C++官方自带可持久化平衡树rope的3000行源码

最新推荐文章于 2024-07-03 12:33:20 发布
最新推荐文章于 2024-07-03 12:33:20 发布
阅读量1k 收藏 3
点赞数
分类专栏： “类”数据结构 - 强大的pbds和rope
版权声明：本文为博主原创文章，遵循 CC 4.0 BY-NC-SA 版权协议，转载请附上原文出处链接和本声明。
本文链接：https://blog.csdn.net/weixin_45697774/article/details/109608491
版权
本文将深入探讨C++标准库中的rope数据结构，这是一种实现了可持久化的平衡树，用于高效处理大量文本操作。通过分析3000行源码，我们将理解其内部工作机制和优势。
摘要由CSDN通过智能技术生成

C++官方rope3000行源码

// SGI's rope class -*- C++ -*-

// Copyright (C) 2001-2015 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/>.

/*
 * Copyright (c) 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.
 */

/** @file ext/rope
 *  This file is a GNU extension to the Standard C++ Library (possibly
 *  containing extensions from the HP/SGI STL subset). 
 */

#ifndef _ROPE
#define _ROPE 1

#pragma GCC system_header

#include <algorithm>
#include <iosfwd>
#include <bits/stl_construct.h>
#include <bits/stl_uninitialized.h>
#include <bits/stl_function.h>
#include <bits/stl_numeric.h>
#include <bits/allocator.h>
#include <bits/gthr.h>
#include <tr1/functional>

# ifdef __GC
#   define __GC_CONST const
# else
#   define __GC_CONST   // constant except for deallocation
# endif

#include <ext/memory> // For uninitialized_copy_n

namespace __gnu_cxx _GLIBCXX_VISIBILITY(default)
{
   
  namespace __detail
  {
   
    enum {
    _S_max_rope_depth = 45 };
    enum _Tag {
   _S_leaf, _S_concat, _S_substringfn, _S_function};
  } // namespace __detail

  using std::size_t;
  using std::ptrdiff_t;
  using std::allocator;
  using std::_Destroy;

_GLIBCXX_BEGIN_NAMESPACE_VERSION

  // See libstdc++/36832.
  template<typename _ForwardIterator, typename _Allocator>
    void
    _Destroy_const(_ForwardIterator __first,
		   _ForwardIterator __last, _Allocator __alloc)
    {
   
      for (; __first != __last; ++__first)
	__alloc.destroy(&*__first);
    }

  template<typename _ForwardIterator, typename _Tp>
    inline void
    _Destroy_const(_ForwardIterator __first,
		   _ForwardIterator __last, allocator<_Tp>)
    {
    _Destroy(__first, __last); }

  // The _S_eos function is used for those functions that
  // convert to/from C-like strings to detect the end of the string.
  
  // The end-of-C-string character.
  // This is what the draft standard says it should be.
  template <class _CharT>
    inline _CharT
    _S_eos(_CharT*)
    {
    return _CharT(); }

  // Test for basic character types.
  // For basic character types leaves having a trailing eos.
  template <class _CharT>
    inline bool
    _S_is_basic_char_type(_CharT*)
    {
    return false; }
  
  template <class _CharT>
    inline bool
    _S_is_one_byte_char_type(_CharT*)
    {
    return false; }

  inline bool
  _S_is_basic_char_type(char*)
  {
    return true; }
  
  inline bool
  _S_is_one_byte_char_type(char*)
  {
    return true; }
  
  inline bool
  _S_is_basic_char_type(wchar_t*)
  {
    return true; }

  // Store an eos iff _CharT is a basic character type.
  // Do not reference _S_eos if it isn't.
  template <class _CharT>
    inline void
    _S_cond_store_eos(_CharT&) {
    }

  inline void
  _S_cond_store_eos(char& __c)
  {
    __c = 0; }

  inline void
  _S_cond_store_eos(wchar_t& __c)
  {
    __c = 0; }

  // char_producers are logically functions that generate a section of
  // a string.  These can be converted to ropes.  The resulting rope
  // invokes the char_producer on demand.  This allows, for example,
  // files to be viewed as ropes without reading the entire file.
  template <class _CharT>
    class char_producer
    {
   
    public:
      virtual ~char_producer() {
    };

      virtual void
      operator()(size_t __start_pos, size_t __len,
		 _CharT* __buffer) = 0;
      // Buffer should really be an arbitrary output iterator.
      // That way we could flatten directly into an ostream, etc.
      // This is thoroughly impossible, since iterator types don't
      // have runtime descriptions.
    };

