stantiate // templates or use macros or inline functions from this file, or you compile // this file and link it with other files to produce an executable, this // file does not by itself cause the resulting executable to be covered by // the GNU General Public License. This exception does not however // invalidate any other reasons why the executable file might be covered by // the GNU General Public License. /** @file atomicity.h * This is an internal header file, included by other library headers. * You should not attempt to use it directly. */ #ifndef _GLIBCXX_ATOMICITY_H #define _GLIBCXX_ATOMICITY_H 1 #include #include #include _GLIBCXX_BEGIN_NAMESPACE(__gnu_cxx) // Functions for portable atomic access. // To abstract locking primitives across all thread policies, use: // __exchange_and_add_dispatch // __atomic_add_dispatch #ifdef _GLIBCXX_ATOMIC_BUILTINS static inline _Atomic_word __exchange_and_add(volatile _Atomic_word* __mem, int __val) { return __sync_fetch_and_add(__mem, __val); } static inline void __atomic_add(volatile _Atomic_word* __mem, int __val) { __sync_fetch_and_add(__mem, __val); } #else _Atomic_word __attribute__ ((__unused__)) __exchange_and_add(volatile _Atomic_word*, int); void __attribute__ ((__unused__)) __atomic_add(volatile _Atomic_word*, int); #endif static inline _Atomic_word __exchange_and_add_single(_Atomic_word* __mem, int __val) { _Atomic_word __result = *__mem; *__mem += __val; return __result; } static inline void __atomic_add_single(_Atomic_word* __mem, int __val) { *__mem += __val; } static inline _Atomic_word __attribute__ ((__unused__)) __exchange_and_add_dispatch(_Atomic_word* __mem, int __val) { #ifdef __GTHREADS if (__gthread_active_p()) return __exchange_and_add(__mem, __val); else return __exchange_and_add_single(__mem, __val); #else return __exchange_and_add_single(__mem, __val); #endif } static inline void __attribute__ ((__unused__)) __atomic_add_dispatch(_Atomic_word* __mem, int __val) { #ifdef __GTHREADS if (__gthread_active_p()) __atomic_add(__mem, __val); else __atomic_add_single(__mem, __val); #else __atomic_add_single(__mem, __val); #endif } _GLIBCXX_END_NAMESPACE // Even if the CPU doesn't need a memory barrier, we need to ensure // that the compiler doesn't reorder memory accesses across the // barriers. #ifndef _GLIBCXX_READ_MEM_BARRIER #define _GLIBCXX_READ_MEM_BARRIER __asm __volatile ("":::"memory") #endif #ifndef _GLIBCXX_WRITE_MEM_BARRIER #define _GLIBCXX_WRITE_MEM_BARRIER __asm __volatile ("":::"memory") #endif #endif // Hashing map implementation -*- C++ -*- // Copyright (C) 2001, 2002, 2004, 2005, 2006 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 2, 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. // You should have received a copy of the GNU General Public License along // with this library; see the file COPYING. If not, write to the Free // Software Foundation, 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, // USA. // As a special exception, you may use this file as part of a free software // library without restriction. Specifically, if other files instantiate // templates or use macros or inline functions from this file, or you compile // this file and link it with other files to produce an executable, this // file does not by itself cause the resulting executable to be covered by // the GNU General Public License. This exception does not however // invalidate any other reasons why the executable file might be covered by // the GNU General Public License. /* * Copyright (c) 1996 * 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. * */ /** @file backward/hash_map * This file is a GNU extension to the Standard C++ Library (possibly * containing extensions from the HP/SGI STL subset). */ #ifndef _HASH_MAP #define _HASH_MAP 1 #include "backward_warning.h" #include #include #include _GLIBCXX_BEGIN_NAMESPACE(__gnu_cxx) using std::equal_to; using std::allocator; using std::pair; using std::_Select1st; /** * This is an SGI extension. * @ingroup SGIextensions * @doctodo */ template, class _EqualKey = equal_to<_Key>, class _Alloc = allocator<_Tp> > class hash_map { private: typedef hashtable,_Key, _HashFn, _Select1st >, _EqualKey, _Alloc> _Ht; _Ht _M_ht; public: typedef typename _Ht::key_type key_type; typedef _Tp data_type; typedef _Tp mapped_type; typedef typename _Ht::value_type value_type; typedef typename _Ht::hasher hasher; typedef typename _Ht::key_equal key_equal; typedef typename _Ht::size_type size_type; typedef typename _Ht::difference_type difference_type; typedef typename _Ht::pointer pointer; typedef typename _Ht::const_pointer const_pointer; typedef typename _Ht::reference reference; typedef typename _Ht::const_reference const_reference; typedef typename _Ht::iterator iterator; typedef typename _Ht::const_iterator const_iterator; typedef typename _Ht::allocator_type allocator_type; hasher hash_funct() const { return _M_ht.hash_funct(); } key_equal key_eq() const { return _M_ht.key_eq(); } allocator_type get_allocator() const { return _M_ht.get_allocator(); } hash_map() : _M_ht(100, hasher(), key_equal(), allocator_type()) {} explicit hash_map(size_type __n) : _M_ht(__n, hasher(), key_equal(), allocator_type()) {} hash_map(size_type __n, const hasher& __hf) : _M_ht(__n, __hf, key_equal(), allocator_type()) {} hash_map(size_type __n, const hasher& __hf, const key_equal& __eql, const allocator_type& __a = allocator_type()) : _M_ht(__n, __hf, __eql, __a) {} template hash_map(_InputIterator __f, _InputIterator __l) : _M_ht(100, hasher(), key_equal(), allocator_type()) { _M_ht.insert_unique(__f, __l); } template hash_map(_InputIterator __f, _InputIterator __l, size_type __n) : _M_ht(__n, hasher(), key_equal(), allocator_type()) { _M_ht.insert_unique(__f, __l); } template hash_map(_InputIterator __f, _InputIterator __l, size_type __n, const hasher& __hf) : _M_ht(__n, __hf, key_equal(), allocator_type()) { _M_ht.insert_unique(__f, __l); } template hash_map(_InputIterator __f, _InputIterator __l, size_type __n, const hasher& __hf, const key_equal& __eql, const allocator_type& __a = allocator_type()) : _M_ht(__n, __hf, __eql, __a) { _M_ht.insert_unique(__f, __l); } size_type size() const { return _M_ht.size(); } size_type max_size() const { return _M_ht.max_size(); } bool empty() const { return _M_ht.empty(); } void swap(hash_map& __hs) { _M_ht.swap(__hs._M_ht); } template friend bool operator== (const hash_map<_K1, _T1, _HF, _EqK, _Al>&, const hash_map<_K1, _T1, _HF, _EqK, _Al>&); iterator begin() { return _M_ht.begin(); } iterator end() { return _M_ht.end(); } const_iterator begin() const { return _M_ht.begin(); } const_iterator end() const { return _M_ht.end(); } pair insert(const value_type& __obj) { return _M_ht.insert_unique(__obj); } template void insert(_InputIterator __f, _InputIterator __l) { _M_ht.insert_unique(__f, __l); } pair insert_noresize(const value_type& __obj) { return _M_ht.insert_unique_noresize(__obj); } iterator find(const key_type& __key) { return _M_ht.find(__key); } const_iterator find(const key_type& __key) const { return _M_ht.find(__key); } _Tp& operator[](const key_type& __key) { return _M_ht.find_or_insert(value_type(__key, _Tp())).second; } size_type count(const key_type& __key) const { return _M_ht.count(__key); } pair equal_range(const key_type& __key) { return _M_ht.equal_range(__key); } pair equal_range(const key_type& __key) const { return _M_ht.equal_range(__key); } size_type erase(const key_type& __key) {return _M_ht.erase(__key); } void erase(iterator __it) { _M_ht.erase(__it); } void erase(iterator __f, iterator __l) { _M_ht.erase(__f, __l); } void clear() { _M_ht.clear(); } void resize(size_type __hint) { _M_ht.resize(__hint); } size_type bucket_count() const { return _M_ht.bucket_count(); } size_type max_bucket_count() const { return _M_ht.max_bucket_count(); } size_type elems_in_bucket(size_type __n) const { return _M_ht.elems_in_bucket(__n); } }; template inline bool operator==(const hash_map<_Key, _Tp, _HashFn, _EqlKey, _Alloc>& __hm1, const hash_map<_Key, _Tp, _HashFn, _EqlKey, _Alloc>& __hm2) { return __hm1._M_ht == __hm2._M_ht; } template inline bool operator!=(const hash_map<_Key, _Tp, _HashFn, _EqlKey, _Alloc>& __hm1, const hash_map<_Key, _Tp, _HashFn, _EqlKey, _Alloc>& __hm2) { return !(__hm1 == __hm2); } template inline void swap(hash_map<_Key, _Tp, _HashFn, _EqlKey, _Alloc>& __hm1, hash_map<_Key, _Tp, _HashFn, _EqlKey, _Alloc>& __hm2) { __hm1.swap(__hm2); } /** * This is an SGI extension. * @ingroup SGIextensions * @doctodo */ template, class _EqualKey = equal_to<_Key>, class _Alloc = allocator<_Tp> > class hash_multimap { // concept requirements __glibcxx_class_requires(_Key, _SGIAssignableConcept) __glibcxx_class_requires(_Tp, _SGIAssignableConcept) __glibcxx_class_requires3(_HashFn, size_t, _Key, _UnaryFunctionConcept) __glibcxx_class_requires3(_EqualKey, _Key, _Key, _BinaryPredicateConcept) private: typedef hashtable, _Key, _HashFn, _Select1st >, _EqualKey, _Alloc> _Ht; _Ht _M_ht; public: typedef typename _Ht::key_type key_type; typedef _Tp data_type; typedef _Tp mapped_type; typedef typename _Ht::value_type value_type; typedef typename _Ht::hasher hasher; typedef typename _Ht::key_equal key_equal; typedef typename _Ht::size_type size_type; typedef typename _Ht::difference_type difference_type; typedef typename _Ht::pointer pointer; typedef typename _Ht::const_pointer const_pointer; typedef typename _Ht::reference reference; typedef typename _Ht::const_reference const_reference; typedef typename _Ht::iterator iterator; typedef typename _Ht::const_iterator const_iterator; typedef typename _Ht::allocator_type allocator_type; hasher hash_funct() const { return _M_ht.hash_funct(); } key_equal key_eq() const { return _M_ht.key_eq(); } allocator_type get_allocator() const { return _M_ht.get_allocator(); } hash_multimap() : _M_ht(100, hasher(), key_equal(), allocator_type()) {} explicit hash_multimap(size_type __n) : _M_ht(__n, hasher(), key_equal(), allocator_type()) {} hash_multimap(size_type __n, const hasher& __hf) : _M_ht(__n, __hf, key_equal(), allocator_type()) {} hash_multimap(size_type __n, const hasher& __hf, const key_equal& __eql, const allocator_type& __a = allocator_type()) : _M_ht(__n, __hf, __eql, __a) {} template hash_multimap(_InputIterator __f, _InputIterator __l) : _M_ht(100, hasher(), key_equal(), allocator_type()) { _M_ht.insert_equal(__f, __l); } template hash_multimap(_InputIterator __f, _InputIterator __l, size_type __n) : _M_ht(__n, hasher(), key_equal(), allocator_type()) { _M_ht.insert_equal(__f, __l); } template hash_multimap(_InputIterator __f, _InputIterator __l, size_type __n, const hasher& __hf) : _M_ht(__n, __hf, key_equal(), allocator_type()) { _M_ht.insert_equal(__f, __l); } template hash_multimap(_InputIterator __f, _InputIterator __l, size_type __n, const hasher& __hf, const key_equal& __eql, const allocator_type& __a = allocator_type()) : _M_ht(__n, __hf, __eql, __a) { _M_ht.insert_equal(__f, __l); } size_type size() const { return _M_ht.size(); 777777} size_type max_size() const { return _M_ht.max_size(); } bool