OpenCores
URL https://opencores.org/ocsvn/openrisc/openrisc/trunk

Subversion Repositories openrisc

[/] [openrisc/] [trunk/] [gnu-stable/] [gcc-4.5.1/] [libstdc++-v3/] [include/] [bits/] [hashtable.h] - Rev 826

Compare with Previous | Blame | View Log

// hashtable.h header -*- C++ -*-
 
// Copyright (C) 2007, 2008, 2009, 2010 Free Software Foundation, Inc.
//
// This file is part of the GNU ISO C++ Library.  This library is free
// software; you can redistribute it and/or modify it under the
// terms of the GNU General Public License as published by the
// Free Software Foundation; either version 3, 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 General Public License for more details.
 
// Under Section 7 of GPL version 3, you are granted additional
// permissions described in the GCC Runtime Library Exception, version
// 3.1, as published by the Free Software Foundation.
 
// You should have received a copy of the GNU General Public License and
// a copy of the GCC Runtime Library Exception along with this program;
// see the files COPYING3 and COPYING.RUNTIME respectively.  If not, see
// <http://www.gnu.org/licenses/>.
 
/** @file bits/hashtable.h
 *  This is an internal header file, included by other library headers.
 *  You should not attempt to use it directly.
 */
 
#ifndef _HASHTABLE_H
#define _HASHTABLE_H 1
 
#pragma GCC system_header
 
#include <bits/hashtable_policy.h>
 
namespace std
{
  // Class template _Hashtable, class definition.
 
  // Meaning of class template _Hashtable's template parameters
 
  // _Key and _Value: arbitrary CopyConstructible types.
 
  // _Allocator: an allocator type ([lib.allocator.requirements]) whose
  // value type is Value.  As a conforming extension, we allow for
  // value type != Value.
 
  // _ExtractKey: function object that takes a object of type Value
  // and returns a value of type _Key.
 
  // _Equal: function object that takes two objects of type k and returns
  // a bool-like value that is true if the two objects are considered equal.
 
  // _H1: the hash function.  A unary function object with argument type
  // Key and result type size_t.  Return values should be distributed
  // over the entire range [0, numeric_limits<size_t>:::max()].
 
  // _H2: the range-hashing function (in the terminology of Tavori and
  // Dreizin).  A binary function object whose argument types and result
  // type are all size_t.  Given arguments r and N, the return value is
  // in the range [0, N).
 
  // _Hash: the ranged hash function (Tavori and Dreizin). A binary function
  // whose argument types are _Key and size_t and whose result type is
  // size_t.  Given arguments k and N, the return value is in the range
  // [0, N).  Default: hash(k, N) = h2(h1(k), N).  If _Hash is anything other
  // than the default, _H1 and _H2 are ignored.
 
  // _RehashPolicy: Policy class with three members, all of which govern
  // the bucket count. _M_next_bkt(n) returns a bucket count no smaller
  // than n.  _M_bkt_for_elements(n) returns a bucket count appropriate
  // for an element count of n.  _M_need_rehash(n_bkt, n_elt, n_ins)
  // determines whether, if the current bucket count is n_bkt and the
  // current element count is n_elt, we need to increase the bucket
  // count.  If so, returns make_pair(true, n), where n is the new
  // bucket count.  If not, returns make_pair(false, <anything>).
 
  // ??? Right now it is hard-wired that the number of buckets never
  // shrinks.  Should we allow _RehashPolicy to change that?
 
  // __cache_hash_code: bool.  true if we store the value of the hash
  // function along with the value.  This is a time-space tradeoff.
  // Storing it may improve lookup speed by reducing the number of times
  // we need to call the Equal function.
 
  // __constant_iterators: bool.  true if iterator and const_iterator are
  // both constant iterator types.  This is true for unordered_set and
  // unordered_multiset, false for unordered_map and unordered_multimap.
 
  // __unique_keys: bool.  true if the return value of _Hashtable::count(k)
  // is always at most one, false if it may be an arbitrary number.  This
  // true for unordered_set and unordered_map, false for unordered_multiset
  // and unordered_multimap.
 
  template<typename _Key, typename _Value, typename _Allocator,
	   typename _ExtractKey, typename _Equal,
	   typename _H1, typename _H2, typename _Hash, 
	   typename _RehashPolicy,
	   bool __cache_hash_code,
	   bool __constant_iterators,
	   bool __unique_keys>
    class _Hashtable
    : public __detail::_Rehash_base<_RehashPolicy,
				    _Hashtable<_Key, _Value, _Allocator,
					       _ExtractKey,
					       _Equal, _H1, _H2, _Hash,
					       _RehashPolicy,
					       __cache_hash_code,
					       __constant_iterators,
					       __unique_keys> >,
      public __detail::_Hash_code_base<_Key, _Value, _ExtractKey, _Equal,
				       _H1, _H2, _Hash, __cache_hash_code>,
      public __detail::_Map_base<_Key, _Value, _ExtractKey, __unique_keys,
				 _Hashtable<_Key, _Value, _Allocator,
					    _ExtractKey,
					    _Equal, _H1, _H2, _Hash,
					    _RehashPolicy,
					    __cache_hash_code,
					    __constant_iterators,
					    __unique_keys> >,
      public __detail::_Equality_base<_ExtractKey, __unique_keys,
				      _Hashtable<_Key, _Value, _Allocator,
						 _ExtractKey,
						 _Equal, _H1, _H2, _Hash,
						 _RehashPolicy,
						 __cache_hash_code,
						 __constant_iterators,
						 __unique_keys> >
    {
    public:
      typedef _Allocator                                  allocator_type;
      typedef _Value                                      value_type;
      typedef _Key                                        key_type;
      typedef _Equal                                      key_equal;
      // mapped_type, if present, comes from _Map_base.
      // hasher, if present, comes from _Hash_code_base.
      typedef typename _Allocator::pointer                pointer;
      typedef typename _Allocator::const_pointer          const_pointer;
      typedef typename _Allocator::reference              reference;
      typedef typename _Allocator::const_reference        const_reference;
 