  // Sequence buffers:
  //
  // Sequence must provide an append operation that appends an
  // array to the sequence.  Sequence buffers are useful only if
  // appending an entire array is cheaper than appending element by element.
  // This is true for many string representations.
  // This should  perhaps inherit from ostream<sequence::value_type>
  // and be implemented correspondingly, so that they can be used
  // for formatted.  For the sake of portability, we don't do this yet.
  //
  // For now, sequence buffers behave as output iterators.  But they also
  // behave a little like basic_ostringstream<sequence::value_type> and a
  // little like containers.

  template<class _Sequence, size_t _Buf_sz = 100>
    class sequence_buffer
    : public std::iterator<std::output_iterator_tag, void, void, void, void>
    {
   
    public:
      typedef typename _Sequence::value_type value_type;
    protected:
      _Sequence* _M_prefix;
      value_type _M_buffer[_Buf_sz];
      size_t     _M_buf_count;
    public:

      void
      flush()
      {
   
	_M_prefix->append(_M_buffer, _M_buffer + _M_buf_count);
	_M_buf_count = 0;
      }
      
      ~sequence_buffer()
      {
    flush(); }
      
      sequence_buffer()
      : _M_prefix(0), _M_buf_count(0) {
    }

      sequence_buffer(const sequence_buffer& __x)
      {
   
	_M_prefix = __x._M_prefix;
	_M_buf_count = __x._M_buf_count;
	std::copy(__x._M_buffer, __x._M_buffer + __x._M_buf_count, _M_buffer);
      }
      
      sequence_buffer(sequence_buffer& __x)
      {
   
	__x.flush();
	_M_prefix = __x._M_prefix;
	_M_buf_count = 0;
      }
      
      sequence_buffer(_Sequence& __s)
      : _M_prefix(&__s), _M_buf_count(0) {
    }
      
      sequence_buffer&
      operator=(sequence_buffer& __x)
      {
   
	__x.flush();
	_M_prefix = __x._M_prefix;
	_M_buf_count = 0;
	return *this;
      }

      sequence_buffer&
      operator=(const sequence_buffer& __x)
      {
   
	_M_prefix = __x._M_prefix;
	_M_buf_count = __x._M_buf_count;
	std::copy(__x._M_buffer, __x._M_buffer + __x._M_buf_count, _M_buffer);
	return *this;
      }
      
      void
      push_back(value_type __x)
      {
   
	if (_M_buf_count < _Buf_sz)
	  {
   
	    _M_buffer[_M_buf_count] = __x;
	    ++_M_buf_count;
	  }
	else
	  {
   
	    flush();
	    _M_buffer[0] = __x;
	    _M_buf_count = 1;
	  }
      }
      
      void
      append(value_type* __s, size_t __len)
      {
   
	if (__len + _M_buf_count <= _Buf_sz)
	  {
   
	    size_t __i = _M_buf_count;
	    for (size_t __j = 0; __j < __len; __i++, __j++)
	      _M_buffer[__i] = __s[__j];
	    _M_buf_count += __len;
	  }
	else if (0 == _M_buf_count)
	  _M_prefix->append(__s, __s + __len);
	else
	  {
   
	    flush();
	    append(__s, __len);
	  }
      }

      sequence_buffer&
      write(value_type* __s, size_t __len)
      {
   
	append(__s, __len);
	return *this;
      }
      
      sequence_buffer&
      put(value_type __x)
      {
   
	push_back(__x);
	return *this;
      }
      
      sequence_buffer&
      operator=(const value_type& __rhs)
      {
   
	push_back(__rhs);
	return *this;
      }
      
      sequence_buffer&
      operator*()
      {
    return *this; }
      
      sequence_buffer&
      operator++()
      {
    return *this; }
      
      sequence_buffer
      operator++(int)
      {
    return *this; }
    };
  