empty() const { return _M_ht.empty(); } void swap(hash_multimap& __hs) { _M_ht.swap(__hs._M_ht); } template friend bool operator==(const hash_multimap<_K1, _T1, _HF, _EqK, _Al>&, const hash_multimap<_K1, _T1, _HF, _EqK, _Al>&); iterator begin() { return _M_ht.begin(); } iterator end() { return _M_ht.end(); } const_iterator begin() const { return _M_ht.begin(); } const_iterator end() const { return _M_ht.end(); } iterator insert(const value_type& __obj) { return _M_ht.insert_equal(__obj); } template void insert(_InputIterator __f, _InputIterator __l) { _M_ht.insert_equal(__f,__l); } iterator insert_noresize(const value_type& __obj) { return _M_ht.insert_equal_noresize(__obj); } iterator find(const key_type& __key) { return _M_ht.find(__key); } const_iterator find(const key_type& __key) const { return _M_ht.find(__key); } size_type count(const key_type& __key) const { return _M_ht.count(__key); } pair equal_range(const key_type& __key) { return _M_ht.equal_range(__key); } pair equal_range(const key_type& __key) const { return _M_ht.equal_range(__key); } size_type erase(const key_type& __key) { return _M_ht.erase(__key); } void erase(iterator __it) { _M_ht.erase(__it); } void erase(iterator __f, iterator __l) { _M_ht.erase(__f, __l); } void clear() { _M_ht.clear(); } void resize(size_type __hint) { _M_ht.resize(__hint); } size_type bucket_count() const { return _M_ht.bucket_count(); } size_type max_bucket_count() const { return _M_ht.max_bucket_count(); } size_type elems_in_bucket(size_type __n) const { return _M_ht.elems_in_bucket(__n); } }; template inline bool operator==(const hash_multimap<_Key, _Tp, _HF, _EqKey, _Alloc>& __hm1, const hash_multimap<_Key, _Tp, _HF, _EqKey, _Alloc>& __hm2) { return __hm1._M_ht == __hm2._M_ht; } template inline bool operator!=(const hash_multimap<_Key, _Tp, _HF, _EqKey, _Alloc>& __hm1, const hash_multimap<_Key, _Tp, _HF, _EqKey, _Alloc>& __hm2) { return !(__hm1 == __hm2); } template inline void swap(hash_multimap<_Key, _Tp, _HashFn, _EqlKey, _Alloc>& __hm1, hash_multimap<_Key, _Tp, _HashFn, _EqlKey, _Alloc>& __hm2) { __hm1.swap(__hm2); } _GLIBCXX_END_NAMESPACE _GLIBCXX_BEGIN_NAMESPACE(std) // Specialization of insert_iterator so that it will work for hash_map // and hash_multimap. template class insert_iterator<__gnu_cxx::hash_map<_Key, _Tp, _HashFn, _EqKey, _Alloc> > { protected: typedef __gnu_cxx::hash_map<_Key, _Tp, _HashFn, _EqKey, _Alloc> _Container; _Container* container; public: typedef _Container container_type; typedef output_iterator_tag iterator_category; typedef void value_type; typedef void difference_type; typedef void pointer; typedef void reference; insert_iterator(_Container& __x) : container(&__x) {} insert_iterator(_Container& __x, typename _Container::iterator) : container(&__x) {} insert_iterator<_Container>& operator=(const typename _Container::value_type& __value) { container->insert(__value); return *this; } insert_iterator<_Container>& operator*() { return *this; } insert_iterator<_Container>& operator++() { return *this; } insert_iterator<_Container>& operator++(int) { return *this; } }; template class insert_iterator<__gnu_cxx::hash_multimap<_Key, _Tp, _HashFn, _EqKey, _Alloc> > { protected: typedef __gnu_cxx::hash_multimap<_Key, _Tp, _HashFn, _EqKey, _Alloc> _Container; _Container* container; typename _Container::iterator iter; public: typedef _Container container_type; typedef output_iterator_tag iterator_category; typedef void value_type; typedef void difference_type; typedef void pointer; typedef void reference; insert_iterator(_Container& __x) : container(&__x) {} insert_iterator(_Container& __x, typename _Container::iterator) : container(&__x) {} insert_iterator<_Container>& operator=(const typename _Container::value_type& __value) { container->insert(__value); return *this; } insert_iterator<_Container>& operator*() { return *this; } insert_iterator<_Container>& operator++() { return *this; } insert_iterator<_Container>& operator++(int) { return *this; } }; _GLIBCXX_END_NAMESPACE #endif // SGI's rope class implementation -*- C++ -*- // Copyright (C) 2001, 2002, 2003, 2004, 2005, 2006, 2007 // 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 2, 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. // You should have received a copy of the GNU General Public License along // with this library; see the file COPYING. If not, write to the Free // Software Foundation, 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, // USA. // As a special exception, you may use this file as part of a free software // library without restriction. Specifically, if other files instantiate // templates or use macros or inline functions from this file, or you compile // this file and link it with other files to produce an executable, this // file does not by itself cause the resulting executable to be covered by // the GNU General Public License. This exception does not however // invalidate any other reasons why the executable file might be covered by // the GNU General Public License. /* * 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 ropeimpl.h * This is an internal header file, included by other library headers. * You should not attempt to use it directly. */ #include #include #include #include // For copy_n and lexicographical_compare_3way #include // For uninitialized_copy_n #include // For power _GLIBCXX_BEGIN_NAMESPACE(__gnu_cxx) using std::size_t; using std::printf; using std::basic_ostream; using std::__throw_length_error; using std::_Destroy; using std::uninitialized_fill_n; // Set