Create HeapLinkedHashSet and LinkedHashSet, ordered heap-friendly hash sets.

Source/wtf/LinkedHashSet.h

Index: Source/wtf/LinkedHashSet.h
diff --git a/Source/wtf/LinkedHashSet.h b/Source/wtf/LinkedHashSet.h
new file mode 100644
index 0000000000000000000000000000000000000000..60ee62d7bac6e3c9e2f38a02b096abada6f3f854
--- /dev/null
+++ b/Source/wtf/LinkedHashSet.h
@@ -0,0 +1,717 @@
+/*
+ * Copyright (C) 2005, 2006, 2007, 2008, 2011, 2012 Apple Inc. All rights reserved.
+ * Copyright (C) 2011, Benjamin Poulain <ikipou@gmail.com>
+ *
+ * This library is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Library General Public
+ * License as published by the Free Software Foundation; either
+ * version 2 of the License, or (at your option) any later version.
+ *
+ * This library is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * Library General Public License for more details.
+ *
+ * You should have received a copy of the GNU Library General Public License
+ * along with this library; see the file COPYING.LIB.  If not, write to
+ * the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor,
+ * Boston, MA 02110-1301, USA.
+ *
+ */
+
+#ifndef WTF_LinkedHashSet_h
+#define WTF_LinkedHashSet_h
+
+#include "wtf/DefaultAllocator.h"
+#include "wtf/HashSet.h"
+#include "wtf/OwnPtr.h"
+#include "wtf/PassOwnPtr.h"
+
+namespace WTF {
+
+// LinkedHashSet: Just like HashSet, this class provides a Set
+// interface - a collection of unique objects with O(1) insertion,
+// removal and test for containership. However, it also has an
+// order - iterating it will always give back values in the order
+// in which they are added.
+
+// Unlike ListHashSet, but like most WTF collections, iteration is NOT safe
+// against mutation of the LinkedHashSet.
+
+template<typename Value, typename HashFunctions, typename HashTraits, typename Allocator> class LinkedHashSet;
+
+template<typename LinkedHashSet> class LinkedHashSetIterator;
+template<typename LinkedHashSet> class LinkedHashSetConstIterator;
+template<typename LinkedHashSet> class LinkedHashSetReverseIterator;
+template<typename LinkedHashSet> class LinkedHashSetConstReverseIterator;
+
+template<typename Value, typename HashFunctions> struct LinkedHashSetTranslator;
+template<typename Value> struct LinkedHashSetExtractor;
+template<typename Value, typename ValueTraits> struct LinkedHashSetTraits;
+
+class LinkedHashSetNodeBase {
+public:
+    LinkedHashSetNodeBase() : m_prev(this), m_next(this) { }
+
+    void unlink()
+    {
+        if (!m_next)
+            return;
+        ASSERT(m_prev);
+        ASSERT(m_next->m_prev == this);
+        ASSERT(m_prev->m_next == this);
+        m_next->m_prev = m_prev;
+        m_prev->m_next = m_next;
+    }
+
+    ~LinkedHashSetNodeBase()
+    {
+        unlink();
+    }
+
+    void insertBefore(LinkedHashSetNodeBase& other)
+    {
+        other.m_next = this;
+        other.m_prev = m_prev;
+        m_prev->m_next = &other;
+        m_prev = &other;
+        ASSERT(other.m_next);
+        ASSERT(other.m_prev);
+        ASSERT(m_next);
+        ASSERT(m_prev);
+    }
+
+    void insertAfter(LinkedHashSetNodeBase& other)
+    {
+        other.m_prev = this;
+        other.m_next = m_next;
+        m_next->m_prev = &other;
+        m_next = &other;
+        ASSERT(other.m_next);
+        ASSERT(other.m_prev);
+        ASSERT(m_next);
+        ASSERT(m_prev);
+    }
+
+    LinkedHashSetNodeBase(LinkedHashSetNodeBase* prev, LinkedHashSetNodeBase* next)
+        : m_prev(prev)
+        , m_next(next)
+    {
+        ASSERT((prev && next) || (!prev && !next));
+    }
+
+    LinkedHashSetNodeBase* m_prev;
+    LinkedHashSetNodeBase* m_next;
+
+protected:
+    // If we take a copy of a node we can't copy the next and prev pointers,
+    // since they point to something that does not point at us. This is used
+    // inside the shouldExpand() "if" in HashTable::add.
+    LinkedHashSetNodeBase(const LinkedHashSetNodeBase& other)
+        : m_prev(0)
+        , m_next(0) { }
+
+private:
+    // Should not be used.
+    LinkedHashSetNodeBase& operator=(const LinkedHashSetNodeBase& other);
+};
+
+template<typename ValueArg>
+class LinkedHashSetNode : public LinkedHashSetNodeBase {
+public:
+    LinkedHashSetNode(const ValueArg& value, LinkedHashSetNodeBase* prev, LinkedHashSetNodeBase* next)
+        : LinkedHashSetNodeBase(prev, next)
+        , m_value(value)
+    {
+    }
+
+    ValueArg m_value;
+
+private:
+    // Not used.
+    LinkedHashSetNode(const LinkedHashSetNode&);