      typedef std::size_t                                 size_type;
      typedef std::ptrdiff_t                              difference_type;
      typedef __detail::_Node_iterator<value_type, __constant_iterators,
				       __cache_hash_code>
                                                          local_iterator;
      typedef __detail::_Node_const_iterator<value_type,
					     __constant_iterators,
					     __cache_hash_code>
                                                          const_local_iterator;
 
      typedef __detail::_Hashtable_iterator<value_type, __constant_iterators,
					    __cache_hash_code>
                                                          iterator;
      typedef __detail::_Hashtable_const_iterator<value_type,
						  __constant_iterators,
						  __cache_hash_code>
                                                          const_iterator;
 
      template<typename _Key2, typename _Value2, typename _Ex2, bool __unique2,
	       typename _Hashtable2>
        friend struct __detail::_Map_base;
 
    private:
      typedef __detail::_Hash_node<_Value, __cache_hash_code> _Node;
      typedef typename _Allocator::template rebind<_Node>::other
                                                        _Node_allocator_type;
      typedef typename _Allocator::template rebind<_Node*>::other
                                                        _Bucket_allocator_type;
 
      typedef typename _Allocator::template rebind<_Value>::other
                                                        _Value_allocator_type;
 
      _Node_allocator_type   _M_node_allocator;
      _Node**                _M_buckets;
      size_type              _M_bucket_count;
      size_type              _M_element_count;
      _RehashPolicy          _M_rehash_policy;
 
      _Node*
      _M_allocate_node(const value_type& __v);
 
      void
      _M_deallocate_node(_Node* __n);
 
      void
      _M_deallocate_nodes(_Node**, size_type);
 
      _Node**
      _M_allocate_buckets(size_type __n);
 
      void
      _M_deallocate_buckets(_Node**, size_type __n);
 
    public:			    
      // Constructor, destructor, assignment, swap
      _Hashtable(size_type __bucket_hint,
		 const _H1&, const _H2&, const _Hash&,
		 const _Equal&, const _ExtractKey&,
		 const allocator_type&);
 
      template<typename _InputIterator>
        _Hashtable(_InputIterator __first, _InputIterator __last,
		   size_type __bucket_hint,
		   const _H1&, const _H2&, const _Hash&, 
		   const _Equal&, const _ExtractKey&,
		   const allocator_type&);
 
      _Hashtable(const _Hashtable&);
 
      _Hashtable(_Hashtable&&);
 
      _Hashtable&
      operator=(const _Hashtable&);
 
      ~_Hashtable();
 
      void swap(_Hashtable&);
 
      // Basic container operations
      iterator
      begin()
      {
	iterator __i(_M_buckets);
	if (!__i._M_cur_node)
	  __i._M_incr_bucket();
	return __i;
      }
 
      const_iterator
      begin() const
      {
	const_iterator __i(_M_buckets);
	if (!__i._M_cur_node)
	  __i._M_incr_bucket();
	return __i;
      }
 
      iterator
      end()
      { return iterator(_M_buckets + _M_bucket_count); }
 
      const_iterator
      end() const
      { return const_iterator(_M_buckets + _M_bucket_count); }
 
      const_iterator
      cbegin() const
      {
	const_iterator __i(_M_buckets);
	if (!__i._M_cur_node)
	  __i._M_incr_bucket();
	return __i;
      }
 
      const_iterator
      cend() const
      { return const_iterator(_M_buckets + _M_bucket_count); }
 
      size_type
      size() const
      { return _M_element_count; }
 
      bool
      empty() const
      { return size() == 0; }
 
      allocator_type
      get_allocator() const
      { return allocator_type(_M_node_allocator); }
 
      _Value_allocator_type
      _M_get_Value_allocator() const
      { return _Value_allocator_type(_M_node_allocator); }
 
      size_type
      max_size() const
      { return _M_node_allocator.max_size(); }
 
      // Observers
      key_equal
      key_eq() const
      { return this->_M_eq; }
 
      // hash_function, if present, comes from _Hash_code_base.
 
      // Bucket operations
      size_type
      bucket_count() const
      { return _M_bucket_count; }
 
      size_type
      max_bucket_count() const
      { return max_size(); }
 
      size_type
      bucket_size(size_type __n) const
      { return std::distance(begin(__n), end(__n)); }
 
      size_type
      bucket(const key_type& __k) const
      { 
	return this->_M_bucket_index(__k, this->_M_hash_code(__k),
				     bucket_count());
      }
 
      local_iterator
      begin(size_type __n)
      { return local_iterator(_M_buckets[__n]); }
 
      local_iterator
      end(size_type)
      { return local_iterator(0); }
 
      const_local_iterator
      begin(size_type __n) const
      { return const_local_iterator(_M_buckets[__n]); }
 
      const_local_iterator
      end(size_type) const
      { return const_local_iterator(0); }
 
      // DR 691.
      const_local_iterator
      cbegin(size_type __n) const
      { return const_local_iterator(_M_buckets[__n]); }
 
      const_local_iterator
      cend(size_type) const
      { return const_local_iterator(0); }
 
      float
      load_factor() const
      { 
	return static_cast<float>(size()) / static_cast<float>(bucket_count());
      }
 
      // max_load_factor, if present, comes from _Rehash_base.
 
      // Generalization of max_load_factor.  Extension, not found in TR1.  Only
      // useful if _RehashPolicy is something other than the default.
      const _RehashPolicy&
      __rehash_policy() const
      { return _M_rehash_policy; }
 
      void 
      __rehash_policy(const _RehashPolicy&);
 
      // Lookup.
      iterator
      find(const key_type& __k);
 
      const_iterator
      find(const key_type& __k) const;
 
      size_type
      count(const key_type& __k) const;
 
      std::pair<iterator, iterator>
      equal_range(const key_type& __k);
 
      std::pair<const_iterator, const_iterator>
      equal_range(const key_type& __k) const;
 
    private:			// Find, insert and erase helper functions
      // ??? This dispatching is a workaround for the fact that we don't
      // have partial specialization of member templates; it would be
      // better to just specialize insert on __unique_keys.  There may be a
      // cleaner workaround.
      typedef typename std::conditional<__unique_keys,
					std::pair<iterator, bool>,
					iterator>::type
        _Insert_Return_Type;
 
      typedef typename std::conditional<__unique_keys,
					std::_Select1st<_Insert_Return_Type>,
					std::_Identity<_Insert_Return_Type>
                                   >::type
        _Insert_Conv_Type;
 
