zoukankan      html  css  js  c++  java
  • stl_tree.h

    stl_tree.h
    G++ 2.91.57,cygnuscygwin-b20includeg++stl_tree.h 完整列表
    /*
     *
     * Copyright (c) 1996,1997
     * Silicon Graphics Computer Systems, Inc.
     *
     * Permission to use, copy, modify, distribute and sell this software
     * and its documentation for any purpose is hereby granted without fee,
     * provided that the above copyright notice appear in all copies and
     * that both that copyright notice and this permission notice appear
     * in supporting documentation.  Silicon Graphics makes no
     * representations about the suitability of this software for any
     * purpose.  It is provided "as is" without express or implied warranty.
     *
     *
     * Copyright (c) 1994
     * Hewlett-Packard Company
     *
     * Permission to use, copy, modify, distribute and sell this software
     * and its documentation for any purpose is hereby granted without fee,
     * provided that the above copyright notice appear in all copies and
     * that both that copyright notice and this permission notice appear
     * in supporting documentation.  Hewlett-Packard Company makes no
     * representations about the suitability of this software for any
     * purpose.  It is provided "as is" without express or implied warranty.
     *
     *
     */
    
    /* NOTE: This is an internal header file, included by other STL headers.
     *   You should not attempt to use it directly.
     */
    
    #ifndef __SGI_STL_INTERNAL_TREE_H
    #define __SGI_STL_INTERNAL_TREE_H
    
    /*
    本檔實作Red-black tree(紅-黑樹)class,用以實作 STL 關聯式容器(如set, 
    multiset, map, multimap)。所用之insertion 和deletion 演算法係以
    Cormen, Leiserson 和 Rivest 所著之 Introduction to Algorithms
    (MIT Press, 1990) 一書為基礎,唯以下兩點不同:
    
    (1) header 不僅指向 root,也指向紅黑樹的最左節點,以便實作出常數時間之
    begin();並且也指向紅黑樹的最右節點,以便set 相關泛型演算法(如set_union 
    等等)有線性時間之表現。
    
    (2) 當一個即將被刪除之節點擁有兩個子節點時,它的successor node is
    relinked into its place, rather than copied, 如此一來唯一失效(invalidated)的迭代器就只是那些referring to the deleted node.
    */
    
    #include <stl_algobase.h>
    #include <stl_alloc.h>
    #include <stl_construct.h>
    #include <stl_function.h>
    
    __STL_BEGIN_NAMESPACE 
    
    typedef bool __rb_tree_color_type;
    const __rb_tree_color_type __rb_tree_red = false;     // 紅色為 0
    const __rb_tree_color_type __rb_tree_black = true; // 黑色為 1
    
    struct __rb_tree_node_base
    {
      typedef __rb_tree_color_type color_type;
      typedef __rb_tree_node_base* base_ptr;
    
      color_type color;     // 節點顏色,非紅即黑。
      base_ptr parent;      // RB 樹的許多操作,必須知道父節點。
      base_ptr left;          // 指向左節點。
      base_ptr right;       // 指向右節點。
    
      static base_ptr minimum(base_ptr x)
      {
        while (x->left != 0) x = x->left;    // 一直向左走,就會找到最小值,
        return x;                            // 這是二元搜尋樹的特性。
      }
    
      static base_ptr maximum(base_ptr x)
      {
        while (x->right != 0) x = x->right;     // 一直向右走,就會找到最大值,
        return x;                            // 這是二元搜尋樹的特性。
      }
    };
    
    template <class Value>
    struct __rb_tree_node : public __rb_tree_node_base
    {
      typedef __rb_tree_node<Value>* link_type;
      Value value_field;    // 節點實值
    };
    
    struct __rb_tree_base_iterator
    {
      typedef __rb_tree_node_base::base_ptr base_ptr;
      typedef bidirectional_iterator_tag iterator_category;
      typedef ptrdiff_t difference_type;
    
      base_ptr node;    // 它用來與容器之間產生一個連結關係(make a reference)
    
      // 以下其實可實作於 operator++ 內,因為再無他處會呼叫此函式了。
      void increment()
      {
        if (node->right != 0) {        // 如果有右子節點。狀況(1)
          node = node->right;        // 就向右走
          while (node->left != 0)    // 然後一直往左子樹走到底
            node = node->left;        // 即是解答
        }
        else {                    // 沒有右子節點。狀況(2)
          base_ptr y = node->parent;    // 找出父節點
          while (node == y->right) {    // 如果現行節點本身是個右子節點,
            node = y;                // 就一直上溯,直到「不為右子節點」止。
            y = y->parent;
          }
          if (node->right != y)        // 「若此時的右子節點不等於此時的父節點」。
            node = y;                // 狀況(3) 此時的父節點即為解答。
                                          // 否則此時的node 為解答。狀況(4)
        }                        
        // 注意,以上判斷「若此時的右子節點不等於此時的父節點」,是為了應付一種
        // 特殊情況:我們欲尋找根節點的下一節點,而恰巧根節點無右子節點。
        // 當然,以上特殊作法必須配合 RB-tree 根節點與特殊之header 之間的
        // 特殊關係。
      }
    
      // 以下其實可實作於 operator-- 內,因為再無他處會呼叫此函式了。
      void decrement()
      {
        if (node->color == __rb_tree_red &&    // 如果是紅節點,且
            node->parent->parent == node)        // 父節點的父節點等於自己,
          node = node->right;                // 狀況(1) 右子節點即為解答。
        // 以上情況發生於node為header時(亦即 node 為 end() 時)。
        // 注意,header 之右子節點即 mostright,指向整棵樹的 max 節點。
        else if (node->left != 0) {            // 如果有左子節點。狀況(2)
          base_ptr y = node->left;            // 令y指向左子節點
          while (y->right != 0)                // 當y有右子節點時
            y = y->right;                    // 一直往右子節點走到底
          node = y;                        // 最後即為答案
        }
        else {                            // 既非根節點,亦無左子節點。
          base_ptr y = node->parent;            // 狀況(3) 找出父節點
          while (node == y->left) {            // 當現行節點身為左子節點
            node = y;                        // 一直交替往上走,直到現行節點
            y = y->parent;                    // 不為左子節點
          }
          node = y;                        // 此時之父節點即為答案
        }
      }
    };
    
    template <class Value, class Ref, class Ptr>
    struct __rb_tree_iterator : public __rb_tree_base_iterator
    {
      typedef Value value_type;
      typedef Ref reference;
      typedef Ptr pointer;
      typedef __rb_tree_iterator<Value, Value&, Value*>     iterator;
      typedef __rb_tree_iterator<Value, const Value&, const Value*> const_iterator;
      typedef __rb_tree_iterator<Value, Ref, Ptr>   self;
      typedef __rb_tree_node<Value>* link_type;
    
