with ASCIICaseInsensitiveHash"
        https://bugs.webkit.org/show_bug.cgi?id=153639
        http://trac.webkit.org/changeset/195911

2016-01-30  Darin Adler  <darin@apple.com>

        Replace CaseFoldingHash with ASCIICaseInsensitiveHash
        https://bugs.webkit.org/show_bug.cgi?id=153639

        Reviewed by Filip Pizlo.

        * wtf/text/StringHash.h: Renamed CaseFoldingHash to ASCIICaseInsensitiveHash.
        (WTF::ASCIICaseInsensitiveHash::foldCase): Use toASCIILower.
        (WTF::ASCIICaseInsensitiveHash::equal): Use equalIgnoringASCIICase.
        Also added some assertions.

        * wtf/text/StringImpl.cpp: Made the latin1CaseFoldTable private to this file,
        since it's no longer needed by CaseFoldingHash. Moved it up before its first use.
        (WTF::equalCompatibilityCaseless): Fixed typo in the name of this function.
        (WTF::equalCompatibiltyCaselessNonNull): Deleted.

        * wtf/text/StringImpl.h: Removed declarations of latin1CaseFoldTable and
        equalCompatibiltyCaselessNonNull

2016-01-29  Ada Chan  <adachan@apple.com>

        Enable VIDEO_PRESENTATION_MODE only in Debug and Release builds on Mac
        https://bugs.webkit.org/show_bug.cgi?id=153665

        Reviewed by Dan Bernstein.

        * wtf/Platform.h:
        Remove this as the flag is already defined in FeatureDefines.xcconfig files.

2016-01-30  Yusuke Suzuki  <utatane.tea@gmail.com>

        Enable SamplingProfiler on POSIX environment
        https://bugs.webkit.org/show_bug.cgi?id=153584

        Reviewed by Michael Saboff.

        Use __GLIBC__ since mcontext_t layout depends on it.

        * wtf/Platform.h:

2016-01-28  Darin Adler  <darin@apple.com>

        Remove equalIgnoringCase since all callers really wanted equalIgnoringASCIICase
        https://bugs.webkit.org/show_bug.cgi?id=153411

        Reviewed by Ryosuke Niwa.

        * wtf/text/AtomicString.h: Removed equalIgnoringCase.
        Added some overloads for equalIgnoringASCIICase and moved the function bodies to
        bottom of the file to make the function declarations easier to read and scan through.
        I plan to do this for more of the functions in this class in the future.

        * wtf/text/StringCommon.h: Added overloads for equalIgnoringASCIICase,
        equalPossiblyIgnoringASCIICase, and a helper named equalIgnoringASCIICaseCommon.
        Added an overload for equalLettersIgnoringASCIICase.

        * wtf/text/StringImpl.cpp:
        (WTF::equalIgnoringCase): Made the few remaining versions of this function private
        to this file. Once we get done moving every client that should be using ASCII case
        instead to new functions, these will almost certainly be deleted.
        (WTF::equalIgnoringASCIICaseNonNull): Tweaked implementation a tiny bit.

        * wtf/text/StringImpl.h: Sorted forward declarations at the top of the file.
        Fixed some small formatting mistakes. Removed equalIgnoringCase, but left
        equalIgnoringCaseNonNull behind for use by CaseFoldingHash until it is replaced
        with ASCIICaseFoldingHash. Improved equalIgnoringASCIICase implementation.
        Removed unneeded using for equalIgnoringASCIICase now that StringCommon.h takes
        care of it.

        * wtf/text/StringView.cpp:
        (WTF::equalIgnoringASCIICase): Deleted. We no longer pass in the length for this,
        so the arguments have changed and the implementation is also now in ASCIICommon.h
        so it's an inline in the header file.

        * wtf/text/StringView.h: Added an overload of equalIgnoringASCIICase,while
        removing another.

        * wtf/text/WTFString.h: Removed some unneeded forward delcarations. Fixed formatting
        of the type name NSString *. Improved some comments. Removed equalIgnoringCase.
        Separated declaration from implementation in a few cases to start making the
        function declarations easier to read and scan through. I plan to do more in the future.

2016-01-27  Chris Dumez  <cdumez@apple.com>

        window.atob() should ignore spaces in input
        https://bugs.webkit.org/show_bug.cgi?id=153522
        <rdar://problem/24357822>

        Reviewed by Benjamin Poulain.

        Turn Base64DecodePolicy enum into flags so that the caller can indicate
        to both validate padding AND ignore spaces.