      _Node*
      _M_find_node(_Node*, const key_type&,
		   typename _Hashtable::_Hash_code_type) const;
 
      iterator
      _M_insert_bucket(const value_type&, size_type,
		       typename _Hashtable::_Hash_code_type);
 
      std::pair<iterator, bool>
      _M_insert(const value_type&, std::true_type);
 
      iterator
      _M_insert(const value_type&, std::false_type);
 
      void
      _M_erase_node(_Node*, _Node**);
 
    public:				
      // Insert and erase
      _Insert_Return_Type
      insert(const value_type& __v) 
      { return _M_insert(__v, std::integral_constant<bool,
			 __unique_keys>()); }
 
      iterator
      insert(const_iterator, const value_type& __v)
      { return iterator(_Insert_Conv_Type()(this->insert(__v))); }
 
      template<typename _InputIterator>
        void
        insert(_InputIterator __first, _InputIterator __last);
 
      void
      insert(initializer_list<value_type> __l)
      { this->insert(__l.begin(), __l.end()); }
 
      iterator
      erase(const_iterator);
 
      size_type
      erase(const key_type&);
 
      iterator
      erase(const_iterator, const_iterator);
 
      void
      clear();
 
      // Set number of buckets to be appropriate for container of n element.
      void rehash(size_type __n);
 
      // DR 1189.
      // reserve, if present, comes from _Rehash_base.
 
    private:
      // Unconditionally change size of bucket array to n.
      void _M_rehash(size_type __n);
    };
 
 
  // Definitions of class template _Hashtable's out-of-line member functions.
  template<typename _Key, typename _Value, 
	   typename _Allocator, typename _ExtractKey, typename _Equal,
	   typename _H1, typename _H2, typename _Hash, typename _RehashPolicy,
	   bool __chc, bool __cit, bool __uk>
    typename _Hashtable<_Key, _Value, _Allocator, _ExtractKey, _Equal,
			_H1, _H2, _Hash, _RehashPolicy,
			__chc, __cit, __uk>::_Node*
    _Hashtable<_Key, _Value, _Allocator, _ExtractKey, _Equal,
	       _H1, _H2, _Hash, _RehashPolicy, __chc, __cit, __uk>::
    _M_allocate_node(const value_type& __v)
    {
      _Node* __n = _M_node_allocator.allocate(1);
      __try
	{
	  _M_node_allocator.construct(__n, __v);
	  __n->_M_next = 0;
	  return __n;
	}
      __catch(...)
	{
	  _M_node_allocator.deallocate(__n, 1);
	  __throw_exception_again;
	}
    }
 
  template<typename _Key, typename _Value, 
	   typename _Allocator, typename _ExtractKey, typename _Equal,
	   typename _H1, typename _H2, typename _Hash, typename _RehashPolicy,
	   bool __chc, bool __cit, bool __uk>
    void
    _Hashtable<_Key, _Value, _Allocator, _ExtractKey, _Equal,
	       _H1, _H2, _Hash, _RehashPolicy, __chc, __cit, __uk>::
    _M_deallocate_node(_Node* __n)
    {
      _M_node_allocator.destroy(__n);
      _M_node_allocator.deallocate(__n, 1);
    }
 
  template<typename _Key, typename _Value, 
	   typename _Allocator, typename _ExtractKey, typename _Equal,
	   typename _H1, typename _H2, typename _Hash, typename _RehashPolicy,
	   bool __chc, bool __cit, bool __uk>
    void
    _Hashtable<_Key, _Value, _Allocator, _ExtractKey, _Equal,
	       _H1, _H2, _Hash, _RehashPolicy, __chc, __cit, __uk>::
    _M_deallocate_nodes(_Node** __array, size_type __n)
    {
      for (size_type __i = 0; __i < __n; ++__i)
	{
	  _Node* __p = __array[__i];
	  while (__p)
	    {
	      _Node* __tmp = __p;
	      __p = __p->_M_next;
	      _M_deallocate_node(__tmp);
	    }
	  __array[__i] = 0;
	}
    }
 
  template<typename _Key, typename _Value, 
	   typename _Allocator, typename _ExtractKey, typename _Equal,
	   typename _H1, typename _H2, typename _Hash, typename _RehashPolicy,
	   bool __chc, bool __cit, bool __uk>
    typename _Hashtable<_Key, _Value, _Allocator, _ExtractKey, _Equal,
			_H1, _H2, _Hash, _RehashPolicy,
			__chc, __cit, __uk>::_Node**
    _Hashtable<_Key, _Value, _Allocator, _ExtractKey, _Equal,
	       _H1, _H2, _Hash, _RehashPolicy, __chc, __cit, __uk>::
    _M_allocate_buckets(size_type __n)
    {
      _Bucket_allocator_type __alloc(_M_node_allocator);
 
      // We allocate one extra bucket to hold a sentinel, an arbitrary
      // non-null pointer.  Iterator increment relies on this.
      _Node** __p = __alloc.allocate(__n + 1);
      std::fill(__p, __p + __n, (_Node*) 0);
      __p[__n] = reinterpret_cast<_Node*>(0x1000);
      return __p;
    }
 
  template<typename _Key, typename _Value, 
	   typename _Allocator, typename _ExtractKey, typename _Equal,
	   typename _H1, typename _H2, typename _Hash, typename _RehashPolicy,
	   bool __chc, bool __cit, bool __uk>
    void
    _Hashtable<_Key, _Value, _Allocator, _ExtractKey, _Equal,
	       _H1, _H2, _Hash, _RehashPolicy, __chc, __cit, __uk>::
    _M_deallocate_buckets(_Node** __p, size_type __n)
    {
      _Bucket_allocator_type __alloc(_M_node_allocator);
      __alloc.deallocate(__p, __n + 1);
    }
 