      __rb_tree_iterator() {}
      __rb_tree_iterator(link_type x) { node = x; }
      __rb_tree_iterator(const iterator& it) { node = it.node; }
    
      reference operator*() const { return link_type(node)->value_field; }
    #ifndef __SGI_STL_NO_ARROW_OPERATOR
      pointer operator->() const { return &(operator*()); }
    #endif /* __SGI_STL_NO_ARROW_OPERATOR */
    
      self& operator++() { increment(); return *this; }
      self operator++(int) {
        self tmp = *this;
        increment();
        return tmp;
      }
        
      self& operator--() { decrement(); return *this; }
      self operator--(int) {
        self tmp = *this;
        decrement();
        return tmp;
      }
    };
    
    inline bool operator==(const __rb_tree_base_iterator& x,
                           const __rb_tree_base_iterator& y) {
      return x.node == y.node;
      // 兩個迭代器相等,意指其所指的節點相等。
    }
    
    inline bool operator!=(const __rb_tree_base_iterator& x,
                           const __rb_tree_base_iterator& y) {
      return x.node != y.node;
      // 兩個迭代器不等,意指其所指的節點不等。
    }
    
    #ifndef __STL_CLASS_PARTIAL_SPECIALIZATION
    
    inline bidirectional_iterator_tag
    iterator_category(const __rb_tree_base_iterator&) {
      return bidirectional_iterator_tag();
    }
    
    inline __rb_tree_base_iterator::difference_type*
    distance_type(const __rb_tree_base_iterator&) {
      return (__rb_tree_base_iterator::difference_type*) 0;
    }
    
    template <class Value, class Ref, class Ptr>
    inline Value* value_type(const __rb_tree_iterator<Value, Ref, Ptr>&) {
      return (Value*) 0;
    }
    
    #endif /* __STL_CLASS_PARTIAL_SPECIALIZATION */
    
    // 以下都是全域函式:__rb_tree_rotate_left(), __rb_tree_rotate_right(),
    // __rb_tree_rebalance(), __rb_tree_rebalance_for_erase()
    
    // 新節點必為紅節點。如果安插處之父節點亦為紅節點,就違反紅黑樹規則,此時必須
    // 做樹形旋轉(及顏色改變,在程式它處)。
    inline void 
    __rb_tree_rotate_left(__rb_tree_node_base* x, __rb_tree_node_base*& root)
    {
      // x 為旋轉點
      __rb_tree_node_base* y = x->right;    // 令y 為旋轉點的右子節點
      x->right = y->left;
      if (y->left !=0)
        y->left->parent = x;        // 別忘了回馬槍設定父節點
      y->parent = x->parent;
    
      // 令 y 完全頂替 x 的地位(必須將 x 對其父節點的關係完全接收過來)
      if (x == root)                    // x 為根節點
        root = y;
      else if (x == x->parent->left)    // x 為其父節點的左子節點
        x->parent->left = y;
      else                            // x 為其父節點的右子節點
        x->parent->right = y;            
      y->left = x;
      x->parent = y;
    }
    
    // 新節點必為紅節點。如果安插處之父節點亦為紅節點,就違反紅黑樹規則,此時必須
    // 做樹形旋轉(及顏色改變,在程式它處)。
    inline void 
    __rb_tree_rotate_right(__rb_tree_node_base* x, __rb_tree_node_base*& root)
    {
      // x 為旋轉點
      __rb_tree_node_base* y = x->left;    // y 為旋轉點的左子節點
      x->left = y->right;
      if (y->right != 0)
        y->right->parent = x;     // 別忘了回馬槍設定父節點
      y->parent = x->parent;
    
      // 令 y 完全頂替 x 的地位(必須將 x 對其父節點的關係完全接收過來)
      if (x == root)                    // x 為根節點
        root = y;
      else if (x == x->parent->right)    // x 為其父節點的右子節點
        x->parent->right = y;
      else                            // x 為其父節點的左子節點
        x->parent->left = y;
      y->right = x;
      x->parent = y;
    }
    
    // 重新令樹形平衡(改變顏色及旋轉樹形)
    // 參數一為新增節點,參數二為 root
    inline void 
    __rb_tree_rebalance(__rb_tree_node_base* x, __rb_tree_node_base*& root)
    {
      x->color = __rb_tree_red;        // 新節點必為紅
      while (x != root && x->parent->color == __rb_tree_red) { // 父節點為紅
        if (x->parent == x->parent->parent->left) { // 父節點為祖父節點之左子節點
          __rb_tree_node_base* y = x->parent->parent->right;    // 令y 為伯父節點
          if (y && y->color == __rb_tree_red) {         // 伯父節點存在,且為紅
            x->parent->color = __rb_tree_black;          // 更改父節點為黑
            y->color = __rb_tree_black;                // 更改伯父節點為黑
            x->parent->parent->color = __rb_tree_red;     // 更改祖父節點為紅
            x = x->parent->parent;
          }
          else {    // 無伯父節點,或伯父節點為黑
            if (x == x->parent->right) { // 如果新節點為父節點之右子節點
              x = x->parent;
              __rb_tree_rotate_left(x, root); // 第一參數為左旋點
            }
            x->parent->color = __rb_tree_black;    // 改變顏色
            x->parent->parent->color = __rb_tree_red;
            __rb_tree_rotate_right(x->parent->parent, root); // 第一參數為右旋點
          }
        }
        else {    // 父節點為祖父節點之右子節點
          __rb_tree_node_base* y = x->parent->parent->left; // 令y 為伯父節點
          if (y && y->color == __rb_tree_red) {        // 有伯父節點,且為紅
            x->parent->color = __rb_tree_black;        // 更改父節點為黑
            y->color = __rb_tree_black;                 // 更改伯父節點為黑
            x->parent->parent->color = __rb_tree_red;     // 更改祖父節點為紅
            x = x->parent->parent;    // 準備繼續往上層檢查...
          }
          else {    // 無伯父節點,或伯父節點為黑
            if (x == x->parent->left) {    // 如果新節點為父節點之左子節點
              x = x->parent;
              __rb_tree_rotate_right(x, root);     // 第一參數為右旋點
            }
            x->parent->color = __rb_tree_black;    // 改變顏色
            x->parent->parent->color = __rb_tree_red;
            __rb_tree_rotate_left(x->parent->parent, root); // 第一參數為左旋點
          }
        }
      }    // while 結束
      root->color = __rb_tree_black;    // 根節點永遠為黑
    }
    