[Mikhail, 2014/04/10 12:25:03]

WTF_MAKE_NONCOPYABLE ?

[Erik Corry, 2014/04/10 13:47:29]

Can't do that because it also overwrites the copy constructor, which I have
customized.

+};
+
+template<
+    typename ValueArg,
+    typename HashFunctions = typename DefaultHash<ValueArg>::Hash,
+    typename TraitsArg = HashTraits<ValueArg>,
+    typename Allocator = DefaultAllocator>
+class LinkedHashSet {
+private:
+    typedef ValueArg Value;
+    typedef TraitsArg Traits;
+    typedef LinkedHashSetNode<Value> Node;
+    typedef LinkedHashSetNodeBase NodeBase;
+    typedef LinkedHashSetTranslator<Value, HashFunctions> NodeHashFunctions;
+    typedef LinkedHashSetTraits<Value, Traits> NodeHashTraits;
+
+    typedef HashTable<Node, Node, IdentityExtractor,
+        NodeHashFunctions, NodeHashTraits, NodeHashTraits, Allocator> ImplType;
+
+public:
+    typedef LinkedHashSetIterator<LinkedHashSet> iterator;
+    friend class LinkedHashSetIterator<LinkedHashSet>;
+    typedef LinkedHashSetConstIterator<LinkedHashSet> const_iterator;
+    friend class LinkedHashSetConstIterator<LinkedHashSet>;
+
+    typedef LinkedHashSetReverseIterator<LinkedHashSet> reverse_iterator;
+    friend class LinkedHashSetReverseIterator<LinkedHashSet>;
+    typedef LinkedHashSetConstReverseIterator<LinkedHashSet> const_reverse_iterator;
+    friend class LinkedHashSetConstReverseIterator<LinkedHashSet>;
+
+    struct AddResult {
+        AddResult(const typename ImplType::AddResult& hashTableAddResult)
+            : storedValue(&hashTableAddResult.storedValue->m_value)
+            , isNewEntry(hashTableAddResult.isNewEntry)
+        {
+        }
+
+        Value* storedValue;
+        bool isNewEntry;
+    };
+
+    typedef typename HashTraits<Value>::PeekInType ValuePeekInType;
+
+    void* operator new(size_t size)

[Mads Ager (chromium), 2014/04/10 11:12:25]

Did we now have a macro for this?

[Erik Corry, 2014/04/10 13:47:29]

Done.