  template<typename _Key, typename _Value, 
	   typename _Allocator, typename _ExtractKey, typename _Equal,
	   typename _H1, typename _H2, typename _Hash, typename _RehashPolicy,
	   bool __chc, bool __cit, bool __uk>
    _Hashtable<_Key, _Value, _Allocator, _ExtractKey, _Equal,
	       _H1, _H2, _Hash, _RehashPolicy, __chc, __cit, __uk>::
    _Hashtable(size_type __bucket_hint,
	       const _H1& __h1, const _H2& __h2, const _Hash& __h,
	       const _Equal& __eq, const _ExtractKey& __exk,
	       const allocator_type& __a)
    : __detail::_Rehash_base<_RehashPolicy, _Hashtable>(),
      __detail::_Hash_code_base<_Key, _Value, _ExtractKey, _Equal,
				_H1, _H2, _Hash, __chc>(__exk, __eq,
							__h1, __h2, __h),
      __detail::_Map_base<_Key, _Value, _ExtractKey, __uk, _Hashtable>(),
      _M_node_allocator(__a),
      _M_bucket_count(0),
      _M_element_count(0),
      _M_rehash_policy()
    {
      _M_bucket_count = _M_rehash_policy._M_next_bkt(__bucket_hint);
      _M_buckets = _M_allocate_buckets(_M_bucket_count);
    }
 
  template<typename _Key, typename _Value, 
	   typename _Allocator, typename _ExtractKey, typename _Equal,
	   typename _H1, typename _H2, typename _Hash, typename _RehashPolicy,
	   bool __chc, bool __cit, bool __uk>
    template<typename _InputIterator>
      _Hashtable<_Key, _Value, _Allocator, _ExtractKey, _Equal,
		 _H1, _H2, _Hash, _RehashPolicy, __chc, __cit, __uk>::
      _Hashtable(_InputIterator __f, _InputIterator __l,
		 size_type __bucket_hint,
		 const _H1& __h1, const _H2& __h2, const _Hash& __h,
		 const _Equal& __eq, const _ExtractKey& __exk,
		 const allocator_type& __a)
      : __detail::_Rehash_base<_RehashPolicy, _Hashtable>(),
	__detail::_Hash_code_base<_Key, _Value, _ExtractKey, _Equal,
				  _H1, _H2, _Hash, __chc>(__exk, __eq,
							  __h1, __h2, __h),
	__detail::_Map_base<_Key, _Value, _ExtractKey, __uk, _Hashtable>(),
	_M_node_allocator(__a),
	_M_bucket_count(0),
	_M_element_count(0),
	_M_rehash_policy()
      {
	_M_bucket_count = std::max(_M_rehash_policy._M_next_bkt(__bucket_hint),
				   _M_rehash_policy.
				   _M_bkt_for_elements(__detail::
						       __distance_fw(__f,
								     __l)));
	_M_buckets = _M_allocate_buckets(_M_bucket_count);
	__try
	  {
	    for (; __f != __l; ++__f)
	      this->insert(*__f);
	  }
	__catch(...)
	  {
	    clear();
	    _M_deallocate_buckets(_M_buckets, _M_bucket_count);
	    __throw_exception_again;
	  }
      }
 
  template<typename _Key, typename _Value, 
	   typename _Allocator, typename _ExtractKey, typename _Equal,
	   typename _H1, typename _H2, typename _Hash, typename _RehashPolicy,
	   bool __chc, bool __cit, bool __uk>
    _Hashtable<_Key, _Value, _Allocator, _ExtractKey, _Equal,
	       _H1, _H2, _Hash, _RehashPolicy, __chc, __cit, __uk>::
    _Hashtable(const _Hashtable& __ht)
    : __detail::_Rehash_base<_RehashPolicy, _Hashtable>(__ht),
      __detail::_Hash_code_base<_Key, _Value, _ExtractKey, _Equal,
				_H1, _H2, _Hash, __chc>(__ht),
      __detail::_Map_base<_Key, _Value, _ExtractKey, __uk, _Hashtable>(__ht),
      _M_node_allocator(__ht._M_node_allocator),
      _M_bucket_count(__ht._M_bucket_count),
      _M_element_count(__ht._M_element_count),
      _M_rehash_policy(__ht._M_rehash_policy)
    {
      _M_buckets = _M_allocate_buckets(_M_bucket_count);
      __try
	{
	  for (size_type __i = 0; __i < __ht._M_bucket_count; ++__i)
	    {
	      _Node* __n = __ht._M_buckets[__i];
	      _Node** __tail = _M_buckets + __i;
	      while (__n)
		{
		  *__tail = _M_allocate_node(__n->_M_v);
		  this->_M_copy_code(*__tail, __n);
		  __tail = &((*__tail)->_M_next);
		  __n = __n->_M_next;
		}
	    }
	}
      __catch(...)
	{
	  clear();
	  _M_deallocate_buckets(_M_buckets, _M_bucket_count);
	  __throw_exception_again;
	}
    }
 
  template<typename _Key, typename _Value, 
	   typename _Allocator, typename _ExtractKey, typename _Equal,
	   typename _H1, typename _H2, typename _Hash, typename _RehashPolicy,
	   bool __chc, bool __cit, bool __uk>
    _Hashtable<_Key, _Value, _Allocator, _ExtractKey, _Equal,
	       _H1, _H2, _Hash, _RehashPolicy, __chc, __cit, __uk>::
    _Hashtable(_Hashtable&& __ht)
    : __detail::_Rehash_base<_RehashPolicy, _Hashtable>(__ht),
      __detail::_Hash_code_base<_Key, _Value, _ExtractKey, _Equal,
				_H1, _H2, _Hash, __chc>(__ht),
      __detail::_Map_base<_Key, _Value, _ExtractKey, __uk, _Hashtable>(__ht),
      _M_node_allocator(__ht._M_node_allocator),
      _M_bucket_count(__ht._M_bucket_count),
      _M_element_count(__ht._M_element_count),
      _M_rehash_policy(__ht._M_rehash_policy),
      _M_buckets(__ht._M_buckets)
    {
      size_type __n_bkt = __ht._M_rehash_policy._M_next_bkt(0);
      __ht._M_buckets = __ht._M_allocate_buckets(__n_bkt);
      __ht._M_bucket_count = __n_bkt;
      __ht._M_element_count = 0;
      __ht._M_rehash_policy = _RehashPolicy();
    }
 