    inline __rb_tree_node_base*
    __rb_tree_rebalance_for_erase(__rb_tree_node_base* z,
                                  __rb_tree_node_base*& root,
                                  __rb_tree_node_base*& leftmost,
                                  __rb_tree_node_base*& rightmost)
    {
      __rb_tree_node_base* y = z;
      __rb_tree_node_base* x = 0;
      __rb_tree_node_base* x_parent = 0;
      if (y->left == 0)             // z has at most one non-null child. y == z.
        x = y->right;               // x might be null.
      else
        if (y->right == 0)          // z has exactly one non-null child.  y == z.
          x = y->left;              // x is not null.
        else {                      // z has two non-null children.  Set y to
          y = y->right;             //   z's successor.  x might be null.
          while (y->left != 0)
            y = y->left;
          x = y->right;
        }
      if (y != z) {                 // relink y in place of z.  y is z's successor
        z->left->parent = y; 
        y->left = z->left;
        if (y != z->right) {
          x_parent = y->parent;
          if (x) x->parent = y->parent;
          y->parent->left = x;      // y must be a left child
          y->right = z->right;
          z->right->parent = y;
        }
        else
          x_parent = y;  
        if (root == z)
          root = y;
        else if (z->parent->left == z)
          z->parent->left = y;
        else 
          z->parent->right = y;
        y->parent = z->parent;
        __STD::swap(y->color, z->color);
        y = z;
        // y now points to node to be actually deleted
      }
      else {                        // y == z
        x_parent = y->parent;
        if (x) x->parent = y->parent;   
        if (root == z)
          root = x;
        else 
          if (z->parent->left == z)
            z->parent->left = x;
          else
            z->parent->right = x;
        if (leftmost == z) 
          if (z->right == 0)        // z->left must be null also
            leftmost = z->parent;
        // makes leftmost == header if z == root
          else
            leftmost = __rb_tree_node_base::minimum(x);
        if (rightmost == z)  
          if (z->left == 0)         // z->right must be null also
            rightmost = z->parent;  
        // makes rightmost == header if z == root
          else                      // x == z->left
            rightmost = __rb_tree_node_base::maximum(x);
      }
      if (y->color != __rb_tree_red) { 
        while (x != root && (x == 0 || x->color == __rb_tree_black))
          if (x == x_parent->left) {
            __rb_tree_node_base* w = x_parent->right;
            if (w->color == __rb_tree_red) {
              w->color = __rb_tree_black;
              x_parent->color = __rb_tree_red;
              __rb_tree_rotate_left(x_parent, root);
              w = x_parent->right;
            }
            if ((w->left == 0 || w->left->color == __rb_tree_black) &&
                (w->right == 0 || w->right->color == __rb_tree_black)) {
              w->color = __rb_tree_red;
              x = x_parent;
              x_parent = x_parent->parent;
            } else {
              if (w->right == 0 || w->right->color == __rb_tree_black) {
                if (w->left) w->left->color = __rb_tree_black;
                w->color = __rb_tree_red;
                __rb_tree_rotate_right(w, root);
                w = x_parent->right;
              }
              w->color = x_parent->color;
              x_parent->color = __rb_tree_black;
              if (w->right) w->right->color = __rb_tree_black;
              __rb_tree_rotate_left(x_parent, root);
              break;
            }
          } else {                  // same as above, with right <-> left.
            __rb_tree_node_base* w = x_parent->left;
            if (w->color == __rb_tree_red) {
              w->color = __rb_tree_black;
              x_parent->color = __rb_tree_red;
              __rb_tree_rotate_right(x_parent, root);
              w = x_parent->left;
            }
            if ((w->right == 0 || w->right->color == __rb_tree_black) &&
                (w->left == 0 || w->left->color == __rb_tree_black)) {
              w->color = __rb_tree_red;
              x = x_parent;
              x_parent = x_parent->parent;
            } else {
              if (w->left == 0 || w->left->color == __rb_tree_black) {
                if (w->right) w->right->color = __rb_tree_black;
                w->color = __rb_tree_red;
                __rb_tree_rotate_left(w, root);
                w = x_parent->left;
              }
              w->color = x_parent->color;
              x_parent->color = __rb_tree_black;
              if (w->left) w->left->color = __rb_tree_black;
              __rb_tree_rotate_right(x_parent, root);
              break;
            }
          }
        if (x) x->color = __rb_tree_black;
      }
      return y;
    }
    
    template <class Key, class Value, class KeyOfValue, class Compare,
              class Alloc = alloc>
    class rb_tree {
    protected:
      typedef void* void_pointer;
      typedef __rb_tree_node_base* base_ptr;
      typedef __rb_tree_node<Value> rb_tree_node;
      typedef simple_alloc<rb_tree_node, Alloc> rb_tree_node_allocator;
      typedef __rb_tree_color_type color_type;
    public:
      // 注意,沒有定義 iterator(喔,不,定義在後面)
      typedef Key key_type;
      typedef Value value_type;
      typedef value_type* pointer;
      typedef const value_type* const_pointer;
      typedef value_type& reference;
      typedef const value_type& const_reference;
      typedef rb_tree_node* link_type;
      typedef size_t size_type;
      typedef ptrdiff_t difference_type;
    protected:
      link_type get_node() { return rb_tree_node_allocator::allocate(); }
      void put_node(link_type p) { rb_tree_node_allocator::deallocate(p); }
    