  // The following should be treated as private, at least for now.
  template<class _CharT>
    class _Rope_char_consumer
    {
   
    public:
      // If we had member templates, these should not be virtual.
      // For now we need to use run-time parametrization where
      // compile-time would do.  Hence this should all be private
      // for now.
      // The symmetry with char_producer is accidental and temporary.
      virtual ~_Rope_char_consumer() {
    };
  
      virtual bool
      operator()(const _CharT* __buffer, size_t __len) = 0;
    };
  
  // First a lot of forward declarations.  The standard seems to require
  // much stricter "declaration before use" than many of the implementations
  // that preceded it.
  template<class _CharT, class _Alloc = allocator<_CharT> >
    class rope;
  
  template<class _CharT, class _Alloc>
    struct _Rope_RopeConcatenation;

  template<class _CharT, class _Alloc>
    struct _Rope_RopeLeaf;
  
  template<class _CharT, class _Alloc>
    struct _Rope_RopeFunction;
  
  template<class _CharT, class _Alloc>
    struct _Rope_RopeSubstring;
  
  template<class _CharT, class _Alloc>
    class _Rope_iterator;
  
  template<class _CharT, class _Alloc>
    class _Rope_const_iterator;
  
  template<class _CharT, class _Alloc>
    class _Rope_char_ref_proxy;
  
  template<class _CharT, class _Alloc>
    class _Rope_char_ptr_proxy;

  template<class _CharT, class _Alloc>
    bool
    operator==(const _Rope_char_ptr_proxy<_CharT, _Alloc>& __x,
	       const _Rope_char_ptr_proxy<_CharT, _Alloc>& __y);

  template<class _CharT, class _Alloc>
    _Rope_const_iterator<_CharT, _Alloc>
    operator-(const _Rope_const_iterator<_CharT, _Alloc>& __x,
	      ptrdiff_t __n);

  template<class _CharT, class _Alloc>
    _Rope_const_iterator<_CharT, _Alloc>
    operator+(const _Rope_const_iterator<_CharT, _Alloc>& __x,
	      ptrdiff_t __n);

  template<class _CharT, class _Alloc>
    _Rope_const_iterator<_CharT, _Alloc>
    operator+(ptrdiff_t __n,
	      const _Rope_const_iterator<_CharT, _Alloc>& __x);

  template<class _CharT, class _Alloc>
    bool
    operator==(const _Rope_const_iterator<_CharT, _Alloc>& __x,
	       const _Rope_const_iterator<_CharT, _Alloc>& __y);

  template<class _CharT, class _Alloc>
    bool
    operator<(const _Rope_const_iterator<_CharT, _Alloc>& __x,
	      const _Rope_const_iterator<_CharT, _Alloc>& __y);
  
  template<class _CharT, class _Alloc>
    ptrdiff_t
    operator-(const _Rope_const_iterator<_CharT, _Alloc>& __x,
	      const _Rope_const_iterator<_CharT, _Alloc>& __y);

  template<class _CharT, class _Alloc>
    _Rope_iterator<_CharT, _Alloc>
    operator-(const _Rope_iterator<_CharT, _Alloc>& __x, ptrdiff_t __n);

  template<class _CharT, class _Alloc>
    _Rope_iterator<_CharT, _Alloc>
    operator+(const _Rope_iterator<_CharT, _Alloc>& __x, ptrdiff_t __n);

  template<class _CharT, class _Alloc>
    _Rope_iterator<_CharT, _Alloc>
    operator+(ptrdiff_t __n, const _Rope_iterator<_CharT, _Alloc>& __x);

  template<class _CharT, class _Alloc>
    bool
    operator==(const _Rope_iterator<_CharT, _Alloc>& __x,
	       const _Rope_iterator<_CharT, _Alloc>& __y);