        Also make sure that the output Vector size is properly shrunk when
        spaces are ignored.

        * wtf/text/Base64.cpp:
        (WTF::base64DecodeInternal):
        (WTF::base64Decode):
        (WTF::base64URLDecode):
        * wtf/text/Base64.h:

2016-01-27  Alexey Proskuryakov  <ap@apple.com>

        Remove ENABLE_CURRENTSRC
        https://bugs.webkit.org/show_bug.cgi?id=153545

        Reviewed by Simon Fraser.

        * wtf/FeatureDefines.h:

2016-01-26  I-Ting Liu  <iting_liu@apple.com>

        Implement wildcard matching for plug-in policy host.
        https://bugs.webkit.org/show_bug.cgi?id=153090

        Reviewed by Darin Adler.

        * wtf/text/AtomicString.h:
        (WTF::AtomicString::AtomicString):
        Add __bridge to allow compilation.

2016-01-26  Joseph Pecoraro  <pecoraro@apple.com>

        Generalize ResourceUsageData gathering to be used outside of ResourceUsageOverlay
        https://bugs.webkit.org/show_bug.cgi?id=153509
        <rdar://problem/24354291>

        Reviewed by Andreas Kling.

        * wtf/Platform.h:
        Rename the ENABLE flag.

2016-01-26  Yusuke Suzuki  <utatane.tea@gmail.com>

        Make HashTable iterators STL iterators compatible
        https://bugs.webkit.org/show_bug.cgi?id=153512

        Reviewed by Alex Christensen.

        While r178581 makes many hash table iterators STL compatible, still several iterators are not.
        This patch fixes that; inheriting std::iterator correctly to meet STL iterator requirements (like iterator_category etc.)
        It could recover Windows build failure.

        * wtf/HashTable.h:

2016-01-26  Anders Carlsson  <andersca@apple.com>

        BlockPtr needs boolean operators
        https://bugs.webkit.org/show_bug.cgi?id=153506

        Reviewed by Tim Horton.

        * wtf/BlockPtr.h:

2016-01-25  Filip Pizlo  <fpizlo@apple.com>

        FTLB3Output should maintain good block order like the LLVM one does
        https://bugs.webkit.org/show_bug.cgi?id=152222

        Reviewed by Geoffrey Garen.

        In the FTL we need to be able to construct a list by inserting elements before other
        specific elements. We didn't already have a scalable way to do this, so this adds such a
        data structure to WTF. This also has changes to SentinelLinkedList to make it support
        these kinds of insertions.

        * WTF.xcodeproj/project.pbxproj:
        * wtf/OrderMaker.h: Added.
        (WTF::OrderMaker::Node::Node):
        (WTF::OrderMaker::OrderMaker):
        (WTF::OrderMaker::prepend):
        (WTF::OrderMaker::append):
        (WTF::OrderMaker::insertBefore):
        (WTF::OrderMaker::insertAfter):
        (WTF::OrderMaker::iterator::iterator):
        (WTF::OrderMaker::iterator::operator*):
        (WTF::OrderMaker::iterator::operator++):
        (WTF::OrderMaker::iterator::operator==):
        (WTF::OrderMaker::iterator::operator!=):
        (WTF::OrderMaker::begin):
        (WTF::OrderMaker::end):
        (WTF::OrderMaker::newNode):
        * wtf/SentinelLinkedList.h:
        (WTF::BasicRawSentinelNode::isOnList):
        (WTF::BasicRawSentinelNode<T>::remove):
        (WTF::BasicRawSentinelNode<T>::prepend):
        (WTF::BasicRawSentinelNode<T>::append):
        (WTF::RawNode>::SentinelLinkedList):
        (WTF::RawNode>::push):
        (WTF::RawNode>::append):
        (WTF::RawNode>::remove):
        (WTF::RawNode>::prepend):
        (WTF::RawNode>::isOnList):

2016-01-25  Alex Christensen  <achristensen@webkit.org>

        [Win] Copy forwarding headers before building a project
        https://bugs.webkit.org/show_bug.cgi?id=153434

        Reviewed by Brent Fulgham.

        * wtf/CMakeLists.txt:
        * wtf/PlatformWin.cmake:

2016-01-20  Ryosuke Niwa  <rniwa@webkit.org>

        HTMLElement::nodeName should not upper case non-ASCII characters
        https://bugs.webkit.org/show_bug.cgi?id=153231

        Reviewed by Darin Adler.

        Added convertToASCIIUppercase to AtomicString, String, and StringImpl. 