      link_type create_node(const value_type& x) {
        link_type tmp = get_node();            // 配置空間
        __STL_TRY {
          construct(&tmp->value_field, x);    // 建構內容
        }
        __STL_UNWIND(put_node(tmp));
        return tmp;
      }
    
      link_type clone_node(link_type x) {    // 複製一個節點(的值和色)
        link_type tmp = create_node(x->value_field);
        tmp->color = x->color;
        tmp->left = 0;
        tmp->right = 0;
        return tmp;
      }
    
      void destroy_node(link_type p) {
        destroy(&p->value_field);        // 解構內容
        put_node(p);                    // 釋還記憶體
      }
    
    protected:
      // RB-tree 只以三筆資料表現。
      size_type node_count; // 追蹤記錄樹的大小(節點數量)
      link_type header;  
      Compare key_compare;     // 節點間的鍵值大小比較準則。應該會是個 function object。
    
      // 以下三個函式用來方便取得 header 的成員
      link_type& root() const { return (link_type&) header->parent; }
      link_type& leftmost() const { return (link_type&) header->left; }
      link_type& rightmost() const { return (link_type&) header->right; }
    
      // 以下六個函式用來方便取得節點 x 的成員
      static link_type& left(link_type x) { return (link_type&)(x->left); }
      static link_type& right(link_type x) { return (link_type&)(x->right); }
      static link_type& parent(link_type x) { return (link_type&)(x->parent); }
      static reference value(link_type x) { return x->value_field; }
      static const Key& key(link_type x) { return KeyOfValue()(value(x)); }
      static color_type& color(link_type x) { return (color_type&)(x->color); }
    
      // 以下六個函式用來方便取得節點 x 的成員
      static link_type& left(base_ptr x) { return (link_type&)(x->left); }
      static link_type& right(base_ptr x) { return (link_type&)(x->right); }
      static link_type& parent(base_ptr x) { return (link_type&)(x->parent); }
      static reference value(base_ptr x) { return ((link_type)x)->value_field; }
      static const Key& key(base_ptr x) { return KeyOfValue()(value(link_type(x)));} 
      static color_type& color(base_ptr x) { return (color_type&)(link_type(x)->color); }
    
      // 求取極大值和極小值。node class 有實作此功能,交給它們完成即可。
      static link_type minimum(link_type x) { 
        return (link_type)  __rb_tree_node_base::minimum(x);
      }
      static link_type maximum(link_type x) {
        return (link_type) __rb_tree_node_base::maximum(x);
      }
    
    public:
      typedef __rb_tree_iterator<value_type, reference, pointer> iterator;
      typedef __rb_tree_iterator<value_type, const_reference, const_pointer> 
              const_iterator;
    
    #ifdef __STL_CLASS_PARTIAL_SPECIALIZATION
      typedef reverse_iterator<const_iterator> const_reverse_iterator;
      typedef reverse_iterator<iterator> reverse_iterator;
    #else /* __STL_CLASS_PARTIAL_SPECIALIZATION */
      typedef reverse_bidirectional_iterator<iterator, value_type, reference,
                                             difference_type>
              reverse_iterator; 
      typedef reverse_bidirectional_iterator<const_iterator, value_type,
                                             const_reference, difference_type>
              const_reverse_iterator;
    #endif /* __STL_CLASS_PARTIAL_SPECIALIZATION */ 
    private:
      iterator __insert(base_ptr x, base_ptr y, const value_type& v);
      link_type __copy(link_type x, link_type p);
      void __erase(link_type x);
      void init() {
        header = get_node();    // 產生一個節點空間,令 header 指向它
        color(header) = __rb_tree_red; // 令 header 為紅色,用來區分 header  
                                       // 和 root(在 iterator.operator++ 中)
        root() = 0;
        leftmost() = header;    // 令 header 的左子節點為自己。
        rightmost() = header;    // 令 header 的右子節點為自己。
      }
    public:
                                    // allocation/deallocation
      rb_tree(const Compare& comp = Compare())
        : node_count(0), key_compare(comp) { init(); }
    
      // 以另一個 rb_tree 物件 x 為初值
      rb_tree(const rb_tree<Key, Value, KeyOfValue, Compare, Alloc>& x) 
        : node_count(0), key_compare(x.key_compare)
      { 
        header = get_node();    // 產生一個節點空間,令 header 指向它
        color(header) = __rb_tree_red;    // 令 header 為紅色
        if (x.root() == 0) {    //  如果 x 是個空白樹
          root() = 0;
          leftmost() = header;     // 令 header 的左子節點為自己。
          rightmost() = header; // 令 header 的右子節點為自己。
        }
        else {    //  x 不是一個空白樹
          __STL_TRY {
            root() = __copy(x.root(), header);        // ??? 
          }
          __STL_UNWIND(put_node(header));
          leftmost() = minimum(root());    // 令 header 的左子節點為最小節點
          rightmost() = maximum(root());    // 令 header 的右子節點為最大節點
        }
        node_count = x.node_count;
      }
      ~rb_tree() {
        clear();
        put_node(header);
      }
      rb_tree<Key, Value, KeyOfValue, Compare, Alloc>& 
      operator=(const rb_tree<Key, Value, KeyOfValue, Compare, Alloc>& x);
    
    public:    
                                    // accessors:
      Compare key_comp() const { return key_compare; }
      iterator begin() { return leftmost(); }        // RB 樹的起頭為最左(最小)節點處
      const_iterator begin() const { return leftmost(); }
      iterator end() { return header; }    // RB 樹的終點為 header所指處
      const_iterator end() const { return header; }
      reverse_iterator rbegin() { return reverse_iterator(end()); }
      const_reverse_iterator rbegin() const { 
        return const_reverse_iterator(end()); 
      }
      reverse_iterator rend() { return reverse_iterator(begin()); }
      const_reverse_iterator rend() const { 
        return const_reverse_iterator(begin());
      } 
      bool empty() const { return node_count == 0; }
      size_type size() const { return node_count; }
      size_type max_size() const { return size_type(-1); }
    