+    {
+        return Allocator::template malloc<void*, LinkedHashSet>(size);
+    }
+    void operator delete(void* p) { Allocator::free(p); }
+    void* operator new[](size_t size) { return Allocator::template newArray<LinkedHashSet>(size); }
+    void operator delete[](void* p) { Allocator::deleteArray(p); }
+    void* operator new(size_t, NotNullTag, void* location)
+    {
+        COMPILE_ASSERT(!Allocator::isGarbageCollected, Garbage_collector_must_be_disabled);
+        ASSERT(location);
+        return location;
+    }
+
+    LinkedHashSet();
+    LinkedHashSet(const LinkedHashSet&);
+    LinkedHashSet& operator=(const LinkedHashSet&);
+    ~LinkedHashSet();
+
+    static void finalize(void* pointer) { reinterpret_cast<LinkedHashSet*>(pointer)->~LinkedHashSet(); }
+
+    void swap(LinkedHashSet&);
+
+    unsigned size() const { return m_impl.size(); }
+    unsigned capacity() const { return m_impl.capacity(); }
+    bool isEmpty() const { return m_impl.isEmpty(); }
+
+    iterator begin() { return makeIterator(firstNode()); }
+    iterator end() { return makeIterator(anchor()); }
+    const_iterator begin() const { return makeConstIterator(firstNode()); }
+    const_iterator end() const { return makeConstIterator(anchor()); }
+
+    reverse_iterator rbegin() { return makeReverseIterator(lastNode()); }
+    reverse_iterator rend() { return makeReverseIterator(anchor()); }
+    const_reverse_iterator rbegin() const { return makeConstReverseIterator(lastNode()); }
+    const_reverse_iterator rend() const { return makeConstReverseIterator(anchor()); }
+
+    Value& first();
+    const Value& first() const;
+    void removeFirst();
+
+    Value& last();
+    const Value& last() const;
+    void removeLast();
+
+    iterator find(ValuePeekInType);
+    const_iterator find(ValuePeekInType) const;
+    bool contains(ValuePeekInType) const;
+
+    // An alternate version of find() that finds the object by hashing and comparing
+    // with some other type, to avoid the cost of type conversion.
+    // The HashTranslator interface is defined in HashSet.
+    template<typename HashTranslator, typename T> iterator find(const T&);
+    template<typename HashTranslator, typename T> const_iterator find(const T&) const;
+    template<typename HashTranslator, typename T> bool contains(const T&) const;
+
+    // The return value of add is a pair of a pointer to the stored value,
+    // and a bool that is true if an new entry was added.
+    AddResult add(ValuePeekInType);
+
+    // Same as add() except that the return value is an
+    // iterator. Useful in cases where it's needed to have the
+    // same return value as find() and where it's not possible to
+    // use a pointer to the storedValue.
+    iterator addReturnIterator(ValuePeekInType);
+
+    // Add the value to the end of the collection. If the value was already in
+    // the list, it is moved to the end.
+    AddResult appendOrMoveToLast(ValuePeekInType);
+
+    // Add the value to the beginning of the collection. If the value was already in
+    // the list, it is moved to the beginning.
+    AddResult prependOrMoveToFirst(ValuePeekInType);
+
+    AddResult insertBefore(ValuePeekInType beforeValue, ValuePeekInType newValue);
+    AddResult insertBefore(iterator it, ValuePeekInType newValue) { return m_impl.template add<NodeHashFunctions>(newValue, it.node()); }
+
+    void remove(ValuePeekInType);
+    void remove(iterator);
+    void clear() { m_impl.clear(); }
+
+    void trace(typename Allocator::Visitor* visitor) { m_impl.trace(visitor); }
+
+    int64_t modifications() const { return m_impl.modifications(); }
+    void checkModifications(int64_t mods) const { m_impl.checkModifications(mods); }

[Mikhail, 2014/04/10 12:25:03]

do we need those in release mode (or you think they will be optimized out)?

[Erik Corry, 2014/04/10 13:47:29]

This one is only called from an assert.
The one with no arguments is an empty function in release mode.