buf_start, buf_end, and buf_ptr appropriately, filling tmp_buf // if necessary. Assumes _M_path_end[leaf_index] and leaf_pos are correct. // Results in a valid buf_ptr if the iterator can be legitimately // dereferenced. template void _Rope_iterator_base<_CharT, _Alloc>:: _S_setbuf(_Rope_iterator_base<_CharT, _Alloc>& __x) { const _RopeRep* __leaf = __x._M_path_end[__x._M_leaf_index]; size_t __leaf_pos = __x._M_leaf_pos; size_t __pos = __x._M_current_pos; switch(__leaf->_M_tag) { case __detail::_S_leaf: __x._M_buf_start = ((_Rope_RopeLeaf<_CharT, _Alloc>*)__leaf)->_M_data; __x._M_buf_ptr = __x._M_buf_start + (__pos - __leaf_pos); __x._M_buf_end = __x._M_buf_start + __leaf->_M_size; break; case __detail::_S_function: case __detail::_S_substringfn: { size_t __len = _S_iterator_buf_len; size_t __buf_start_pos = __leaf_pos; size_t __leaf_end = __leaf_pos + __leaf->_M_size; char_producer<_CharT>* __fn = ((_Rope_RopeFunction<_CharT, _Alloc>*)__leaf)->_M_fn; if (__buf_start_pos + __len <= __pos) { __buf_start_pos = __pos - __len / 4; if (__buf_start_pos + __len > __leaf_end) __buf_start_pos = __leaf_end - __len; } if (__buf_start_pos + __len > __leaf_end) __len = __leaf_end - __buf_start_pos; (*__fn)(__buf_start_pos - __leaf_pos, __len, __x._M_tmp_buf); __x._M_buf_ptr = __x._M_tmp_buf + (__pos - __buf_start_pos); __x._M_buf_start = __x._M_tmp_buf; __x._M_buf_end = __x._M_tmp_buf + __len; } break; default: break; } } // Set path and buffer inside a rope iterator. We assume that // pos and root are already set. template void _Rope_iterator_base<_CharT, _Alloc>:: _S_setcache(_Rope_iterator_base<_CharT, _Alloc>& __x) { const _RopeRep* __path[int(__detail::_S_max_rope_depth) + 1]; const _RopeRep* __curr_rope; int __curr_depth = -1; /* index into path */ size_t __curr_start_pos = 0; size_t __pos = __x._M_current_pos; unsigned char __dirns = 0; // Bit vector marking right turns in the path if (__pos >= __x._M_root->_M_size) { __x._M_buf_ptr = 0; return; } __curr_rope = __x._M_root; if (0 != __curr_rope->_M_c_string) { /* Treat the root as a leaf. */ __x._M_buf_start = __curr_rope->_M_c_string; __x._M_buf_end = __curr_rope->_M_c_string + __curr_rope->_M_size; __x._M_buf_ptr = __curr_rope->_M_c_string + __pos; __x._M_path_end[0] = __curr_rope; __x._M_leaf_index = 0; __x._M_leaf_pos = 0; return; } for(;;) { ++__curr_depth; __path[__curr_depth] = __curr_rope; switch(__curr_rope->_M_tag) { case __detail::_S_leaf: case __detail::_S_function: case __detail::_S_substringfn: __x._M_leaf_pos = __curr_start_pos; goto done; case __detail::_S_concat: { _Rope_RopeConcatenation<_CharT, _Alloc>* __c = (_Rope_RopeConcatenation<_CharT, _Alloc>*)__curr_rope; _RopeRep* __left = __c->_M_left; size_t __left_len = __left->_M_size; __dirns <<= 1; if (__pos >= __curr_start_pos + __left_len) { __dirns |= 1; __curr_rope = __c->_M_right; __curr_start_pos += __left_len; } else __curr_rope = __left; } break; } } done: // Copy last section of path into _M_path_end. { int __i = -1; int __j = __curr_depth + 1 - int(_S_path_cache_len); if (__j < 0) __j = 0; while (__j <= __curr_depth) __x._M_path_end[++__i] = __path[__j++]; __x._M_leaf_index = __i; } __x._M_path_directions = __dirns; _S_setbuf(__x); } // Specialized version of the above. Assumes that // the path cache is valid for the previous position. template void _Rope_iterator_base<_CharT, _Alloc>:: _S_setcache_for_incr(_Rope_iterator_base<_CharT, _Alloc>& __x) { int __current_index = __x._M_leaf_index; const _RopeRep* __current_node = __x._M_path_end[__current_index]; size_t __len = __current_node->_M_size; size_t __node_start_pos = __x._M_leaf_pos; unsigned char __dirns = __x._M_path_directions; _Rope_RopeConcatenation<_CharT, _Alloc>* __c; if (__x._M_current_pos - __node_start_pos < __len) { /* More stuff in this leaf, we just didn't cache it. */ _S_setbuf(__x); return; } // node_start_pos is starting position of last_node. while (--__current_index >= 0) { if (!(__dirns & 1) /* Path turned left */) break; __current_node = __x._M_path_end[__current_index]; __c = (_Rope_RopeConcatenation<_CharT, _Alloc>*)__current_node; // Otherwise we were in the right child. Thus we should pop // the concatenation node. __node_start_pos -= __c->_M_left->_M_size; __dirns >>= 1; } if (__current_index < 0) { // We underflowed the cache. Punt. _S_setcache(__x); return; } __current_node = __x._M_path_end[__current_index]; __c = (_Rope_RopeConcatenation<_CharT, _Alloc>*)__current_node; // current_node is a concatenation node. We are positioned on the first // character in its right child. // node_start_pos is starting position of current_node. __node_start_pos += __c->_M_left->_M_size; __current_node = __c->_M_right; __x._M_path_end[++__current_index] = __current_node; __dirns |= 1; while (__detail::_S_concat == __current_node->_M_tag) { ++__current_index; if (int(_S_path_cache_len) == __current_index) { int __i; for (__i = 0; __i < int(_S_path_cache_len) - 1; __i++) __x._M_path_end[__i] = __x._M_path_end[__i+1]; --__current_index; } __current_node = ((_Rope_RopeConcatenation<_CharT, _Alloc>*)__current_node)->_M_left; __x._M_path_end[__current_index] = __current_node; __dirns <<= 1; // node_start_pos is unchanged. } __x._M_leaf_index = __current_index; __x._M_leaf_pos = __node_start_pos; __x._M_path_directions = __dirns; _S_setbuf(__x); } template void _Rope_iterator_base<_CharT, _Alloc>:: _M_incr(size_t __n) { _M_current_pos += __n; if (0 != _M_buf_ptr) { size_t __chars_left = _M_buf_end - _M_buf_ptr; if (__chars_left > __n) _M_buf_ptr += __n; else