        * wtf/text/AtomicString.cpp:
        (WTF::AtomicString::convertASCIICase): Generalized from convertToASCIILowercase.
        (WTF::AtomicString::convertToASCIILowercase):
        (WTF::AtomicString::convertToASCIIUppercase):
        * wtf/text/AtomicString.h:
        * wtf/text/StringImpl.cpp:
        (WTF::StringImpl::convertASCIICase): Generalized from convertToASCIILowercase.
        (WTF::StringImpl::convertToASCIILowercase):
        (WTF::StringImpl::convertToASCIIUppercase):
        * wtf/text/StringImpl.h:
        * wtf/text/WTFString.cpp:
        (WTF::String::convertToASCIIUppercase): Added.
        * wtf/text/WTFString.h:

2016-01-22  Chris Dumez  <cdumez@apple.com>

        Unreviewed attempt to fix the Windows build after r195452.

        * wtf/text/WTFString.h:

2016-01-22  Darin Adler  <darin@apple.com>

        Reduce use of equalIgnoringCase to just ignore ASCII case
        https://bugs.webkit.org/show_bug.cgi?id=153266

        Reviewed by Ryosuke Niwa.

        * wtf/ASCIICType.h:
        (WTF::isASCIIAlphaCaselessEqual): Loosened the assertion in this function
        so it can work with ASCII spaces, numeric digits, and some punctuation.
        Commented that we might want to change its name later.

        * wtf/Assertions.cpp:
        (WTFInitializeLogChannelStatesFromString): Use equalLettersIgnoringASCIICase.

        * wtf/text/AtomicString.h:
        (WTF::equalLettersIgnoringASCIICase): Added. Calls the version that takes a String.

        * wtf/text/StringCommon.h:
        (WTF::equalLettersIgnoringASCIICase): Added. Takes a pointer to characters.
        (WTF::equalLettersIgnoringASCIICaseCommonWithoutLength): Added.
        (WTF::equalLettersIgnoringASCIICaseCommon): Added.

        * wtf/text/StringImpl.h:
        (WTF::equalLettersIgnoringASCIICase): Added. Calls equalLettersIgnoringASCIICaseCommon.

        * wtf/text/StringView.h:
        (WTF::equalLettersIgnoringASCIICase): Added. Calls equalLettersIgnoringASCIICaseCommon.

        * wtf/text/WTFString.h: Reordered/reformatted a little.
        (WTF::equalIgnoringASCIICase): Added. Calls the version that takes a StringImpl.

2016-01-22  Csaba Osztrogonác  <ossy@webkit.org>

        Buildfix for older GCCs after r195142
        https://bugs.webkit.org/show_bug.cgi?id=153351

        Unreviewed buildfix.

        * wtf/SystemTracing.h:

2016-01-21  Filip Pizlo  <fpizlo@apple.com>

        REGRESSION(r195417): many tests crash
        https://bugs.webkit.org/show_bug.cgi?id=153316

        Reviewed by Saam Barati.

        This rolls out the StdLibExtras.h change, and simplifies RangeSet to not use binary search.
        That's fine for now, since B3 doesn't stress RangeSet enough right now.

        * wtf/RangeSet.h:
        (WTF::RangeSet::contains):
        (WTF::RangeSet::overlaps):
        (WTF::RangeSet::clear):
        (WTF::RangeSet::findRange):
        * wtf/StdLibExtras.h:
        (WTF::binarySearchImpl):
        (WTF::binarySearch):
        (WTF::tryBinarySearch):
        (WTF::approximateBinarySearch):

2016-01-21  Filip Pizlo  <fpizlo@apple.com>

        B3 should have load elimination
        https://bugs.webkit.org/show_bug.cgi?id=153288

        Reviewed by Geoffrey Garen.

        I needed a way to track sets of ranges, where there is a high likelihood that all of the
        ranges overlap. So I created RangeSet. It's a usually-sorted list of coalesced ranges.
        Practically this means that right now, FTL B3 will end up with two kinds of range sets: a set
        that just contains top and a set that contains nothing. In the future, most sets will either
        be top of empty but some of them will contain a handful of other things.