+
+private:
+    Node* anchor() { return reinterpret_cast<Node*>(&m_anchor); }
+    const Node* anchor() const { return reinterpret_cast<const Node*>(&m_anchor); }
+    Node* firstNode() { return reinterpret_cast<Node*>(m_anchor.m_next); }
+    const Node* firstNode() const { return reinterpret_cast<const Node*>(m_anchor.m_next); }
+    Node* lastNode() { return reinterpret_cast<Node*>(m_anchor.m_prev); }
+    const Node* lastNode() const { return reinterpret_cast<const Node*>(m_anchor.m_prev); }
+
+    iterator makeIterator(const Node* position) { return iterator(position, this); }
+    const_iterator makeConstIterator(const Node* position) const { return const_iterator(position, this); }
+    reverse_iterator makeReverseIterator(const Node* position) { return reverse_iterator(position, this); }
+    const_reverse_iterator makeConstReverseIterator(const Node* position) const { return const_reverse_iterator(position, this); }
+
+    ImplType m_impl;
+    NodeBase m_anchor;
+#ifndef ASSERT_ENABLED
+    uint64_t m_modifications;
+#endif
+};
+
+template<typename Value, typename HashFunctions>
+struct LinkedHashSetTranslator {
+    typedef LinkedHashSetNode<Value> Node;
+    typedef LinkedHashSetNodeBase NodeBase;
+    typedef typename HashTraits<Value>::PeekInType ValuePeekInType;
+    static unsigned hash(const Node& node) { return HashFunctions::hash(node.m_value); }
+    static unsigned hash(const ValuePeekInType& key) { return HashFunctions::hash(key); }
+    static bool equal(const Node& a, const ValuePeekInType& b) { return HashFunctions::equal(a.m_value, b); }
+    static bool equal(const Node& a, const Node& b) { return HashFunctions::equal(a.m_value, b.m_value); }
+    static void translate(Node& location, ValuePeekInType key, NodeBase* anchor)
+    {
+        location.m_value = key;
+        anchor->insertBefore(location);
+    }
+
+    // Empty (or deleted) slots have the m_next pointer set to null, but we
+    // don't do anything to the other fields, which may contain junk.
+    // Therefore you can't compare a newly constructed empty value with a
+    // slot and get the right answer.
+    static const bool safeToCompareToEmptyOrDeleted = false;
+};
+
+template<typename Value>
+struct LinkedHashSetExtractor {
+    static const Value& extract(const LinkedHashSetNode<Value>& node) { return node.m_value; }
+};
+
+template<typename Value, typename ValueTraitsArg>
+struct LinkedHashSetTraits : public SimpleClassHashTraits<LinkedHashSetNode<Value> > {
+    typedef LinkedHashSetNode<Value> Node;
+    typedef ValueTraitsArg ValueTraits;
+
+    // The slot is empty when the m_next field is zero so it's safe to zero
+    // the backing.
+    static const bool emptyValueIsZero = true;
+
+    static const bool hasIsEmptyValueFunction = true;
+    static bool isEmptyValue(const Node& value) { return !value.m_next; }
+
+    static const int deletedValue = -1;
+
+    static void constructDeletedValue(Node& slot) { slot.m_next = reinterpret_cast<Node*>(deletedValue); }
+    static bool isDeletedValue(const Node& slot) { return slot.m_next == reinterpret_cast<Node*>(deletedValue); }
+
+    // We always need to call destructors, that's how we get linked and
+    // unlinked from the chain.
+    static const bool needsDestruction = true;
+
+    // Whether we need to trace and do weak processing depends on the traits of
+    // the type inside the node.
+    template<typename U = void>
+    struct NeedsTracingLazily {
+        static const bool value = ValueTraits::template NeedsTracingLazily<>::value;
+    };
+    static const bool isWeak = ValueTraits::isWeak;
+};
+
+template<typename LinkedHashSetType>
+class LinkedHashSetIterator {
+private:
+    typedef typename LinkedHashSetType::Node Node;
+    typedef typename LinkedHashSetType::Traits Traits;
+
+    typedef typename LinkedHashSetType::Value& ReferenceType;
+    typedef typename LinkedHashSetType::Value* PointerType;
+
+    typedef LinkedHashSetConstIterator<LinkedHashSetType> const_iterator;
+
+    Node* node() const { return const_cast<Node*>(m_iterator.node()); }

[Mikhail, 2014/04/10 12:25:03]

looks like it should not be a const method

[Erik Corry, 2014/04/10 13:47:29]

Done.