      void swap(rb_tree<Key, Value, KeyOfValue, Compare, Alloc>& t) {
        // RB-tree 只以三個資料成員表現。所以互換兩個 RB-trees時,
        // 只需將這三個成員互換即可。
        __STD::swap(header, t.header);
        __STD::swap(node_count, t.node_count);
        __STD::swap(key_compare, t.key_compare);
      }
        
    public:
                                    // insert/erase
      // 將 x 安插到 RB-tree 中(保持節點值獨一無二)。
      pair<iterator,bool> insert_unique(const value_type& x);
      // 將 x 安插到 RB-tree 中(允許節點值重複)。
      iterator insert_equal(const value_type& x);
    
      iterator insert_unique(iterator position, const value_type& x);
      iterator insert_equal(iterator position, const value_type& x);
    
    #ifdef __STL_MEMBER_TEMPLATES  
      template <class InputIterator>
      void insert_unique(InputIterator first, InputIterator last);
      template <class InputIterator>
      void insert_equal(InputIterator first, InputIterator last);
    #else /* __STL_MEMBER_TEMPLATES */
      void insert_unique(const_iterator first, const_iterator last);
      void insert_unique(const value_type* first, const value_type* last);
      void insert_equal(const_iterator first, const_iterator last);
      void insert_equal(const value_type* first, const value_type* last);
    #endif /* __STL_MEMBER_TEMPLATES */
    
      void erase(iterator position);
      size_type erase(const key_type& x);
      void erase(iterator first, iterator last);
      void erase(const key_type* first, const key_type* last);
      void clear() {
        if (node_count != 0) {
          __erase(root());
          leftmost() = header;
          root() = 0;
          rightmost() = header;
          node_count = 0;
        }
      }      
    
    public:
                                    // 集合(set)的各種操作行為:
      iterator find(const key_type& x);
      const_iterator find(const key_type& x) const;
      size_type count(const key_type& x) const;
      iterator lower_bound(const key_type& x);
      const_iterator lower_bound(const key_type& x) const;
      iterator upper_bound(const key_type& x);
      const_iterator upper_bound(const key_type& x) const;
      pair<iterator,iterator> equal_range(const key_type& x);
      pair<const_iterator, const_iterator> equal_range(const key_type& x) const;
    
    public:
                                    // Debugging.
      bool __rb_verify() const;
    };
    
    template <class Key, class Value, class KeyOfValue, class Compare, class Alloc>
    inline bool operator==(const rb_tree<Key, Value, KeyOfValue, Compare, Alloc>& x, 
                           const rb_tree<Key, Value, KeyOfValue, Compare, Alloc>& y) {
      return x.size() == y.size() && equal(x.begin(), x.end(), y.begin());
    }
    
    template <class Key, class Value, class KeyOfValue, class Compare, class Alloc>
    inline bool operator<(const rb_tree<Key, Value, KeyOfValue, Compare, Alloc>& x, 
                          const rb_tree<Key, Value, KeyOfValue, Compare, Alloc>& y) {
      return lexicographical_compare(x.begin(), x.end(), y.begin(), y.end());
    }
    
    #ifdef __STL_FUNCTION_TMPL_PARTIAL_ORDER
    
    template <class Key, class Value, class KeyOfValue, class Compare, class Alloc>
    inline void swap(rb_tree<Key, Value, KeyOfValue, Compare, Alloc>& x, 
                     rb_tree<Key, Value, KeyOfValue, Compare, Alloc>& y) {
      x.swap(y);
    }
    
    #endif /* __STL_FUNCTION_TMPL_PARTIAL_ORDER */
    
    
    template <class Key, class Value, class KeyOfValue, class Compare, class Alloc>
    rb_tree<Key, Value, KeyOfValue, Compare, Alloc>& 
    rb_tree<Key, Value, KeyOfValue, Compare, Alloc>::
    operator=(const rb_tree<Key, Value, KeyOfValue, Compare, Alloc>& x) {
      if (this != &x) {
                                    // Note that Key may be a constant type.
        clear();
        node_count = 0;
        key_compare = x.key_compare;        
        if (x.root() == 0) {
          root() = 0;
          leftmost() = header;
          rightmost() = header;
        }
        else {
          root() = __copy(x.root(), header);
          leftmost() = minimum(root());
          rightmost() = maximum(root());
          node_count = x.node_count;
        }
      }
      return *this;
    }
    
    template <class Key, class Value, class KeyOfValue, class Compare, class Alloc>
    typename rb_tree<Key, Value, KeyOfValue, Compare, Alloc>::iterator
    rb_tree<Key, Value, KeyOfValue, Compare, Alloc>::
    __insert(base_ptr x_, base_ptr y_, const Value& v) {
    // 參數x_ 為新值安插點,參數y_ 為安插點之父節點,參數v 為新值。
      link_type x = (link_type) x_;
      link_type y = (link_type) y_;
      link_type z;
    
      // key_compare 是鍵值大小比較準則。應該會是個 function object。
      if (y == header || x != 0 || key_compare(KeyOfValue()(v), key(y))) {
        z = create_node(v);  // 產生一個新節點
        left(y) = z;          // 這使得當 y 即為 header時,leftmost() = z
        if (y == header) {
          root() = z;
          rightmost() = z;
        }
        else if (y == leftmost())    // 如果y為最左節點
          leftmost() = z;               // 維護leftmost(),使它永遠指向最左節點
      }
      else {
        z = create_node(v);        // 產生一個新節點
        right(y) = z;                // 令新節點成為安插點之父節點 y 的右子節點
        if (y == rightmost())
          rightmost() = z;              // 維護rightmost(),使它永遠指向最右節點
      }
      parent(z) = y;        // 設定新節點的父節點
      left(z) = 0;        // 設定新節點的左子節點
      right(z) = 0;         // 設定新節點的右子節點
                              // 新節點的顏色將在 __rb_tree_rebalance() 設定(並調整)
      __rb_tree_rebalance(z, header->parent);    // 參數一為新增節點,參數二為 root
      ++node_count;        // 節點數累加
      return iterator(z);    // 傳回一個迭代器,指向新增節點
    }
    