  template<class _CharT, class _Alloc>
    bool
    operator<(const _Rope_iterator<_CharT, _Alloc>& __x,
	      const _Rope_iterator<_CharT, _Alloc>& __y);

  template<class _CharT, class _Alloc>
    ptrdiff_t
    operator-(const _Rope_iterator<_CharT, _Alloc>& __x,
	      const _Rope_iterator<_CharT, _Alloc>& __y);

  template<class _CharT, class _Alloc>
    rope<_CharT, _Alloc>
    operator+(const rope<_CharT, _Alloc>& __left,
	      const rope<_CharT, _Alloc>& __right);

  template<class _CharT, class _Alloc>
    rope<_CharT, _Alloc>
    operator+(const rope<_CharT, _Alloc>& __left, const _CharT* __right);

  template<class _CharT, class _Alloc>
    rope<_CharT, _Alloc>
    operator+(const rope<_CharT, _Alloc>& __left, _CharT __right);

  // Some helpers, so we can use power on ropes.
  // See below for why this isn't local to the implementation.
  
  // This uses a nonstandard refcount convention.
  // The result has refcount 0.
  template<class _CharT, class _Alloc>
    struct _Rope_Concat_fn
    : public std::binary_function<rope<_CharT, _Alloc>, rope<_CharT, _Alloc>,
				  rope<_CharT, _Alloc> >
    {
   
      rope<_CharT, _Alloc>
      operator()(const rope<_CharT, _Alloc>& __x,
		 const rope<_CharT, _Alloc>& __y)
      {
    return __x + __y; }
    };

  template <class _CharT, class _Alloc>
    inline rope<_CharT, _Alloc>
    identity_element(_Rope_Concat_fn<_CharT, _Alloc>)
    {
    return rope<_CharT, _Alloc>(); }

  // Class _Refcount_Base provides a type, _RC_t, a data member,
  // _M_ref_count, and member functions _M_incr and _M_decr, which perform
  // atomic preincrement/predecrement.  The constructor initializes
  // _M_ref_count.
  struct _Refcount_Base
  {
   
    // The type _RC_t
    typedef size_t _RC_t;
    
    // The data member _M_ref_count
    volatile _RC_t _M_ref_count;

    // Constructor
#ifdef __GTHREAD_MUTEX_INIT
    __gthread_mutex_t _M_ref_count_lock = __GTHREAD_MUTEX_INIT;
#else
    __gthread_mutex_t _M_ref_count_lock;
#endif

    _Refcount_Base(_RC_t __n) : _M_ref_count(__n)
    {
   
#ifndef __GTHREAD_MUTEX_INIT
#ifdef __GTHREAD_MUTEX_INIT_FUNCTION
      __GTHREAD_MUTEX_INIT_FUNCTION (&_M_ref_count_lock);
#else
#error __GTHREAD_MUTEX_INIT or __GTHREAD_MUTEX_INIT_FUNCTION should be defined by gthr.h abstraction layer, report problem to libstdc++@gcc.gnu.org.
#endif
#endif
    }

#ifndef __GTHREAD_MUTEX_INIT
    ~_Refcount_Base()
    {
    __gthread_mutex_destroy(&_M_ref_count_lock); }
#endif

    void
    _M_incr()
    {
   
      __gthread_mutex_lock(&_M_ref_count_lock);
      ++_M_ref_count;
      __gthread_mutex_unlock(&_M_ref_count_lock);
    }

    _RC_t
    _M_decr()
    {
   
      __gthread_mutex_lock(&_M_ref_count_lock);
      volatile _RC_t __tmp = --_M_ref_count;
      __gthread_mutex_unlock(&_M_ref_count_lock);
      return __tmp;
    }
  };

  //
  // What follows should really be local to rope.  Unfortunately,
  // that doesn't work, since it makes it impossible to define generic
  // equality on rope iterators.  According to the draft standard, the
  // template parameters for such an equality operator cannot be inferred
  // from the occurrence of a member class as a parameter.
  // (SGI compilers in fact allow this, but the __result wouldn't be
  // portable.)
  // Similarly, some of the static member functions are member functions
  // only to avoid polluting the global namespace, and to circumvent
  // restrictions on type inference for template functions.
  //