+
+protected:
+    LinkedHashSetIterator(const Node* position, LinkedHashSetType* m_container)
+        : m_iterator(position , m_container)
+    {
+    }
+
+public:
+    // Default copy, assignment and destructor are OK.
+
+    PointerType get() const { return const_cast<PointerType>(m_iterator.get()); }
+    ReferenceType operator*() const { return *get(); }
+    PointerType operator->() const { return get(); }
+
+    LinkedHashSetIterator& operator++() { ++m_iterator; return *this; }
+    LinkedHashSetIterator& operator--() { --m_iterator; return *this; }
+
+    // Postfix ++ and -- intentionally omitted.
+
+    // Comparison.
+    bool operator==(const LinkedHashSetIterator& other) const { return m_iterator == other.m_iterator; }
+    bool operator!=(const LinkedHashSetIterator& other) const { return m_iterator != other.m_iterator; }
+
+    operator const_iterator() const { return m_iterator; }
+
+protected:
+    const_iterator m_iterator;
+    template<typename T, typename U, typename V, typename W> friend class LinkedHashSet;
+};
+
+template<typename LinkedHashSetType>
+class LinkedHashSetConstIterator {
+private:
+    typedef typename LinkedHashSetType::Node Node;
+    typedef typename LinkedHashSetType::Traits Traits;
+
+    typedef const typename LinkedHashSetType::Value& ReferenceType;
+    typedef const typename LinkedHashSetType::Value* PointerType;
+
+    const Node* node() const { return static_cast<const Node*>(m_position); }
+
+protected:
+    LinkedHashSetConstIterator(const LinkedHashSetNodeBase* position, const LinkedHashSetType* container)
+        : m_position(position)
+#ifdef ASSERT_ENABLED
+        , m_container(container)
+        , m_containerModifications(container->modifications())
+#endif
+    {
+    }
+
+public:
+    PointerType get() const
+    {
+        checkModifications();
+        return &static_cast<const Node*>(m_position)->m_value;
+    }
+    ReferenceType operator*() const { return *get(); }
+    PointerType operator->() const { return get(); }
+
+    LinkedHashSetConstIterator& operator++()
+    {
+        ASSERT(m_position);
+        checkModifications();
+        m_position = m_position->m_next;
+        return *this;
+    }
+
+    LinkedHashSetConstIterator& operator--()
+    {
+        ASSERT(m_position);
+        checkModifications();
+        m_position = m_position->m_prev;
+        return *this;
+    }
+
+    // Postfix ++ and -- intentionally omitted.
+
+    // Comparison.
+    bool operator==(const LinkedHashSetConstIterator& other) const
+    {
+        return m_position == other.m_position;
+    }
+    bool operator!=(const LinkedHashSetConstIterator& other) const
+    {
+        return m_position != other.m_position;
+    }
+
+private:
+    const LinkedHashSetNodeBase* m_position;
+#ifdef ASSERT_ENABLED
+    void checkModifications() const { m_container->checkModifications(m_containerModifications); }
+    const LinkedHashSetType* m_container;
+    int64_t m_containerModifications;
+#else
+    void checkModifications() const { }
+#endif
+    template<typename T, typename U, typename V, typename W> friend class LinkedHashSet;
+    friend class LinkedHashSetIterator<LinkedHashSetType>;
+};
+
+template<typename LinkedHashSetType>
+class LinkedHashSetReverseIterator : public LinkedHashSetIterator<LinkedHashSetType> {
+    typedef LinkedHashSetIterator<LinkedHashSetType> Superclass;
+    typedef LinkedHashSetConstReverseIterator<LinkedHashSetType> const_reverse_iterator;
+    typedef typename LinkedHashSetType::Node Node;
+
+protected:
+    LinkedHashSetReverseIterator(const Node* position, LinkedHashSetType* container)
+        : Superclass(position, container) { }
+
+public:
+    LinkedHashSetReverseIterator& operator++() { Superclass::operator--(); return *this; }
+    LinkedHashSetReverseIterator& operator--() { Superclass::operator++(); return *this; }
+
+    // Postfix ++ and -- intentionally omitted.
+
+    operator const_reverse_iterator() const { return *reinterpret_cast<const_reverse_iterator*>(this); }
+