  template<typename _Key, typename _Value, 
	   typename _Allocator, typename _ExtractKey, typename _Equal,
	   typename _H1, typename _H2, typename _Hash, typename _RehashPolicy,
	   bool __chc, bool __cit, bool __uk>
    _Hashtable<_Key, _Value, _Allocator, _ExtractKey, _Equal,
	       _H1, _H2, _Hash, _RehashPolicy, __chc, __cit, __uk>&
    _Hashtable<_Key, _Value, _Allocator, _ExtractKey, _Equal,
	       _H1, _H2, _Hash, _RehashPolicy, __chc, __cit, __uk>::
    operator=(const _Hashtable& __ht)
    {
      _Hashtable __tmp(__ht);
      this->swap(__tmp);
      return *this;
    }
 
  template<typename _Key, typename _Value, 
	   typename _Allocator, typename _ExtractKey, typename _Equal,
	   typename _H1, typename _H2, typename _Hash, typename _RehashPolicy,
	   bool __chc, bool __cit, bool __uk>
    _Hashtable<_Key, _Value, _Allocator, _ExtractKey, _Equal,
	       _H1, _H2, _Hash, _RehashPolicy, __chc, __cit, __uk>::
    ~_Hashtable()
    {
      clear();
      _M_deallocate_buckets(_M_buckets, _M_bucket_count);
    }
 
  template<typename _Key, typename _Value, 
	   typename _Allocator, typename _ExtractKey, typename _Equal,
	   typename _H1, typename _H2, typename _Hash, typename _RehashPolicy,
	   bool __chc, bool __cit, bool __uk>
    void
    _Hashtable<_Key, _Value, _Allocator, _ExtractKey, _Equal,
	       _H1, _H2, _Hash, _RehashPolicy, __chc, __cit, __uk>::
    swap(_Hashtable& __x)
    {
      // The only base class with member variables is hash_code_base.  We
      // define _Hash_code_base::_M_swap because different specializations
      // have different members.
      __detail::_Hash_code_base<_Key, _Value, _ExtractKey, _Equal,
	_H1, _H2, _Hash, __chc>::_M_swap(__x);
 
      // _GLIBCXX_RESOLVE_LIB_DEFECTS
      // 431. Swapping containers with unequal allocators.
      std::__alloc_swap<_Node_allocator_type>::_S_do_it(_M_node_allocator,
							__x._M_node_allocator);
 
      std::swap(_M_rehash_policy, __x._M_rehash_policy);
      std::swap(_M_buckets, __x._M_buckets);
      std::swap(_M_bucket_count, __x._M_bucket_count);
      std::swap(_M_element_count, __x._M_element_count);
    }
 
  template<typename _Key, typename _Value, 
	   typename _Allocator, typename _ExtractKey, typename _Equal,
	   typename _H1, typename _H2, typename _Hash, typename _RehashPolicy,
	   bool __chc, bool __cit, bool __uk>
    void
    _Hashtable<_Key, _Value, _Allocator, _ExtractKey, _Equal,
	       _H1, _H2, _Hash, _RehashPolicy, __chc, __cit, __uk>::
    __rehash_policy(const _RehashPolicy& __pol)
    {
      _M_rehash_policy = __pol;
      size_type __n_bkt = __pol._M_bkt_for_elements(_M_element_count);
      if (__n_bkt > _M_bucket_count)
	_M_rehash(__n_bkt);
    }
 
  template<typename _Key, typename _Value, 
	   typename _Allocator, typename _ExtractKey, typename _Equal,
	   typename _H1, typename _H2, typename _Hash, typename _RehashPolicy,
	   bool __chc, bool __cit, bool __uk>
    typename _Hashtable<_Key, _Value, _Allocator, _ExtractKey, _Equal,
			_H1, _H2, _Hash, _RehashPolicy,
			__chc, __cit, __uk>::iterator
    _Hashtable<_Key, _Value, _Allocator, _ExtractKey, _Equal,
	       _H1, _H2, _Hash, _RehashPolicy, __chc, __cit, __uk>::
    find(const key_type& __k)
    {
      typename _Hashtable::_Hash_code_type __code = this->_M_hash_code(__k);
      std::size_t __n = this->_M_bucket_index(__k, __code, _M_bucket_count);
      _Node* __p = _M_find_node(_M_buckets[__n], __k, __code);
      return __p ? iterator(__p, _M_buckets + __n) : this->end();
    }
 
  template<typename _Key, typename _Value, 
	   typename _Allocator, typename _ExtractKey, typename _Equal,
	   typename _H1, typename _H2, typename _Hash, typename _RehashPolicy,
	   bool __chc, bool __cit, bool __uk>
    typename _Hashtable<_Key, _Value, _Allocator, _ExtractKey, _Equal,
			_H1, _H2, _Hash, _RehashPolicy,
			__chc, __cit, __uk>::const_iterator
    _Hashtable<_Key, _Value, _Allocator, _ExtractKey, _Equal,
	       _H1, _H2, _Hash, _RehashPolicy, __chc, __cit, __uk>::
    find(const key_type& __k) const
    {
      typename _Hashtable::_Hash_code_type __code = this->_M_hash_code(__k);
      std::size_t __n = this->_M_bucket_index(__k, __code, _M_bucket_count);
      _Node* __p = _M_find_node(_M_buckets[__n], __k, __code);
      return __p ? const_iterator(__p, _M_buckets + __n) : this->end();
    }
 
  template<typename _Key, typename _Value, 
	   typename _Allocator, typename _ExtractKey, typename _Equal,
	   typename _H1, typename _H2, typename _Hash, typename _RehashPolicy,
	   bool __chc, bool __cit, bool __uk>
    typename _Hashtable<_Key, _Value, _Allocator, _ExtractKey, _Equal,
			_H1, _H2, _Hash, _RehashPolicy,
			__chc, __cit, __uk>::size_type
    _Hashtable<_Key, _Value, _Allocator, _ExtractKey, _Equal,
	       _H1, _H2, _Hash, _RehashPolicy, __chc, __cit, __uk>::
    count(const key_type& __k) const
    {
      typename _Hashtable::_Hash_code_type __code = this->_M_hash_code(__k);
      std::size_t __n = this->_M_bucket_index(__k, __code, _M_bucket_count);
      std::size_t __result = 0;
      for (_Node* __p = _M_buckets[__n]; __p; __p = __p->_M_next)
	if (this->_M_compare(__k, __code, __p))
	  ++__result;
      return __result;
    }
 