  //
  // The internal data structure for representing a rope.  This is
  // private to the implementation.  A rope is really just a pointer
  // to one of these.
  //
  // A few basic functions for manipulating this data structure
  // are members of _RopeRep.  Most of the more complex algorithms
  // are implemented as rope members.
  //
  // Some of the static member functions of _RopeRep have identically
  // named functions in rope that simply invoke the _RopeRep versions.

#define __ROPE_DEFINE_ALLOCS(__a) \
        __ROPE_DEFINE_ALLOC(_CharT,_Data) /* character data */ \
        typedef _Rope_RopeConcatenation<_CharT,__a> __C; \
        __ROPE_DEFINE_ALLOC(__C,_C) \
        typedef _Rope_RopeLeaf<_CharT,__a> __L; \
        __ROPE_DEFINE_ALLOC(__L,_L) \
        typedef _Rope_RopeFunction<_CharT,__a> __F; \
        __ROPE_DEFINE_ALLOC(__F,_F) \
        typedef _Rope_RopeSubstring<_CharT,__a> __S; \
        __ROPE_DEFINE_ALLOC(__S,_S)

  //  Internal rope nodes potentially store a copy of the allocator
  //  instance used to allocate them.  This is mostly redundant.
  //  But the alternative would be to pass allocator instances around
  //  in some form to nearly all internal functions, since any pointer
  //  assignment may result in a zero reference count and thus require
  //  deallocation.

#define __STATIC_IF_SGI_ALLOC  /* not static */

  template <class _CharT, class _Alloc>
    struct _Rope_rep_base
    : public _Alloc
    {
   
      typedef _Alloc allocator_type;

      allocator_type
      get_allocator() const
      {
    return *static_cast<const _Alloc*>(this); }

      allocator_type&
      _M_get_allocator()
      {
    return *static_cast<_Alloc*>(this); }

      const allocator_type&
      _M_get_allocator() const
      {
    return *static_cast<const _Alloc*>(this); }

      _Rope_rep_base(size_t __size, const allocator_type&)
      : _M_size(__size) {
    }

      size_t _M_size;

# define __ROPE_DEFINE_ALLOC(_Tp, __name) \
        typedef typename \
          _Alloc::template rebind<_Tp>::other __name##Alloc; \
        static _Tp* __name##_allocate(size_t __n) \
          { return __name##Alloc().allocate(__n); } \
        static void __name##_deallocate(_Tp *__p, size_t __n) \
          { __name##Alloc().deallocate(__p, __n); }
      __ROPE_DEFINE_ALLOCS(_Alloc)
# undef __ROPE_DEFINE_ALLOC
    };

  template<class _CharT, class _Alloc>
    struct _Rope_RopeRep
    : public _Rope_rep_base<_CharT, _Alloc>
# ifndef __GC
	     , _Refcount_Base
# endif
    {
   
    public:
      __detail::_Tag _M_tag:8;
      bool _M_is_balanced:8;
      unsigned char _M_depth;
      __GC_CONST _CharT* _M_c_string;
#ifdef __GTHREAD_MUTEX_INIT
      __gthread_mutex_t _M_c_string_lock = __GTHREAD_MUTEX_INIT;
#else
      __gthread_mutex_t _M_c_string_lock;
#endif
                        /* Flattened version of string, if needed.  */
                        /* typically 0.                             */
                        /* If it's not 0, then the memory is owned  */
                        /* by this node.                            */
                        /* In the case of a leaf, this may point to */
                        /* the same memory as the data field.       */
      typedef typename _Rope_rep_base<_CharT, _Alloc>::allocator_type
        allocator_type;

      using _Rope_rep_base<_CharT, _Alloc>::get_allocator;
      using _Rope_rep_base<_CharT, _Alloc>::_M_get_allocator;