if (__chars_left == __n) { _M_buf_ptr += __n; _S_setcache_for_incr(*this); } else _M_buf_ptr = 0; } } template void _Rope_iterator_base<_CharT, _Alloc>:: _M_decr(size_t __n) { if (0 != _M_buf_ptr) { size_t __chars_left = _M_buf_ptr - _M_buf_start; if (__chars_left >= __n) _M_buf_ptr -= __n; else _M_buf_ptr = 0; } _M_current_pos -= __n; } template void _Rope_iterator<_CharT, _Alloc>:: _M_check() { if (_M_root_rope->_M_tree_ptr != this->_M_root) { // _Rope was modified. Get things fixed up. _RopeRep::_S_unref(this->_M_root); this->_M_root = _M_root_rope->_M_tree_ptr; _RopeRep::_S_ref(this->_M_root); this->_M_buf_ptr = 0; } } template inline _Rope_const_iterator<_CharT, _Alloc>:: _Rope_const_iterator(const _Rope_iterator<_CharT, _Alloc>& __x) : _Rope_iterator_base<_CharT, _Alloc>(__x) { } template inline _Rope_iterator<_CharT, _Alloc>:: _Rope_iterator(rope<_CharT, _Alloc>& __r, size_t __pos) : _Rope_iterator_base<_CharT,_Alloc>(__r._M_tree_ptr, __pos), _M_root_rope(&__r) { _RopeRep::_S_ref(this->_M_root); } template inline size_t rope<_CharT, _Alloc>:: _S_char_ptr_len(const _CharT* __s) { const _CharT* __p = __s; while (!_S_is0(*__p)) ++__p; return (__p - __s); } #ifndef __GC template inline void _Rope_RopeRep<_CharT, _Alloc>:: _M_free_c_string() { _CharT* __cstr = _M_c_string; if (0 != __cstr) { size_t __size = this->_M_size + 1; _Destroy(__cstr, __cstr + __size, _M_get_allocator()); this->_Data_deallocate(__cstr, __size); } } template inline void _Rope_RopeRep<_CharT, _Alloc>:: _S_free_string(_CharT* __s, size_t __n, allocator_type& __a) { if (!_S_is_basic_char_type((_CharT*)0)) _Destroy(__s, __s + __n, __a); // This has to be a static member, so this gets a bit messy __a.deallocate(__s, _Rope_RopeLeaf<_CharT, _Alloc>::_S_rounded_up_size(__n)); } // There are several reasons for not doing this with virtual destructors // and a class specific delete operator: // - A class specific delete operator can't easily get access to // allocator instances if we need them. // - Any virtual function would need a 4 or byte vtable pointer; // this only requires a one byte tag per object. template void _Rope_RopeRep<_CharT, _Alloc>:: _M_free_tree() { switch(_M_tag) { case __detail::_S_leaf: { _Rope_RopeLeaf<_CharT, _Alloc>* __l = (_Rope_RopeLeaf<_CharT, _Alloc>*)this; __l->_Rope_RopeLeaf<_CharT, _Alloc>::~_Rope_RopeLeaf(); _L_deallocate(__l, 1); break; } case __detail::_S_concat: { _Rope_RopeConcatenation<_CharT,_Alloc>* __c = (_Rope_RopeConcatenation<_CharT, _Alloc>*)this; __c->_Rope_RopeConcatenation<_CharT, _Alloc>:: 777777777777777777777777777777777777 ~_Rope_RopeConcatenation(); _C_deallocate(__c, 1); break; } case __detail::_S_function: { _Rope_RopeFunction<_CharT, _Alloc>* __f = (_Rope_RopeFunction<_CharT, _Alloc>*)this; __f->_Rope_RopeFunction<_CharT, _Alloc>::~_Rope_RopeFunction(); _F_deallocate(__f, 1); break; } case __detail::_S_substringfn: { _Rope_RopeSubstring<_CharT, _Alloc>* __ss = (_Rope_RopeSubstring<_CharT, _Alloc>*)this; __ss->_Rope_RopeSubstring<_CharT, _Alloc>:: ~_Rope_RopeSubstring(); _S_deallocate(__ss, 1); break; } } } #else template inline void _Rope_RopeRep<_CharT, _Alloc>:: _S_free_string(const _CharT*, size_t, allocator_type) { } #endif // Concatenate a C string onto a leaf rope by copying the rope data. // Used for short ropes. template typename rope<_CharT, _Alloc>::_RopeLeaf* rope<_CharT, _Alloc>:: _S_leaf_concat_char_iter(_RopeLeaf* __r, const _CharT* __iter, size_t __len) { size_t __old_len = __r->_M_size; _CharT* __new_data = (_CharT*) _Data_allocate(_S_rounded_up_size(__old_len + __len)); _RopeLeaf* __result; uninitialized_copy_n(__r->_M_data, __old_len, __new_data); uninitialized_copy_n(__iter, __len, __new_data + __old_len); _S_cond_store_eos(__new_data[__old_len + __len]); try { __result = _S_new_RopeLeaf(__new_data, __old_len + __len, __r->_M_get_allocator()); } catch(...) { _RopeRep::__STL_FREE_STRING(__new_data, __old_len + __len, __r->_M_get_allocator()); __throw_exception_again; } return __result; } #ifndef __GC // As above, but it's OK to clobber original if refcount is 1 template typename rope<_CharT,_Alloc>::_RopeLeaf* rope<_CharT, _Alloc>:: _S_destr_leaf_concat_char_iter(_RopeLeaf* __r, const _CharT* __iter, size_t __len) { if (__r->_M_ref_count > 1) return _S_leaf_concat_char_iter(__r, __iter, __len); size_t __old_len = __r->_M_size; if (_S_allocated_capacity(__old_len) >= __old_len + __len) { // The space has been partially initialized for the standard // character types. But that doesn't matter for those types. uninitialized_copy_n(__iter, __len, __r->_M_data + __old_len); if (_S_is_basic_char_type((_CharT*)0)) _S_cond_store_eos(__r->_M_data[__old_len + __len]); else if (__r->_M_c_string != __r->_M_data && 0 != __r->_M_c_string) { __r->_M_free_c_string(); __r->_M_c_string = 0; } __r->_M_size = __old_len + __len; __r->_M_ref_count = 2; return __r; } else { _RopeLeaf* __result = _S_leaf_concat_char_iter(__r, __iter, __len); return __result; } } #endif // Assumes left and right are not 0. // Does not increment (nor decrement on exception) child reference counts. // Result has ref count 1. template typename rope<_CharT, _Alloc>::_RopeRep* rope<_CharT, _Alloc>:: _S_tree_concat(_RopeRep* __left, _RopeRep* __right) { _RopeConcatenation* __result = _S_new_RopeConcatenation(__left, __right, __left-> _M_get_allocator()); size_t __depth = __result->_M_depth; if (__depth > 20 && (__result->_M_size < 1000 || __depth > size_t(__detail::_S_max_rope_depth))) { _RopeRep* __balanced; try { __balanced = _S_balance(__result); __result->_M_unref_nonnil(); } catch(...) { _C_deallocate(__result,1); __throw_exception_again; } // In case of exception, we need to deallocate // otherwise dangling result node. But caller // still owns its children. Thus unref is // inappropriate. return __balanced; } else return __result; } template typename rope<_CharT, _Alloc>::_RopeRep* rope<_CharT, _Alloc>:: _S_concat_char_iter(_RopeRep* __r, const _CharT*__s, size_t __slen) { _RopeRep* __result; if (0 == __slen) { _S_ref(__r); return __r; } if (0 == __r) return __STL_ROPE_FROM_UNOWNED_CHAR_PTR(__s, __slen, __r->_M_get_allocator()); if (__r->_M_tag == __detail::_S_leaf && __r->_M_size + __slen <= size_t(_S_copy_max)) { __result = _S_leaf_concat_char_iter((_RopeLeaf*)__r, __s, __slen); return __result; } if (__detail::_S_concat == __r->_M_tag && __detail::_S_leaf == ((_RopeConcatenation*) __r)->_M_right->_M_tag) { _RopeLeaf* __right = (_RopeLeaf* )(((_RopeConcatenation* )__r)->_M_right); if (__right->_M_size + __slen <= size_t(_S_copy_max)) { _RopeRep* __left = ((_RopeConcatenation*)__r)->_M_left; _RopeRep* __nright = _S_leaf_concat_char_iter((_RopeLeaf*)__right, __s, __slen); __left->_M_ref_nonnil(); try { __result = _S_tree_concat(__left, __nright); } catch(...) { _S_unref(__left); _S_unref(__nright); __throw_exception_again; } return __result; } } _RopeRep* __nright = __STL_ROPE_FROM_UNOWNED_CHAR_PTR(__s, __slen, __r->_M_get_allocator()); try { __r->_M_ref_nonnil(); __result = _S_tree_concat(__r, __nright); } catch(...) { _S_unref(__r); _S_unref(__nright); __throw_exception_again; } return __result; } #ifndef __GC template typename rope<_CharT,_Alloc>::_RopeRep* rope<_CharT,_Alloc>:: _S_destr_concat_char_iter(_RopeRep* __r, const _CharT* __s, size_t __slen) { _RopeRep* __result; if (0 == __r) return __STL_ROPE_FROM_UNOWNED_CHAR_PTR(__s, __slen, __r->_M_get_allocator()); size_t __count = __r->_M_ref_count; size_t __orig_size = __r->_M_size; if (__count > 1) return _S_concat_char_iter(__r, __s, __slen); if (0 == __slen) { __r->_M_ref_count = 2; // One more than before return __r; } if (__orig_size + __slen <= size_t(_S_copy_max) && __detail::_S_leaf == __r->_M_tag) { __result = _S_destr_leaf_concat_char_iter((_RopeLeaf*)__r, __s, __slen); return __result; } if (__detail::_S_concat == __r->_M_tag) { _RopeLeaf* __right = (_RopeLeaf*)(((_RopeConcatenation*) __r)->_M_right); if (__detail::_S_leaf == __right->_M_tag && __right->_M_size + __slen <= size_t(_S_copy_max)) { _RopeRep* __new_right = _S_destr_leaf_concat_char_iter(__right, __s, __slen); if (__right == __new_right) __new_right->_M_ref_count = 1; else __right->_M_unref_nonnil(); __r->_M_ref_count = 2; // One more than before. ((_RopeConcatenation*)__r)->_M_right = __new_right; __r->_M_size = __orig_size + __slen; if (0 != __r->_M_c_string) { __r->_M_free_c_string(); __r->_M_c_string = 0; } return __r; } } _RopeRep* __right = __STL_ROPE_FROM_UNOWNED_CHAR_PTR(__s, __slen, __r->_M_get_allocator()); __r->_M_ref_nonnil(); try { __result = _S_tree_concat(__r, __right); } catch(...) { _S_unref(__r); _S_unref(__right); __throw_exception_again; } return __result; } #endif /* !__GC */ template typename rope<_CharT, _Alloc>::_RopeRep* rope<_CharT, _Alloc>:: _S_concat(_RopeRep* __left, _RopeRep* __right) { if (0 == __left) { _S_ref(__right); return __right; } if (0 == __right) { __left->_M_ref_nonnil(); return __left; } if (__detail::_S_leaf == __right->_M_tag) { if (__detail::_S_leaf == __left->_M_tag) { if (__right->_M_size + __left->_M_size <= size_t(_S_copy_max)) return _S_leaf_concat_char_iter((_RopeLeaf*)__left, ((_RopeLeaf*)__right)->_M_data, __right->_M_size); } else if (__detail::_S_concat == __left->_M_tag && __detail::_S_leaf == ((_RopeConcatenation*) __left)->_M_right->_M_tag) { _RopeLeaf* __leftright = (_RopeLeaf*)(((_RopeConcatenation*)__left)->_M_right); if (__leftright->_M_size + __right->_M_size <= size_t(_S_copy_max)) { _RopeRep* __leftleft = ((_RopeConcatenation*)__left)->_M_left; _RopeRep* __rest = _S_leaf_concat_char_iter(__leftright, ((_RopeLeaf*) __right)-> _M_data, __right->_M_size); __leftleft->_M_ref_nonnil(); try { return(_S_tree_concat(__leftleft, __rest)); } catch(...) { _S_unref(__leftleft); _S_unref(__rest); __throw_exception_again; } } } } __left->_M_ref_nonnil(); __right->_M_ref_nonnil(); try { return(_S_tree_concat(__left, __right)); } catch(...) { _S_unref(__left); _S_unref(__right); __throw_exception_again; } } template typename rope<_CharT, _Alloc>::_RopeRep* rope<_CharT, _Alloc>:: _S_substring(_RopeRep* __base, size_t __start, size_t __endp1) { if (0 == __base) return 0; size_t __len = __base->_M_size; size_t __adj_endp1; const size_t __lazy_threshold = 128; if (__endp1 >= __len) { if (0 == __start) { __base->_M_ref_nonnil(); return __base; } else __adj_endp1 = __len; } else __adj_endp1 = __endp1; switch(__base->_M_tag) { case __detail::_S_concat: { _RopeConcatenation* __c = (_RopeConcatenation*)__base; _RopeRep* __left = __c->_M_left; _RopeRep* __right = __c->_M_right; size_t __left_len = __left->_M_size; _RopeRep* __result; if (__adj_endp1 <= __left_len) return _S_substring(__left, __start, __endp1); else if (__start >= __left_len) return _S_substring(__right, __start - __left_len, __adj_endp1 - __left_len); _Self_destruct_ptr __left_result(_S_substring(__left, __start, __left_len)); _Self_destruct_ptr __right_result(_S_substring(__right, 0, __endp1 - __left_len)); __result = _S_concat(__left_result, __right_result); return __result; } case __detail::_S_leaf: { _RopeLeaf* __l = (_RopeLeaf*)__base; _RopeLeaf* __result; size_t __result_len; if (__start >= __adj_endp1) return 0; __result_len = __adj_endp1 - __start; if (__result_len > __lazy_threshold) goto lazy; #ifdef __GC const _CharT* __section = __l->_M_data + __start; __result = _S_new_RopeLeaf(__section, __result_len, __base->_M_get_allocator()); __result->_M_c_string = 0; // Not eos terminated. #else // We should sometimes create substring node instead. __result = __STL_ROPE_FROM_UNOWNED_CHAR_PTR(__l->_M_data + __start, __result_len, __base-> _M_get_allocator()); #endif return __result; } case __detail::_S_substringfn: // Avoid introducing multiple layers of substring nodes. { _RopeSubstring* __old = (_RopeSubstring*)__base; size_t __result_len; if (__start >= __adj_endp1) return 0; __result_len = __adj_endp1 - __start; if (__result_len > __lazy_threshold) { _RopeSubstring* __result = _S_new_RopeSubstring(__old->_M_base, __start + __old->_M_start, __adj_endp1 - __start, __base->_M_get_allocator()); return __result; } // *** else fall through: *** } case __detail::_S_function: { _RopeFunction* __f = (_RopeFunction*)__base; _CharT* __section; size_t __result_len; if (__start >= __adj_endp1) return 0; __result_len = __adj_endp1 - __start; if (__result_len > __lazy_threshold) goto lazy; __section = (_CharT*) _Data_allocate(_S_rounded_up_size(__result_len)); try { (*(__f->_M_fn))(__start, __result_len, __section); } catch(...) { _RopeRep::__STL_FREE_STRING(__section, __result_len, __base->_M_get_allocator()); __throw_exception_again; } _S_cond_store_eos(__section[__result_len]); return _S_new_RopeLeaf(__section, __result_len, __base->_M_get_allocator()); } } lazy: { // Create substring node. return _S_new_RopeSubstring(__base, __start, __adj_endp1 - __start, __base->_M_get_allocator()); } } template class _Rope_flatten_char_consumer : public _Rope_char_consumer<_CharT> { private: _CharT* _M_buf_ptr; public: _Rope_flatten_char_consumer(_CharT* __buffer) { _M_buf_ptr = __buffer; }; ~_Rope_flatten_char_consumer() {} bool operator()(const _CharT* __leaf, size_t __n) { uninitialized_copy_n(__leaf, __n, _M_buf_ptr); _M_buf_ptr += __n; return true; } }; template class _Rope_find_char_char_consumer : public _Rope_char_consumer<_CharT> { private: _CharT _M_pattern; public: size_t _M_count; // Number of nonmatching characters _Rope_find_char_char_consumer(_CharT __p) : _M_pattern(__p), _M_count(0) {} ~_Rope_find_char_char_consumer() {} bool operator()(const _CharT* __leaf, size_t __n) { size_t __i; for (__i = 0; __i < __n; __i++) { if (__leaf[__i] == _M_pattern) { _M_count += __i; return false; } } _M_count += __n; return true; } }; template // Here _CharT is both the stream and rope character type. class _Rope_insert_char_consumer : public _Rope_char_consumer<_CharT> { private: typedef basic_ostream<_CharT,_Traits> _Insert_ostream; _Insert_ostream& _M_o; public: _Rope_insert_char_consumer(_Insert_ostream& __writer) : _M_o(__writer) {}; ~_Rope_insert_char_consumer() { }; // Caller is presumed to own the ostream bool operator() (const _CharT* __leaf, size_t __n); // Returns true to continue traversal. }; template bool _Rope_insert_char_consumer<_CharT, _Traits>:: operator()(const _CharT* __leaf, size_t __n) { size_t __i; // We assume that formatting is set up correctly for each element. for (__i = 0; __i < __n; __i++) _M_o.put(__leaf[__i]); return true; } template bool rope<_CharT, _Alloc>:: _S_apply_to_pieces(_Rope_char_consumer<_CharT>& __c, const _RopeRep* __r, size_t __begin, size_t __end) { if (0 == __r) return true; switch(__r->_M_tag) { case __detail::_S_concat: { _RopeConcatenation* __conc = (_RopeConcatenation*)__r; _RopeRep* __left = __conc->_M_left; size_t __left_len = __left->_M_size; if (__begin < __left_len) { size_t __left_end = std::min(__left_len, __end); if (!_S_apply_to_pieces(__c, __left, __begin, __left_end)) return false; } if (__end > __left_len) { _RopeRep* __right = __conc->_M_right; size_t __right_start = std::max(__left_len, __begin); if (!_S_apply_to_pieces(__c, __right, __right_start - __left_len, __end - __left_len)) return false; } } return true; case __detail::_S_leaf: { _RopeLeaf* __l = (_RopeLeaf*)__r; return __c(__l->_M_data + __begin, __end - __begin); } case __detail::_S_function: case __detail::_S_substringfn: { _RopeFunction* __f = (_RopeFunction*)__r; size_t __len = __end - __begin; bool __result; _CharT* __buffer = (_CharT*)_Alloc().allocate(__len * sizeof(_CharT)); try { (*(__f->_M_fn))(__begin, __len, __buffer); __result = __c(__buffer, __len); _Alloc().deallocate(__buffer, __len * sizeof(_CharT)); } catch(...) { _Alloc().deallocate(__buffer, __len * sizeof(_CharT)); __throw_exception_again; } return __result; } default: return false; } } template inline void _Rope_fill(basic_ostream<_CharT, _Traits>& __o, size_t __n) { char __f = __o.fill(); size_t __i; for (__i = 0; __i < __n; __i++) __o.put(__f); } template inline bool _Rope_is_simple(_CharT*) { return false; } inline bool _Rope_is_simple(char*) { return true; } inline bool _Rope_is_simple(wchar_t*) { return true; } template basic_ostream<_CharT, _Traits>& operator<<(basic_ostream<_CharT, _Traits>& __o, const rope<_CharT, _Alloc>& __r) { size_t __w = __o.width(); bool __left = bool(__o.flags() & std::ios::left); size_t __pad_len; size_t __rope_len = __r.size(); _Rope_insert_char_consumer<_CharT, _Traits> __c(__o); bool __is_simple = _Rope_is_simple((_CharT*)0); if (__rope_len < __w) __pad_len = __w - __rope_len; else __pad_len = 0; if (!__is_simple) __o.width(__w / __rope_len); try { if (__is_simple && !