  template<typename _Key, typename _Value, 
	   typename _Allocator, typename _ExtractKey, typename _Equal,
	   typename _H1, typename _H2, typename _Hash, typename _RehashPolicy,
	   bool __chc, bool __cit, bool __uk>
    std::pair<typename _Hashtable<_Key, _Value, _Allocator,
				  _ExtractKey, _Equal, _H1,
				  _H2, _Hash, _RehashPolicy,
				  __chc, __cit, __uk>::iterator,
	      typename _Hashtable<_Key, _Value, _Allocator,
				  _ExtractKey, _Equal, _H1,
				  _H2, _Hash, _RehashPolicy,
				  __chc, __cit, __uk>::iterator>
    _Hashtable<_Key, _Value, _Allocator, _ExtractKey, _Equal,
	       _H1, _H2, _Hash, _RehashPolicy, __chc, __cit, __uk>::
    equal_range(const key_type& __k)
    {
      typename _Hashtable::_Hash_code_type __code = this->_M_hash_code(__k);
      std::size_t __n = this->_M_bucket_index(__k, __code, _M_bucket_count);
      _Node** __head = _M_buckets + __n;
      _Node* __p = _M_find_node(*__head, __k, __code);
 
      if (__p)
	{
	  _Node* __p1 = __p->_M_next;
	  for (; __p1; __p1 = __p1->_M_next)
	    if (!this->_M_compare(__k, __code, __p1))
	      break;
 
	  iterator __first(__p, __head);
	  iterator __last(__p1, __head);
	  if (!__p1)
	    __last._M_incr_bucket();
	  return std::make_pair(__first, __last);
	}
      else
	return std::make_pair(this->end(), this->end());
    }
 
  template<typename _Key, typename _Value, 
	   typename _Allocator, typename _ExtractKey, typename _Equal,
	   typename _H1, typename _H2, typename _Hash, typename _RehashPolicy,
	   bool __chc, bool __cit, bool __uk>
    std::pair<typename _Hashtable<_Key, _Value, _Allocator,
				  _ExtractKey, _Equal, _H1,
				  _H2, _Hash, _RehashPolicy,
				  __chc, __cit, __uk>::const_iterator,
	      typename _Hashtable<_Key, _Value, _Allocator,
				  _ExtractKey, _Equal, _H1,
				  _H2, _Hash, _RehashPolicy,
				  __chc, __cit, __uk>::const_iterator>
    _Hashtable<_Key, _Value, _Allocator, _ExtractKey, _Equal,
	       _H1, _H2, _Hash, _RehashPolicy, __chc, __cit, __uk>::
    equal_range(const key_type& __k) const
    {
      typename _Hashtable::_Hash_code_type __code = this->_M_hash_code(__k);
      std::size_t __n = this->_M_bucket_index(__k, __code, _M_bucket_count);
      _Node** __head = _M_buckets + __n;
      _Node* __p = _M_find_node(*__head, __k, __code);
 
      if (__p)
	{
	  _Node* __p1 = __p->_M_next;
	  for (; __p1; __p1 = __p1->_M_next)
	    if (!this->_M_compare(__k, __code, __p1))
	      break;
 
	  const_iterator __first(__p, __head);
	  const_iterator __last(__p1, __head);
	  if (!__p1)
	    __last._M_incr_bucket();
	  return std::make_pair(__first, __last);
	}
      else
	return std::make_pair(this->end(), this->end());
    }
 
  // Find the node whose key compares equal to k, beginning the search
  // at p (usually the head of a bucket).  Return nil if no node is found.
  template<typename _Key, typename _Value, 
	   typename _Allocator, typename _ExtractKey, typename _Equal,
	   typename _H1, typename _H2, typename _Hash, typename _RehashPolicy,
	   bool __chc, bool __cit, bool __uk>
    typename _Hashtable<_Key, _Value, _Allocator, _ExtractKey,
			_Equal, _H1, _H2, _Hash, _RehashPolicy,
			__chc, __cit, __uk>::_Node* 
    _Hashtable<_Key, _Value, _Allocator, _ExtractKey, _Equal,
	       _H1, _H2, _Hash, _RehashPolicy, __chc, __cit, __uk>::
    _M_find_node(_Node* __p, const key_type& __k,
		typename _Hashtable::_Hash_code_type __code) const
    {
      for (; __p; __p = __p->_M_next)
	if (this->_M_compare(__k, __code, __p))
	  return __p;
      return false;
    }
 
  // Insert v in bucket n (assumes no element with its key already present).
  template<typename _Key, typename _Value, 
	   typename _Allocator, typename _ExtractKey, typename _Equal,
	   typename _H1, typename _H2, typename _Hash, typename _RehashPolicy,
	   bool __chc, bool __cit, bool __uk>
    typename _Hashtable<_Key, _Value, _Allocator, _ExtractKey, _Equal,
			_H1, _H2, _Hash, _RehashPolicy,
			__chc, __cit, __uk>::iterator
    _Hashtable<_Key, _Value, _Allocator, _ExtractKey, _Equal,
	       _H1, _H2, _Hash, _RehashPolicy, __chc, __cit, __uk>::
    _M_insert_bucket(const value_type& __v, size_type __n,
		     typename _Hashtable::_Hash_code_type __code)
    {
      std::pair<bool, std::size_t> __do_rehash
	= _M_rehash_policy._M_need_rehash(_M_bucket_count,
					  _M_element_count, 1);
 
      // Allocate the new node before doing the rehash so that we don't
      // do a rehash if the allocation throws.
      _Node* __new_node = _M_allocate_node(__v);
 
      __try
	{
	  if (__do_rehash.first)
	    {
	      const key_type& __k = this->_M_extract(__v);
	      __n = this->_M_bucket_index(__k, __code, __do_rehash.second);
	      _M_rehash(__do_rehash.second);
	    }
 