      _Rope_RopeRep(__detail::_Tag __t, int __d, bool __b, size_t __size,
		    const allocator_type& __a)
      : _Rope_rep_base<_CharT, _Alloc>(__size, __a),
#ifndef __GC
	_Refcount_Base(1),
#endif
	_M_tag(__t), _M_is_balanced(__b), _M_depth(__d), _M_c_string(0)
#ifdef __GTHREAD_MUTEX_INIT
      {
    }
#else
      {
    __GTHREAD_MUTEX_INIT_FUNCTION (&_M_c_string_lock); }
      ~_Rope_RopeRep()
      {
    __gthread_mutex_destroy (&_M_c_string_lock); }
#endif
#ifdef __GC
      void
      _M_incr () {
    }
#endif
      static void
      _S_free_string(__GC_CONST _CharT*, size_t __len,
		     allocator_type& __a);
#define __STL_FREE_STRING(__s, __l, __a) _S_free_string(__s, __l, __a);
                        // Deallocate data section of a leaf.
                        // This shouldn't be a member function.
                        // But its hard to do anything else at the
                        // moment, because it's templatized w.r.t.
                        // an allocator.
                        // Does nothing if __GC is defined.
#ifndef __GC
      void _M_free_c_string();
      void _M_free_tree();
      // Deallocate t. Assumes t is not 0.
      void
      _M_unref_nonnil()
      {
   
	if (0 == _M_decr())
	  _M_free_tree();
      }

      void
      _M_ref_nonnil()
      {
    _M_incr(); }

      static void
      _S_unref(_Rope_RopeRep* __t)
      {
   
	if (0 != __t)
	  __t->_M_unref_nonnil();
      }

      static void
      _S_ref(_Rope_RopeRep* __t)
      {
   
	if (0 != __t)
	  __t->_M_incr();
      }
      
      static void
      _S_free_if_unref(_Rope_RopeRep* __t)
      {
   
	if (0 != __t && 0 == __t->_M_ref_count)
	  __t->_M_free_tree();
      }
#   else /* __GC */
      void _M_unref_nonnil() {
    }
      void _M_ref_nonnil() {
    }
      static void _S_unref(_Rope_RopeRep*) {
    }
      static void _S_ref(_Rope_RopeRep*) {
    }
      static void _S_free_if_unref(_Rope_RopeRep*) {
    }
#   endif
protected:
      _Rope_RopeRep&
      operator=(const _Rope_RopeRep&);

      _Rope_RopeRep(const _Rope_RopeRep&);
    };

  template<class _CharT, class _Alloc>
    struct _Rope_RopeLeaf
    : public _Rope_RopeRep<_CharT, _Alloc>
    {
   
    public:
      // Apparently needed by VC++
      // The data fields of leaves are allocated with some
      // extra space, to accommodate future growth and for basic
      // character types, to hold a trailing eos character.
      enum {
    _S_alloc_granularity = 8 };
      
      static size_t
      _S_rounded_up_size(size_t __n)
      {
   
        size_t __size_with_eos;
	