    // 安插新值;節點鍵值允許重複。
    // 注意,傳回值是一個 RB-tree 迭代器,指向新增節點
    template <class Key, class Value, class KeyOfValue, class Compare, class Alloc>
    typename rb_tree<Key, Value, KeyOfValue, Compare, Alloc>::iterator
    rb_tree<Key, Value, KeyOfValue, Compare, Alloc>::insert_equal(const Value& v)
    {
      link_type y = header;
      link_type x = root();    // 從根節點開始
      while (x != 0) {        // 從根節點開始,往下尋找適當的安插點
        y = x;
        x = key_compare(KeyOfValue()(v), key(x)) ? left(x) : right(x);
        // 以上,遇「大」則往左,遇「小於或等於」則往右
      }
      return __insert(x, y, v);
    }
    
    // 安插新值;節點鍵值不允許重複,若重複則安插無效。
    // 注意,傳回值是個pair,第一元素是個 RB-tree 迭代器,指向新增節點,
    // 第二元素表示安插成功與否。
    template <class Key, class Value, class KeyOfValue, class Compare, class Alloc>
    pair<typename rb_tree<Key, Value, KeyOfValue, Compare, Alloc>::iterator, bool>
    rb_tree<Key, Value, KeyOfValue, Compare, Alloc>::insert_unique(const Value& v)
    {
      link_type y = header;
      link_type x = root();    // 從根節點開始
      bool comp = true;
      while (x != 0) {         // 從根節點開始,往下尋找適當的安插點
        y = x;
        comp = key_compare(KeyOfValue()(v), key(x)); // v 鍵值小於目前節點之鍵值?
        x = comp ? left(x) : right(x);    // 遇「大」則往左,遇「小於或等於」則往右
      }
      // 離開 while 迴圈之後,y 所指即安插點之父節點(此時的它必為葉節點)
    
      iterator j = iterator(y);   // 令迭代器j指向安插點之父節點 y
      if (comp)    // 如果離開 while 迴圈時 comp 為真(表示遇「大」,將安插於左側)
        if (j == begin())   // 如果安插點之父節點為最左節點
          return pair<iterator,bool>(__insert(x, y, v), true);
          // 以上,x 為安插點,y 為安插點之父節點,v 為新值。
        else    // 否則(安插點之父節點不為最左節點)
          --j;    // 調整 j,回頭準備測試...
      if (key_compare(key(j.node), KeyOfValue()(v)))    
        // 小於新值(表示遇「小」,將安插於右側)
        return pair<iterator,bool>(__insert(x, y, v), true);
    
      // 進行至此,表示新值一定與樹中鍵值重複,那麼就不該插入新值。
      return pair<iterator,bool>(j, false);
    }
    
    
    template <class Key, class Val, class KeyOfValue, class Compare, class Alloc>
    typename rb_tree<Key, Val, KeyOfValue, Compare, Alloc>::iterator 
    rb_tree<Key, Val, KeyOfValue, Compare, Alloc>::insert_unique(iterator position,
                                                                 const Val& v) {
      if (position.node == header->left) // begin()
        if (size() > 0 && key_compare(KeyOfValue()(v), key(position.node)))
          return __insert(position.node, position.node, v);
      // first argument just needs to be non-null 
        else
          return insert_unique(v).first;
      else if (position.node == header) // end()
        if (key_compare(key(rightmost()), KeyOfValue()(v)))
          return __insert(0, rightmost(), v);
        else
          return insert_unique(v).first;
      else {
        iterator before = position;
        --before;
        if (key_compare(key(before.node), KeyOfValue()(v))
            && key_compare(KeyOfValue()(v), key(position.node)))
          if (right(before.node) == 0)
            return __insert(0, before.node, v); 
          else
            return __insert(position.node, position.node, v);
        // first argument just needs to be non-null 
        else
          return insert_unique(v).first;
      }
    }
    
    template <class Key, class Val, class KeyOfValue, class Compare, class Alloc>
    typename rb_tree<Key, Val, KeyOfValue, Compare, Alloc>::iterator 
    rb_tree<Key, Val, KeyOfValue, Compare, Alloc>::insert_equal(iterator position,
                                                                const Val& v) {
      if (position.node == header->left) // begin()
        if (size() > 0 && key_compare(KeyOfValue()(v), key(position.node)))
          return __insert(position.node, position.node, v);
      // first argument just needs to be non-null 
        else
          return insert_equal(v);
      else if (position.node == header) // end()
        if (!key_compare(KeyOfValue()(v), key(rightmost())))
          return __insert(0, rightmost(), v);
        else
          return insert_equal(v);
      else {
        iterator before = position;
        --before;
        if (!key_compare(KeyOfValue()(v), key(before.node))
            && !key_compare(key(position.node), KeyOfValue()(v)))
          if (right(before.node) == 0)
            return __insert(0, before.node, v); 
          else
            return __insert(position.node, position.node, v);
        // first argument just needs to be non-null 
        else
          return insert_equal(v);
      }
    }
    
    #ifdef __STL_MEMBER_TEMPLATES  
    
    template <class K, class V, class KoV, class Cmp, class Al> template<class II>
    void rb_tree<K, V, KoV, Cmp, Al>::insert_equal(II first, II last) {
      for ( ; first != last; ++first)
        insert_equal(*first);
    }
    
    template <class K, class V, class KoV, class Cmp, class Al> template<class II>
    void rb_tree<K, V, KoV, Cmp, Al>::insert_unique(II first, II last) {
      for ( ; first != last; ++first)
        insert_unique(*first);
    }
    