	  __new_node->_M_next = _M_buckets[__n];
	  this->_M_store_code(__new_node, __code);
	  _M_buckets[__n] = __new_node;
	  ++_M_element_count;
	  return iterator(__new_node, _M_buckets + __n);
	}
      __catch(...)
	{
	  _M_deallocate_node(__new_node);
	  __throw_exception_again;
	}
    }
 
  // Insert v if no element with its key is already present.
  template<typename _Key, typename _Value, 
	   typename _Allocator, typename _ExtractKey, typename _Equal,
	   typename _H1, typename _H2, typename _Hash, typename _RehashPolicy,
	   bool __chc, bool __cit, bool __uk>
    std::pair<typename _Hashtable<_Key, _Value, _Allocator,
				  _ExtractKey, _Equal, _H1,
				  _H2, _Hash, _RehashPolicy,
				  __chc, __cit, __uk>::iterator, bool>
    _Hashtable<_Key, _Value, _Allocator, _ExtractKey, _Equal,
	       _H1, _H2, _Hash, _RehashPolicy, __chc, __cit, __uk>::
    _M_insert(const value_type& __v, std::true_type)
    {
      const key_type& __k = this->_M_extract(__v);
      typename _Hashtable::_Hash_code_type __code = this->_M_hash_code(__k);
      size_type __n = this->_M_bucket_index(__k, __code, _M_bucket_count);
 
      if (_Node* __p = _M_find_node(_M_buckets[__n], __k, __code))
	return std::make_pair(iterator(__p, _M_buckets + __n), false);
      return std::make_pair(_M_insert_bucket(__v, __n, __code), true);
    }
 
  // Insert v unconditionally.
  template<typename _Key, typename _Value, 
	   typename _Allocator, typename _ExtractKey, typename _Equal,
	   typename _H1, typename _H2, typename _Hash, typename _RehashPolicy,
	   bool __chc, bool __cit, bool __uk>
    typename _Hashtable<_Key, _Value, _Allocator, _ExtractKey, _Equal,
			_H1, _H2, _Hash, _RehashPolicy,
			__chc, __cit, __uk>::iterator
    _Hashtable<_Key, _Value, _Allocator, _ExtractKey, _Equal,
	       _H1, _H2, _Hash, _RehashPolicy, __chc, __cit, __uk>::
    _M_insert(const value_type& __v, std::false_type)
    {
      std::pair<bool, std::size_t> __do_rehash
	= _M_rehash_policy._M_need_rehash(_M_bucket_count,
					  _M_element_count, 1);
      if (__do_rehash.first)
	_M_rehash(__do_rehash.second);
 
      const key_type& __k = this->_M_extract(__v);
      typename _Hashtable::_Hash_code_type __code = this->_M_hash_code(__k);
      size_type __n = this->_M_bucket_index(__k, __code, _M_bucket_count);
 
      // First find the node, avoid leaking new_node if compare throws.
      _Node* __prev = _M_find_node(_M_buckets[__n], __k, __code);
      _Node* __new_node = _M_allocate_node(__v);
 
      if (__prev)
	{
	  __new_node->_M_next = __prev->_M_next;
	  __prev->_M_next = __new_node;
	}
      else
	{
	  __new_node->_M_next = _M_buckets[__n];
	  _M_buckets[__n] = __new_node;
	}
      this->_M_store_code(__new_node, __code);
 
      ++_M_element_count;
      return iterator(__new_node, _M_buckets + __n);
    }
 
  // For erase(iterator) and erase(const_iterator).
  template<typename _Key, typename _Value, 
	   typename _Allocator, typename _ExtractKey, typename _Equal,
	   typename _H1, typename _H2, typename _Hash, typename _RehashPolicy,
	   bool __chc, bool __cit, bool __uk>
    void
    _Hashtable<_Key, _Value, _Allocator, _ExtractKey, _Equal,
	       _H1, _H2, _Hash, _RehashPolicy, __chc, __cit, __uk>::
    _M_erase_node(_Node* __p, _Node** __b)
    {
      _Node* __cur = *__b;
      if (__cur == __p)
	*__b = __cur->_M_next;
      else
	{
	  _Node* __next = __cur->_M_next;
	  while (__next != __p)
	    {
	      __cur = __next;
	      __next = __cur->_M_next;
	    }
	  __cur->_M_next = __next->_M_next;
	}
 
      _M_deallocate_node(__p);
      --_M_element_count;
    }
 
  template<typename _Key, typename _Value, 
	   typename _Allocator, typename _ExtractKey, typename _Equal,
	   typename _H1, typename _H2, typename _Hash, typename _RehashPolicy,
	   bool __chc, bool __cit, bool __uk>
    template<typename _InputIterator>
      void 
      _Hashtable<_Key, _Value, _Allocator, _ExtractKey, _Equal,
		 _H1, _H2, _Hash, _RehashPolicy, __chc, __cit, __uk>::
      insert(_InputIterator __first, _InputIterator __last)
      {
	size_type __n_elt = __detail::__distance_fw(__first, __last);
	std::pair<bool, std::size_t> __do_rehash
	  = _M_rehash_policy._M_need_rehash(_M_bucket_count,
					    _M_element_count, __n_elt);
	if (__do_rehash.first)
	  _M_rehash(__do_rehash.second);
 
	for (; __first != __last; ++__first)
	  this->insert(*__first);
      }
 
  template<typename _Key, typename _Value, 
	   typename _Allocator, typename _ExtractKey, typename _Equal,
	   typename _H1, typename _H2, typename _Hash, typename _RehashPolicy,
	   bool __chc, bool __cit, bool __uk>
    typename _Hashtable<_Key, _Value, _Allocator, _ExtractKey, _Equal,
			_H1, _H2, _Hash, _RehashPolicy,
			__chc, __cit, __uk>::iterator
    _Hashtable<_Key, _Value, _Allocator, _ExtractKey, _Equal,
	       _H1, _H2, _Hash, _RehashPolicy, __chc, __cit, __uk>::
    erase(const_iterator __it)
    {
      iterator __result(__it._M_cur_node, __it._M_cur_bucket);
      ++__result;
      _M_erase_node(__it._M_cur_node, __it._M_cur_bucket);
      return __result;
    }
 