        if (_S_is_basic_char_type((_CharT*)0))
	  __size_with_eos = __n + 1;
	else
	  __size_with_eos = __n;
#ifdef __GC
	return __size_with_eos;
#else
最低0.47元/天 解锁文章
繁凡さん
关注
专栏目录
【COGS】 [HZOI 2015] Persistable Editor-可持久化平衡树
远行客
07-29 320
传送门:cogs2314 题解 非旋treap的模板题。把字符串的信息按输入顺序记在treap上，每个节点存一个字符。插入字符串和输出都直接暴力找（长度保证） 代码 #include&amp;amp;amp;lt;bits/stdc++.h&amp;amp;amp;gt; using namespace std; const int N=1e5+1000; int rt[N],tim,cnt; int n,op,ix,iy,...
ACM入门之【可持续化数组(Rope)】
辉小歌的博客
05-18 452
一个黑科技，有时候可以用来偷鸡。万一过了呢？ rope(可持续化平衡树) 需要的头文件#include <ext/rope> 命名空间声明:using namespace __gnu_cxx; 常用的函数： rope<int> a; a.push_back(x) // 在尾部插入 a.append(x) //在尾部插入 a.pop_back(x) // 删除最后一个元素 a.size() // 返回长度 a.insert(int pos, x) // 在pos插入.
可持久化线段树/平衡树
最新发布
Se_ren_di_pity的博客
07-03 546
对于一棵树来说，我们每次单点修改的时候，无非就是修改了它的左子树或者是右子树，那么对于新的版本来说，当他修改的是左子树的节点的时候，显然我们可以继续沿用历史版本的右子树，也就是只要复制并修改一棵左子树就可以了；这就是可持久化的核心思想。操作和可持久化线段树类似，我这里采用的是FHQ-Treap实现的平衡树，在split的过程中会修改树的值，因此在split的时候需要开新点，由于merge的时候连接的都是split开的新点，所以merge就不用再开新点了。
C++ STL rope介绍----可持久化平衡树
07-14 672
大致介绍： rope这个东西，我刚刚知道这玩意，用的不是很多，做个简单的介绍。 官方说明：我是刘邦（我估计你是看不懂的）。 rope就是一个用可持久化平衡树实现的“重型”string（然而它也可以保存int或其他的类型），它不是标准STL里的东西，属于STL扩展。 crope即rope<char>，就是一个“重型”string，且可以用cin/cout直接输入输出。 ...
【模板】可持久化平衡树
loooooog的博客
02-01 273
【模板】可持久化平衡树 题目背景 本题为题目 普通平衡树 的可持久化加强版。 数据已经经过强化 感谢@Kelin 提供的一组hack数据 题目描述 您需要写一种数据结构（可参考题目标题），来维护一些数，其中需要提供以下操作（对于各个以往的历史版本）： 插入x数 删除x数(若有多个相同的数，因只删除一个，如果没有请忽略该操作) 查询x数...
「模板」 可持久化平衡树
weixin_30362083的博客
05-17 72
「模板」 可持久化平衡树 (There's suddenly something wrong with my Chinese Input Method...) Persistable FHQ_Treap. MLE 2 points after debugging for a whole evening for the pointers. At last I chose to use arra...
c++自带的可持久化平衡树？rope大法好！（超详细解答 + 5道例题讲解，可直接替代可持久化的线段树、并查集、平衡树！）
繁凡さん的博客
11-10 5366
c++自带的可持久化平衡树？rope大法好！ C++官方说明 C++11及以后开始支持 可以当作可持久化平衡树使用，内部构造是一个块状链表 1. 声明 1）头文件 #include<ext/rope> 2)调用命名空间 using namespace __gnu_cxx; 3)声明使用 rope<int>Irp; rope<long long>Lrp; crope crp;//相当于定义成rope<char>，即定义为string类型 rope&lt
Rope大法（可持久化平衡树）
vocaloid01的博客
07-27 1541
2008年OI集训论文上有介绍&lt;对块状链表的一点研究&gt;，其主要是结合了链表和数组各自的优点，链表中的节点指向每个数据 块，即数组，并且记录数据的个数，然后分块查找和插入。 头文件：#include &lt;ext/rope&gt; 命名空间：using namespace __gnu_cxx 基本操作： rope test; test.push_back(x);//在末尾添...
【模板】可持久化数组
夕弦
05-17 921