    #else /* __STL_MEMBER_TEMPLATES */
    
    template <class K, class V, class KoV, class Cmp, class Al>
    void
    rb_tree<K, V, KoV, Cmp, Al>::insert_equal(const V* first, const V* last) {
      for ( ; first != last; ++first)
        insert_equal(*first);
    }
    
    template <class K, class V, class KoV, class Cmp, class Al>
    void
    rb_tree<K, V, KoV, Cmp, Al>::insert_equal(const_iterator first,
                                              const_iterator last) {
      for ( ; first != last; ++first)
        insert_equal(*first);
    }
    
    template <class K, class V, class KoV, class Cmp, class A>
    void 
    rb_tree<K, V, KoV, Cmp, A>::insert_unique(const V* first, const V* last) {
      for ( ; first != last; ++first)
        insert_unique(*first);
    }
    
    template <class K, class V, class KoV, class Cmp, class A>
    void 
    rb_tree<K, V, KoV, Cmp, A>::insert_unique(const_iterator first,
                                              const_iterator last) {
      for ( ; first != last; ++first)
        insert_unique(*first);
    }
    
    #endif /* __STL_MEMBER_TEMPLATES */
             
    template <class Key, class Value, class KeyOfValue, class Compare, class Alloc>
    inline void
    rb_tree<Key, Value, KeyOfValue, Compare, Alloc>::erase(iterator position) {
      link_type y = (link_type) __rb_tree_rebalance_for_erase(position.node,
                                                              header->parent,
                                                              header->left,
                                                              header->right);
      destroy_node(y);
      --node_count;
    }
    
    template <class Key, class Value, class KeyOfValue, class Compare, class Alloc>
    typename rb_tree<Key, Value, KeyOfValue, Compare, Alloc>::size_type 
    rb_tree<Key, Value, KeyOfValue, Compare, Alloc>::erase(const Key& x) {
      pair<iterator,iterator> p = equal_range(x);
      size_type n = 0;
      distance(p.first, p.second, n);
      erase(p.first, p.second);
      return n;
    }
    
    template <class K, class V, class KeyOfValue, class Compare, class Alloc>
    typename rb_tree<K, V, KeyOfValue, Compare, Alloc>::link_type 
    rb_tree<K, V, KeyOfValue, Compare, Alloc>::__copy(link_type x, link_type p) {
                                    // structural copy.  x and p must be non-null.
      link_type top = clone_node(x);
      top->parent = p;
     
      __STL_TRY {
        if (x->right)
          top->right = __copy(right(x), top);
        p = top;
        x = left(x);
    
        while (x != 0) {
          link_type y = clone_node(x);
          p->left = y;
          y->parent = p;
          if (x->right)
            y->right = __copy(right(x), y);
          p = y;
          x = left(x);
        }
      }
      __STL_UNWIND(__erase(top));
    
      return top;
    }
    
    template <class Key, class Value, class KeyOfValue, class Compare, class Alloc>
    void rb_tree<Key, Value, KeyOfValue, Compare, Alloc>::__erase(link_type x) {
                                    // erase without rebalancing
      while (x != 0) {
        __erase(right(x));
        link_type y = left(x);
        destroy_node(x);
        x = y;
      }
    }
    
    template <class Key, class Value, class KeyOfValue, class Compare, class Alloc>
    void rb_tree<Key, Value, KeyOfValue, Compare, Alloc>::erase(iterator first, 
                                                                iterator last) {
      if (first == begin() && last == end())
        clear();
      else
        while (first != last) erase(first++);
    }
    
    template <class Key, class Value, class KeyOfValue, class Compare, class Alloc>
    void rb_tree<Key, Value, KeyOfValue, Compare, Alloc>::erase(const Key* first, 
                                                                const Key* last) {
      while (first != last) erase(*first++);
    }
    
    // 尋找 RB 樹中是否有鍵值為 k 的節點
    template <class Key, class Value, class KeyOfValue, class Compare, class Alloc>
    typename rb_tree<Key, Value, KeyOfValue, Compare, Alloc>::iterator 
    rb_tree<Key, Value, KeyOfValue, Compare, Alloc>::find(const Key& k) {
      link_type y = header;        // Last node which is not less than k. 
      link_type x = root();        // Current node. 
    
      while (x != 0) 
        // 以下,key_compare 是節點鍵值大小比較準則。應該會是個 function object。
        if (!key_compare(key(x), k)) 
          // 進行到這裡,表示 x 鍵值大於 k。遇到大值就向左走。
          y = x, x = left(x);    // 注意語法!
        else
          // 進行到這裡,表示 x 鍵值小於 k。遇到小值就向右走。
          x = right(x);
    
      iterator j = iterator(y);   
      return (j == end() || key_compare(k, key(j.node))) ? end() : j;
    }
    
    // 尋找 RB 樹中是否有鍵值為 k 的節點
    template <class Key, class Value, class KeyOfValue, class Compare, class Alloc>
    typename rb_tree<Key, Value, KeyOfValue, Compare, Alloc>::const_iterator 
    rb_tree<Key, Value, KeyOfValue, Compare, Alloc>::find(const Key& k) const {
      link_type y = header; /* Last node which is not less than k. */
      link_type x = root(); /* Current node. */
    
      while (x != 0) {
        // 以下,key_compare 是節點鍵值大小比較準則。應該會是個 function object。
        if (!key_compare(key(x), k))
          // 進行到這裡,表示 x 鍵值大於 k。遇到大值就向左走。
          y = x, x = left(x);    // 注意語法!
        else
          // 進行到這裡,表示 x 鍵值小於 k。遇到小值就向右走。
          x = right(x);
      }
      const_iterator j = const_iterator(y);   
      return (j == end() || key_compare(k, key(j.node))) ? end() : j;
    }
    