  template<typename _Key, typename _Value, 
	   typename _Allocator, typename _ExtractKey, typename _Equal,
	   typename _H1, typename _H2, typename _Hash, typename _RehashPolicy,
	   bool __chc, bool __cit, bool __uk>
    typename _Hashtable<_Key, _Value, _Allocator, _ExtractKey, _Equal,
			_H1, _H2, _Hash, _RehashPolicy,
			__chc, __cit, __uk>::size_type
    _Hashtable<_Key, _Value, _Allocator, _ExtractKey, _Equal,
	       _H1, _H2, _Hash, _RehashPolicy, __chc, __cit, __uk>::
    erase(const key_type& __k)
    {
      typename _Hashtable::_Hash_code_type __code = this->_M_hash_code(__k);
      std::size_t __n = this->_M_bucket_index(__k, __code, _M_bucket_count);
      size_type __result = 0;
 
      _Node** __slot = _M_buckets + __n;
      while (*__slot && !this->_M_compare(__k, __code, *__slot))
	__slot = &((*__slot)->_M_next);
 
      _Node** __saved_slot = 0;
      while (*__slot && this->_M_compare(__k, __code, *__slot))
	{
	  // _GLIBCXX_RESOLVE_LIB_DEFECTS
	  // 526. Is it undefined if a function in the standard changes
	  // in parameters?
	  if (&this->_M_extract((*__slot)->_M_v) != &__k)
	    {
              _Node* __p = *__slot;
              *__slot = __p->_M_next;
	      _M_deallocate_node(__p);
	      --_M_element_count;
	      ++__result;
	    }
	  else
	    {
	      __saved_slot = __slot;
	      __slot = &((*__slot)->_M_next);
	    }
	}
 
      if (__saved_slot)
	{
	  _Node* __p = *__saved_slot;
	  *__saved_slot = __p->_M_next;
	  _M_deallocate_node(__p);
	  --_M_element_count;
	  ++__result;
	}
 
      return __result;
    }
 
  // ??? This could be optimized by taking advantage of the bucket
  // structure, but it's not clear that it's worth doing.  It probably
  // wouldn't even be an optimization unless the load factor is large.
  template<typename _Key, typename _Value, 
	   typename _Allocator, typename _ExtractKey, typename _Equal,
	   typename _H1, typename _H2, typename _Hash, typename _RehashPolicy,
	   bool __chc, bool __cit, bool __uk>
    typename _Hashtable<_Key, _Value, _Allocator, _ExtractKey, _Equal,
			_H1, _H2, _Hash, _RehashPolicy,
			__chc, __cit, __uk>::iterator
    _Hashtable<_Key, _Value, _Allocator, _ExtractKey, _Equal,
	       _H1, _H2, _Hash, _RehashPolicy, __chc, __cit, __uk>::
    erase(const_iterator __first, const_iterator __last)
    {
      while (__first != __last)
	__first = this->erase(__first);
      return iterator(__last._M_cur_node, __last._M_cur_bucket);
    }
 
  template<typename _Key, typename _Value, 
	   typename _Allocator, typename _ExtractKey, typename _Equal,
	   typename _H1, typename _H2, typename _Hash, typename _RehashPolicy,
	   bool __chc, bool __cit, bool __uk>
    void
    _Hashtable<_Key, _Value, _Allocator, _ExtractKey, _Equal,
	       _H1, _H2, _Hash, _RehashPolicy, __chc, __cit, __uk>::
    clear()
    {
      _M_deallocate_nodes(_M_buckets, _M_bucket_count);
      _M_element_count = 0;
    }
 
  template<typename _Key, typename _Value, 
	   typename _Allocator, typename _ExtractKey, typename _Equal,
	   typename _H1, typename _H2, typename _Hash, typename _RehashPolicy,
	   bool __chc, bool __cit, bool __uk>
    void
    _Hashtable<_Key, _Value, _Allocator, _ExtractKey, _Equal,
	       _H1, _H2, _Hash, _RehashPolicy, __chc, __cit, __uk>::
    rehash(size_type __n)
    {
      _M_rehash(std::max(_M_rehash_policy._M_next_bkt(__n),
			 _M_rehash_policy._M_bkt_for_elements(_M_element_count
							      + 1)));
    }
 
  template<typename _Key, typename _Value, 
	   typename _Allocator, typename _ExtractKey, typename _Equal,
	   typename _H1, typename _H2, typename _Hash, typename _RehashPolicy,
	   bool __chc, bool __cit, bool __uk>
    void
    _Hashtable<_Key, _Value, _Allocator, _ExtractKey, _Equal,
	       _H1, _H2, _Hash, _RehashPolicy, __chc, __cit, __uk>::
    _M_rehash(size_type __n)
    {
      _Node** __new_array = _M_allocate_buckets(__n);
      __try
	{
	  for (size_type __i = 0; __i < _M_bucket_count; ++__i)
	    while (_Node* __p = _M_buckets[__i])
	      {
		std::size_t __new_index = this->_M_bucket_index(__p, __n);
		_M_buckets[__i] = __p->_M_next;
		__p->_M_next = __new_array[__new_index];
		__new_array[__new_index] = __p;
	      }
	  _M_deallocate_buckets(_M_buckets, _M_bucket_count);
	  _M_bucket_count = __n;
	  _M_buckets = __new_array;
	}
      __catch(...)
	{
	  // A failure here means that a hash function threw an exception.
	  // We can't restore the previous state without calling the hash
	  // function again, so the only sensible recovery is to delete
	  // everything.
	  _M_deallocate_nodes(__new_array, __n);
	  _M_deallocate_buckets(__new_array, __n);
	  _M_deallocate_nodes(_M_buckets, _M_bucket_count);
	  _M_element_count = 0;
	  __throw_exception_again;
	}
    }
}
 
#endif // _HASHTABLE_H
 

Compare with Previous | Blame | View Log

powered by: WebSVN 2.1.0

© copyright 1999-2024 OpenCores.org, equivalent to Oliscience, all rights reserved. OpenCores®, registered trademark.