+    template<typename T, typename U, typename V, typename W> friend class LinkedHashSet;
+};
+
+template<typename LinkedHashSetType>
+class LinkedHashSetConstReverseIterator : public LinkedHashSetConstIterator<LinkedHashSetType> {
+    typedef LinkedHashSetConstIterator<LinkedHashSetType> Superclass;
+    typedef typename LinkedHashSetType::Node Node;
+
+public:
+    LinkedHashSetConstReverseIterator(const Node* position, const LinkedHashSetType* container)
+        : Superclass(position, container) { }
+
+    LinkedHashSetConstReverseIterator& operator++() { Superclass::operator--(); return *this; }
+    LinkedHashSetConstReverseIterator& operator--() { Superclass::operator++(); return *this; }
+
+    // Postfix ++ and -- intentionally omitted.
+
+    template<typename T, typename U, typename V, typename W> friend class LinkedHashSet;
+};
+
+template<typename T, typename U, typename V, typename W>
+inline LinkedHashSet<T, U, V, W>::LinkedHashSet() { }
+
+template<typename T, typename U, typename V, typename W>
+inline LinkedHashSet<T, U, V, W>::LinkedHashSet(const LinkedHashSet& other)
+    : m_anchor()
+{
+    const_iterator end = other.end();
+    for (const_iterator it = other.begin(); it != end; ++it)
+        add(*it);
+}
+
+template<typename T, typename U, typename V, typename W>
+inline LinkedHashSet<T, U, V, W>& LinkedHashSet<T, U, V, W>::operator=(const LinkedHashSet& other)
+{
+    LinkedHashSet tmp(other);
+    swap(tmp);
+    return *this;
+}
+
+template<typename T, typename U, typename V, typename W>
+inline void LinkedHashSet<T, U, V, W>::swap(LinkedHashSet& other)
+{
+    m_impl.swap(other.m_impl);
+    swap(m_anchor, other.m_anchor);
+}
+
+template<typename T, typename U, typename V, typename Allocator>
+inline LinkedHashSet<T, U, V, Allocator>::~LinkedHashSet()
+{
+    // The destructor of m_anchor will implicitly be called here, which will
+    // unlink the anchor from the collection.
+}
+
+template<typename T, typename U, typename V, typename W>
+inline T& LinkedHashSet<T, U, V, W>::first()
+{
+    ASSERT(!isEmpty());
+    return firstNode()->m_value;
+}
+
+template<typename T, typename U, typename V, typename W>
+inline const T& LinkedHashSet<T, U, V, W>::first() const
+{
+    ASSERT(!isEmpty());
+    return firstNode()->m_value;
+}
+
+template<typename T, typename U, typename V, typename W>
+inline void LinkedHashSet<T, U, V, W>::removeFirst()
+{
+    ASSERT(!isEmpty());
+    m_impl.remove(static_cast<Node*>(m_anchor.m_next));
+}
+
+template<typename T, typename U, typename V, typename W>
+inline T& LinkedHashSet<T, U, V, W>::last()
+{
+    ASSERT(!isEmpty());
+    return lastNode()->m_value;
+}
+
+template<typename T, typename U, typename V, typename W>
+inline const T& LinkedHashSet<T, U, V, W>::last() const
+{
+    ASSERT(!isEmpty());
+    return lastNode()->m_value;
+}
+
+template<typename T, typename U, typename V, typename W>
+inline void LinkedHashSet<T, U, V, W>::removeLast()
+{
+    ASSERT(!isEmpty());
+    m_impl.remove(static_cast<Node*>(m_anchor.m_prev));
+}
+
+template<typename T, typename U, typename V, typename W>
+inline typename LinkedHashSet<T, U, V, W>::iterator LinkedHashSet<T, U, V, W>::find(ValuePeekInType value)
+{
+    LinkedHashSet::Node* node = m_impl.template lookup<LinkedHashSet::NodeHashFunctions, ValuePeekInType>(value);
+    if (!node)
+        return end();
+    return makeIterator(node);
+}
+
+template<typename T, typename U, typename V, typename W>
+inline typename LinkedHashSet<T, U, V, W>::const_iterator LinkedHashSet<T, U, V, W>::find(ValuePeekInType value) const
+{
+    const LinkedHashSet::Node* node = m_impl.template lookup<LinkedHashSet::NodeHashFunctions, ValuePeekInType>(value);
+    if (!node)
+        return end();
+    return makeConstIterator(node);
+}
+
+template<typename Value, typename U, typename V, typename W>
+template<typename HashTranslator, typename T>
+inline typename LinkedHashSet<Value, U, V, W>::iterator LinkedHashSet<Value, U, V, W>::find(const T& value)
+{
+    iterator it = begin();
+    iterator endIterator = end();