        * WTF.xcodeproj/project.pbxproj:
        * wtf/CMakeLists.txt:
        * wtf/MathExtras.h:
        (WTF::leftShiftWithSaturation):
        (WTF::nonEmptyRangesOverlap):
        (WTF::rangesOverlap):
        * wtf/RangeSet.h: Added.
        (WTF::RangeSet::RangeSet):
        (WTF::RangeSet::~RangeSet):
        (WTF::RangeSet::add):
        (WTF::RangeSet::contains):
        (WTF::RangeSet::overlaps):
        (WTF::RangeSet::clear):
        (WTF::RangeSet::dump):
        (WTF::RangeSet::dumpRaw):
        (WTF::RangeSet::compact):
        (WTF::RangeSet::overlapsNonEmpty):
        (WTF::RangeSet::subsumesNonEmpty):
        (WTF::RangeSet::findRange):
        * wtf/StdLibExtras.h:
        (WTF::binarySearchImpl):
        (WTF::binarySearch):
        (WTF::tryBinarySearch):
        (WTF::approximateBinarySearch):

2016-01-19  Ada Chan  <adachan@apple.com>

        Make it possible to enable VIDEO_PRESENTATION_MODE on other Cocoa platforms.
        https://bugs.webkit.org/show_bug.cgi?id=153218

        Reviewed by Eric Carlson.

        * wtf/Platform.h:

2016-01-19  Filip Pizlo  <fpizlo@apple.com>

        B3 should have basic path specialization
        https://bugs.webkit.org/show_bug.cgi?id=153200

        Reviewed by Benjamin Poulain.

        * wtf/GraphNodeWorklist.h:
        (WTF::GraphNodeWorklist::push):
        (WTF::GraphNodeWorklist::pushAll):
        (WTF::GraphNodeWorklist::isEmpty):
        (WTF::GraphNodeWorklist::notEmpty):

2016-01-20  Said Abou-Hallawa  <sabouhallawa@apple.com>

        Refactor AtomicStringKeyedMRUCache to be a generic LRU cache
        https://bugs.webkit.org/show_bug.cgi?id=153109

        Reviewed by Darin Adler.

        Refactor AtomicStringKeyedMRUCache, move it to WTF project and rename it
        to be TinyLRUCache. Define another template and call it TinyLRUCachePolicy.
        This one can be overridden for different keys and values. Its function is 
        creating the cached values.

        * WTF.xcodeproj/project.pbxproj:
        * wtf/TinyLRUCache.h: Added.
        (WebCore::TinyLRUCachePolicy::isKeyNull):
        (WebCore::TinyLRUCachePolicy::createValueForNullKey):
        (WebCore::TinyLRUCachePolicy::createValueForKey):
        (WebCore::TinyLRUCache::get):

2016-01-19  Saam barati  <sbarati@apple.com>

        WTF::Bag should be non-copyable
        https://bugs.webkit.org/show_bug.cgi?id=153253

        Reviewed by Filip Pizlo.

        * wtf/Bag.h:
        * wtf/SegmentedVector.h:
        (WTF::SegmentedVector::append):
        (WTF::SegmentedVector::alloc):

2016-01-19  Enrica Casucci  <enrica@apple.com>

        Add support for DataDetectors in WK (iOS).
        https://bugs.webkit.org/show_bug.cgi?id=152989
        rdar://problem/22855960

        Reviewed by Tim Horton.

        Adding feature definition for data detection.

        * wtf/FeatureDefines.h:

2016-01-19  Commit Queue  <commit-queue@webkit.org>

        Unreviewed, rolling out r195300.
        https://bugs.webkit.org/show_bug.cgi?id=153244

        enrica wants more time to fix Windows (Requested by thorton on
        #webkit).

        Reverted changeset:

        "Add support for DataDetectors in WK (iOS)."
        https://bugs.webkit.org/show_bug.cgi?id=152989
        http://trac.webkit.org/changeset/195300

2016-01-19  Chris Dumez  <cdumez@apple.com>

        Unreviewed, rolling out r195141.

        Seems to cause crashes on iOS9 64bit

        Reverted changeset:

        "Fragmentation-free allocator for timeless and/or coupled
        allocations."
        https://bugs.webkit.org/show_bug.cgi?id=152696
        http://trac.webkit.org/changeset/195141

2016-01-19  Enrica Casucci  <enrica@apple.com>

        Add support for DataDetectors in WK (iOS).
        https://bugs.webkit.org/show_bug.cgi?id=152989
        rdar://problem/22855960

        Reviewed by Tim Horton.

        Adding feature definition for data detection.

        * wtf/FeatureDefines.h:

2016-01-17  Filip Pizlo  <fpizlo@apple.com>

        FTL B3 should be just as fast as FTL LLVM on Octane/crypto
        https://bugs.webkit.org/show_bug.cgi?id=153113

        Reviewed by Saam Barati.