    // 計算 RB 樹中鍵值為 k 的節點個數
    template <class Key, class Value, class KeyOfValue, class Compare, class Alloc>
    typename rb_tree<Key, Value, KeyOfValue, Compare, Alloc>::size_type 
    rb_tree<Key, Value, KeyOfValue, Compare, Alloc>::count(const Key& k) const {
      pair<const_iterator, const_iterator> p = equal_range(k);
      size_type n = 0;
      distance(p.first, p.second, n);
      return n;
    }
    
    template <class Key, class Value, class KeyOfValue, class Compare, class Alloc>
    typename rb_tree<Key, Value, KeyOfValue, Compare, Alloc>::iterator 
    rb_tree<Key, Value, KeyOfValue, Compare, Alloc>::lower_bound(const Key& k) {
      link_type y = header; /* Last node which is not less than k. */
      link_type x = root(); /* Current node. */
    
      while (x != 0) 
        if (!key_compare(key(x), k))
          y = x, x = left(x);
        else
          x = right(x);
    
      return iterator(y);
    }
    
    template <class Key, class Value, class KeyOfValue, class Compare, class Alloc>
    typename rb_tree<Key, Value, KeyOfValue, Compare, Alloc>::const_iterator 
    rb_tree<Key, Value, KeyOfValue, Compare, Alloc>::lower_bound(const Key& k) const {
      link_type y = header; /* Last node which is not less than k. */
      link_type x = root(); /* Current node. */
    
      while (x != 0) 
        if (!key_compare(key(x), k))
          y = x, x = left(x);
        else
          x = right(x);
    
      return const_iterator(y);
    }
    
    template <class Key, class Value, class KeyOfValue, class Compare, class Alloc>
    typename rb_tree<Key, Value, KeyOfValue, Compare, Alloc>::iterator 
    rb_tree<Key, Value, KeyOfValue, Compare, Alloc>::upper_bound(const Key& k) {
      link_type y = header; /* Last node which is greater than k. */
      link_type x = root(); /* Current node. */
    
       while (x != 0) 
         if (key_compare(k, key(x)))
           y = x, x = left(x);
         else
           x = right(x);
    
       return iterator(y);
    }
    
    template <class Key, class Value, class KeyOfValue, class Compare, class Alloc>
    typename rb_tree<Key, Value, KeyOfValue, Compare, Alloc>::const_iterator 
    rb_tree<Key, Value, KeyOfValue, Compare, Alloc>::upper_bound(const Key& k) const {
      link_type y = header; /* Last node which is greater than k. */
      link_type x = root(); /* Current node. */
    
       while (x != 0) 
         if (key_compare(k, key(x)))
           y = x, x = left(x);
         else
           x = right(x);
    
       return const_iterator(y);
    }
    
    template <class Key, class Value, class KeyOfValue, class Compare, class Alloc>
    inline pair<typename rb_tree<Key, Value, KeyOfValue, Compare, Alloc>::iterator,
                typename rb_tree<Key, Value, KeyOfValue, Compare, Alloc>::iterator>
    rb_tree<Key, Value, KeyOfValue, Compare, Alloc>::equal_range(const Key& k) {
      return pair<iterator, iterator>(lower_bound(k), upper_bound(k));
    }
    
    template <class Key, class Value, class KoV, class Compare, class Alloc>
    inline pair<typename rb_tree<Key, Value, KoV, Compare, Alloc>::const_iterator,
                typename rb_tree<Key, Value, KoV, Compare, Alloc>::const_iterator>
    rb_tree<Key, Value, KoV, Compare, Alloc>::equal_range(const Key& k) const {
      return pair<const_iterator,const_iterator>(lower_bound(k), upper_bound(k));
    }
    
    // 計算從 node 至 root 路徑中的黑節點數量。
    inline int __black_count(__rb_tree_node_base* node, __rb_tree_node_base* root)
    {
      if (node == 0)
        return 0;
      else {
        int bc = node->color == __rb_tree_black ? 1 : 0;
        if (node == root)
          return bc;
        else
          return bc + __black_count(node->parent, root); // 累加
      }
    }
    
    // 驗證己身這棵樹是否符合 RB 樹的條件
    template <class Key, class Value, class KeyOfValue, class Compare, class Alloc>
    bool 
    rb_tree<Key, Value, KeyOfValue, Compare, Alloc>::__rb_verify() const
    {
      // 空樹,符合RB樹標準
      if (node_count == 0 || begin() == end())
        return node_count == 0 && begin() == end() &&
          header->left == header && header->right == header;
    
      // 最左(葉)節點至 root 路徑內的黑節點數
      int len = __black_count(leftmost(), root()); 
      // 以下走訪整個RB樹,針對每個節點(從最小到最大)...
      for (const_iterator it = begin(); it != end(); ++it) { 
        link_type x = (link_type) it.node; // __rb_tree_base_iterator::node
        link_type L = left(x);        // 這是左子節點
        link_type R = right(x);     // 這是右子節點
    
        if (x->color == __rb_tree_red)
          if ((L && L->color == __rb_tree_red) ||
              (R && R->color == __rb_tree_red))
            return false;    // 父子節點同為紅色,不符合 RB 樹的要求。
    
        if (L && key_compare(key(x), key(L))) // 目前節點的鍵值小於左子節點鍵值
          return false;             // 不符合二元搜尋樹的要求。
        if (R && key_compare(key(R), key(x))) // 目前節點的鍵值大於右子節點鍵值
          return false;        // 不符合二元搜尋樹的要求。
    
        // 「葉節點至 root」路徑內的黑節點數,與「最左節點至 root」路徑內的黑節點數不同。
        // 這不符合 RB 樹的要求。
        if (!L && !R && __black_count(x, root()) != len) 
          return false;
      }
    
      if (leftmost() != __rb_tree_node_base::minimum(root()))
        return false;    // 最左節點不為最小節點,不符合二元搜尋樹的要求。
      if (rightmost() != __rb_tree_node_base::maximum(root()))
        return false;    // 最右節點不為最大節點,不符合二元搜尋樹的要求。
    
      return true;
    }
    
    __STL_END_NAMESPACE 
    
    #endif /* __SGI_STL_INTERNAL_TREE_H */
    
    // Local Variables:
    // mode:C++
    // End:
  • 相关阅读:
    野指针
    MFC动态创建控件及其消息响应函数
    关于CString总结
    关于char, wchar_t, TCHAR, _T(),L,宏 _T、TEXT,_TEXT、L
    VS2005、VS2008中的快捷键、组合键大全
    could not be found (are you missing a using directive or an assembly reference?)
    项目任务管理系统
    VSDBCMD
    sqlcmd
    Microsoft® Command Line Utilities 11 for SQL Server®
  • 原文地址:https://www.cnblogs.com/zendu/p/4987813.html
Copyright © 2011-2022 走看看