戳我 可持久化数组是由可持久化线段树或可持久化平衡树实现的。这里先给出可持久化线段树的实现方法。 为了方便起见，处理的数组长度为5， 起始的数组元素为1~5，修改是将第一个位置的数组元素改为2。 建树规则很简单，只要在叶子节点上写上该点的值就可以了。 先根据原数组建一棵线段树 第一次修改之后的线段树 我们发现，这两棵线段树中只有一个叶子节点的值发生了改变，而操作数非常多，假...
平衡树的C++实现
11-04
用C++实现的平衡树的插入和删除操作，查询操作没实现。测试数据量比较大，可以根据需要在insert和remove方法里面修改打印的东西
cpp-rope:C ++中的绳索实现
05-09
绳索 绳索是一种基于二叉树的数据结构，旨在对大量文本进行有效的字符串操作。 这是C ++中的绳索实现。 尽管C ++ STL中已有绳索的实现，但是该实现仅限于使用char或wchar作为基础类型的绳索。 此外，STL实现不支持Unicode，这增加了使用UTF-8字符串的复杂性（例如，按单词/代码点进行迭代）。 此实现同时提供了两者。 实施细节 该绳索未指定基础类型，必须将基础类型作为模板参数提供。 绳索上的随机访问和切片操作受“度量”概念的支持，“度量”与类半身像相似。 几乎所有交互都是通过迭代器完成的，该迭代器是由度量标识符构成的。 措施 “ Measure”类型是一个抽象类，具有以下要求： 它必须可以根据绳索的基础类型的向量进行构造 它必须是可组合的（即，必须有（T，T）-> T类型的函数，例如'+'） 它必须具有“索引”功能，在给定向量和目标度量的情况下，该函数将能够在向
C++STL源码分析 SGI版
10-30
**C++ STL源码分析 SGI版** C++标准模板库（Standard Template Library，STL）是C++编程中不可或缺的一部分，它提供了高效且灵活的数据结构和算法。STL的实现有多种，其中SGI（Stanford Graphics Interface）版本以...
【洛谷P3835】【模板】—可持久化平衡树（可持久化FHQ_Treap）
weixin_30500663的博客
01-24 110
传送门 其实感觉和主席树没什么区别，考虑到TreapTreapTreap期望的树高大约为logloglog 所以每次只会有logloglog被改变 不变的指向儿子就可以了 注意引用地址符 #include<bits/stdc++.h> using namespace std; inline int read(){ char ch=...
可持久化平衡树
weixin_30599769的博客
01-21 290
Kirin:可持久化数据结构可以保留历史版本，用新建代替修改。 洛谷模板题 题目描述 您需要写一种数据结构（可参考题目标题），来维护一些数，其中需要提供以下操作（对于各个以往的历史版本）： 1.插入x数 2.删除x数(若有多个相同的数，因只删除一个，如果没有请忽略该操作) 3.查询x数的排名(排名定义为比当前数小的数的个数+1。若有多个相同的数，因输出最小的排名) 4.查询排名为x的数 5....
可持久化平衡树（fhq Treap）
lcl1234567890lcl的博客
08-02 465
P3835 【模板】可持久化平衡树 其实和一般的fhq Treap差别不是很大，只有在分裂，合并的时候加入可持久化的操作 具体操作如下 void split(int p,int val,int &x,int &y) { if(!p) x=y=0; else { if(e[p].val<=val) { x=++tot; e[x]=e[p]; split(e[x].r,val,e[x].r,y); update(x); } else {
[模板]可持久化平衡树
C202044zxy的博客
11-30 146
一、题目 点此看题 二、解法 其实可持久化的思路很简单，就是你永远不改原版，而是在新建点上修改。 所以在merge\text{merge}merge #include <cstdio> #include <cstdlib> #include <ctime> #define inf 2147483647 const int MAXN = 500005; const...
C++ Effective STL指南：容器选择与迭代器优化
C++ Effective STL笔记是一份针对C++初学者的重要参考资料，旨在帮助理解和优化在C++中使用标准模板库(STL)的实践。这份笔记详细讨论了如何根据具体需求选择合适的容器类型，以便最大化程序性能和代码效率。 1. **...
评论
添加红包

请填写红包祝福语或标题

个

红包个数最小为10个

元

红包金额最低5元

当前余额3.43元 前往充值 >
需支付：10.00元
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包

打赏作者

繁凡さん

你的鼓励将是我创作的最大动力

¥1 ¥2 ¥4 ¥6 ¥10 ¥20
扫码支付：¥1
获取中
扫码支付

您的余额不足，请更换扫码支付或充值

打赏作者

实付元
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明：

1.余额是钱包充值的虚拟货币，按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载，可以购买VIP、付费专栏及课程。

余额充值