[Erik Corry, 2014/04/10 10:53:06]

I need to fix this implementation to not loop.

[Erik Corry, 2014/04/10 13:47:29]

Done.

+    while (it != endIterator && !HashTranslator::template equal<Value, T>(it->m_value, value))
+        ++it;
+    return it;
+}
+
+template<typename Value, typename U, typename V, typename W>
+template<typename HashTranslator, typename T>
+inline typename LinkedHashSet<Value, U, V, W>::const_iterator LinkedHashSet<Value, U, V, W>::find(const T& value) const
+{
+    const_iterator it = begin();
+    const_iterator endIterator = end();
+    while (it != endIterator && !HashTranslator::template equal<Value, T>(it->m_value, value))
+        ++it;
+    return it;
+}
+
+template<typename Translator>
+struct LinkedHashSetTranslatorAdapter {
+    template<typename T> static unsigned hash(const T& key) { return Translator::hash(key); }
+    template<typename T, typename U> static bool equal(const T& a, const U& b) { return Translator::equal(a->m_value, b); }
+};
+
+template<typename Value, typename U, typename V, typename W>
+template<typename HashTranslator, typename T>
+inline bool LinkedHashSet<Value, U, V, W>::contains(const T& value) const
+{
+    return m_impl.template contains<LinkedHashSetTranslatorAdapter<HashTranslator> >(value);
+}
+
+template<typename T, typename U, typename V, typename W>
+inline bool LinkedHashSet<T, U, V, W>::contains(ValuePeekInType value) const
+{
+    return m_impl.template contains<NodeHashFunctions>(value);
+}
+
+template<typename Value, typename HashFunctions, typename Traits, typename Allocator>
+typename LinkedHashSet<Value, HashFunctions, Traits, Allocator>::AddResult LinkedHashSet<Value, HashFunctions, Traits, Allocator>::add(ValuePeekInType value)
+{
+    return m_impl.template add<NodeHashFunctions>(value, &m_anchor);
+}
+
+template<typename T, typename U, typename V, typename W>
+typename LinkedHashSet<T, U, V, W>::iterator LinkedHashSet<T, U, V, W>::addReturnIterator(ValuePeekInType value)
+{
+    typename ImplType::AddResult result = m_impl.template add<NodeHashFunctions>(value, &m_anchor);
+    return makeIterator(result.storedValue);
+}
+
+template<typename T, typename U, typename V, typename W>
+typename LinkedHashSet<T, U, V, W>::AddResult LinkedHashSet<T, U, V, W>::appendOrMoveToLast(ValuePeekInType value)
+{
+    typename ImplType::AddResult result = m_impl.template add<NodeHashFunctions>(value, &m_anchor);
+    Node* node = result.storedValue;
+    if (!result.isNewEntry) {
+        node->unlink();
+        m_anchor.insertBefore(*node);
+    }
+    return result;
+}
+
+template<typename T, typename U, typename V, typename W>
+typename LinkedHashSet<T, U, V, W>::AddResult LinkedHashSet<T, U, V, W>::prependOrMoveToFirst(ValuePeekInType value)
+{
+    typename ImplType::AddResult result = m_impl.template add<NodeHashFunctions>(value, m_anchor.m_next);
+    Node* node = result.storedValue;
+    if (!result.isNewEntry) {
+        node->unlink();
+        m_anchor.insertAfter(*node);
+    }
+    return result;
+}
+
+template<typename T, typename U, typename V, typename W>
+typename LinkedHashSet<T, U, V, W>::AddResult LinkedHashSet<T, U, V, W>::insertBefore(ValuePeekInType beforeValue, ValuePeekInType newValue)
+{
+    return insertBefore(find(beforeValue), newValue);
+}
+
+template<typename T, typename U, typename V, typename W>
+inline void LinkedHashSet<T, U, V, W>::remove(iterator it)
+{
+    if (it == end())
+        return;
+    m_impl.remove(it.node());
+}
+
+template<typename T, typename U, typename V, typename W>
+inline void LinkedHashSet<T, U, V, W>::remove(ValuePeekInType value)
+{
+    remove(find(value));
+}
+
+template<typename T>
+struct IsWeak<LinkedHashSetNode<T> > {
+    static const bool value = IsWeak<T>::value;
+};
+
+inline void swap(LinkedHashSetNodeBase& a, LinkedHashSetNodeBase& b)
+{
+    typedef LinkedHashSetNodeBase Base;
+    swap(a.m_next, b.m_next);
+    swap(a.m_prev, b.m_prev);
+    if (b.m_next) {
+        b.m_next->m_prev = &b;
+        b.m_prev->m_next = &b;
+    }
+    if (a.m_next) {
+        a.m_next->m_prev = &a;
+        a.m_prev->m_next = &a;
+    }
+}
+
+template<typename T>
+inline void swap(LinkedHashSetNode<T>& a, LinkedHashSetNode<T>& b)
+{
+    typedef LinkedHashSetNodeBase Base;
+
+    swap(static_cast<Base&>(a), static_cast<Base&>(b));
+
+    using namespace std;

[Mads Ager (chromium), 2014/04/10 11:12:25]

I thought we were not doing this?

[Erik Corry, 2014/04/10 13:47:29]

I was a bit confused too, but the check-webkit-style script says

Source/wtf/LinkedHashSet.h:696:  Use 'using namespace std;' instead of 'using
std::swap;'.  [build/using_std] [4]

[Erik Corry, 2014/04/10 13:53:07]

Or I could just remove the using directive completely, which is what I did now.

+    swap(a.m_value, b.m_value);
+}
+
+}
+
+using WTF::LinkedHashSet;
+
+#endif /* WTF_LinkedHashSet_h */