__left && __pad_len > 0) _Rope_fill(__o, __pad_len); __r.apply_to_pieces(0, __r.size(), __c); if (__is_simple && __left && __pad_len > 0) _Rope_fill(__o, __pad_len); if (!__is_simple) __o.width(__w); } catch(...) { if (!__is_simple) __o.width(__w); __throw_exception_again; } return __o; } template _CharT* rope<_CharT, _Alloc>:: _S_flatten(_RopeRep* __r, size_t __start, size_t __len, _CharT* __buffer) { _Rope_flatten_char_consumer<_CharT> __c(__buffer); _S_apply_to_pieces(__c, __r, __start, __start + __len); return(__buffer + __len); } template size_t rope<_CharT, _Alloc>:: find(_CharT __pattern, size_t __start) const { _Rope_find_char_char_consumer<_CharT> __c(__pattern); _S_apply_to_pieces(__c, this->_M_tree_ptr, __start, size()); size_type __result_pos = __start + __c._M_count; #ifndef __STL_OLD_ROPE_SEMANTICS if (__result_pos == size()) __result_pos = npos; #endif return __result_pos; } template _CharT* rope<_CharT, _Alloc>:: _S_flatten(_RopeRep* __r, _CharT* __buffer) { if (0 == __r) return __buffer; switch(__r->_M_tag) { case __detail::_S_concat: { _RopeConcatenation* __c = (_RopeConcatenation*)__r; _RopeRep* __left = __c->_M_left; _RopeRep* __right = __c->_M_right; _CharT* __rest = _S_flatten(__left, __buffer); return _S_flatten(__right, __rest); } case __detail::_S_leaf: { _RopeLeaf* __l = (_RopeLeaf*)__r; return copy_n(__l->_M_data, __l->_M_size, __buffer).second; } case __detail::_S_function: case __detail::_S_substringfn: // We don't yet do anything with substring nodes. // This needs to be fixed before ropefiles will work well. { _RopeFunction* __f = (_RopeFunction*)__r; (*(__f->_M_fn))(0, __f->_M_size, __buffer); return __buffer + __f->_M_size; } default: return 0; } } // This needs work for _CharT != char template void rope<_CharT, _Alloc>:: _S_dump(_RopeRep* __r, int __indent) { for (int __i = 0; __i < __indent; __i++) putchar(' '); if (0 == __r) { printf("NULL\n"); return; } if (_S_concat == __r->_M_tag) { _RopeConcatenation* __c = (_RopeConcatenation*)__r; _RopeRep* __left = __c->_M_left; _RopeRep* __right = __c->_M_right; #ifdef __GC printf("Concatenation %p (depth = %d, len = %ld, %s balanced)\n", __r, __r->_M_depth, __r->_M_size, __r->_M_is_balanced? "" : "not"); #else printf("Concatenation %p (rc = %ld, depth = %d, " "len = %ld, %s balanced)\n", __r, __r->_M_ref_count, __r->_M_depth, __r->_M_size, __r->_M_is_balanced? "" : "not"); #endif _S_dump(__left, __indent + 2); _S_dump(__right, __indent + 2); return; } else { char* __kind; switch (__r->_M_tag) { case __detail::_S_leaf: __kind = "Leaf"; break; case __detail::_S_function: __kind = "Function"; break; case __detail::_S_substringfn: __kind = "Function representing substring"; break; default: __kind = "(corrupted kind field!)"; } #ifdef __GC printf("%s %p (depth = %d, len = %ld) ", __kind, __r, __r->_M_depth, __r->_M_size); #else printf("%s %p (rc = %ld, depth = %d, len = %ld) ", __kind, __r, __r->_M_ref_count, __r->_M_depth, __r->_M_size); #endif if (_S_is_one_byte_char_type((_CharT*)0)) { const int __max_len = 40; _Self_destruct_ptr __prefix(_S_substring(__r, 0, __max_len)); _CharT __buffer[__max_len + 1]; bool __too_big = __r->_M_size > __prefix->_M_size; _S_flatten(__prefix, __buffer); __buffer[__prefix->_M_size] = _S_eos((_CharT*)0); printf("%s%s\n", (char*)__buffer, __too_big? "...\n" : "\n"); } else printf("\n"); } } template const unsigned long rope<_CharT, _Alloc>:: _S_min_len[int(__detail::_S_max_rope_depth) + 1] = { /* 0 */1, /* 1 */2, /* 2 */3, /* 3 */5, /* 4 */8, /* 5 */13, /* 6 */21, /* 7 */34, /* 8 */55, /* 9 */89, /* 10 */144, /* 11 */233, /* 12 */377, /* 13 */610, /* 14 */987, /* 15 */1597, /* 16 */2584, /* 17 */4181, /* 18 */6765, /* 19 */10946, /* 20 */17711, /* 21 */28657, /* 22 */46368, /* 23 */75025, /* 24 */121393, /* 25 */196418, /* 26 */317811, /* 27 */514229, /* 28 */832040, /* 29 */1346269, /* 30 */2178309, /* 31 */3524578, /* 32 */5702887, /* 33 */9227465, /* 34 */14930352, /* 35 */24157817, /* 36 */39088169, /* 37 */63245986, /* 38 */102334155, /* 39 */165580141, /* 40 */267914296, /* 41 */433494437, /* 42 */701408733, /* 43 */1134903170, /* 44 */1836311903, /* 45 */2971215073u }; // These are Fibonacci numbers < 2**32. template typename rope<_CharT, _Alloc>::_RopeRep* rope<_CharT, _Alloc>:: _S_balance(_RopeRep* __r) { _RopeRep* __forest[int(__detail::_S_max_rope_depth) + 1]; _RopeRep* __result = 0; int __i; // Invariant: // The concatenation of forest in descending order is equal to __r. // __forest[__i]._M_size >= _S_min_len[__i] // __forest[__i]._M_depth = __i // References from forest are included in refcount. for (__i = 0; __i <= int(__detail::_S_max_rope_depth); ++__i) __forest[__i] = 0; try { _S_add_to_forest(__r, __forest); for (__i = 0; __i <= int(__detail::_S_max_rope_depth); ++__i) if (0 != __forest[__i]) { #ifndef __GC _Self_destruct_ptr __old(__result); #endif __result = _S_concat(__forest[__i], __result); __forest[__i]->_M_unref_nonnil(); #if !defined(__GC) && defined(__EXCEPTIONS) __forest[__i] = 0; #endif } } catch(...) { for(__i = 0; __i <= int(__detail::_S_max_rope_depth); __i++) _S_unref(__forest[__i]); __throw_exception_again; } if (__result->_M_depth > int(__detail::_S_max_rope_depth)) __throw_length_error(__N("rope::_S_balance")); return(__result); } template void rope<_CharT, _Alloc>:: _S_add_to_forest(_RopeRep* __r, _RopeRep** __forest) { if (__r->_M_is_balanced) { _S_add_leaf_to_forest(__r, __forest); return; } { _RopeConcatenation* __c = (_RopeConcatenation*)__r; _S_add_to_forest(__c->_M_left, __forest); _S_add_to_forest(__c->_M_right, __forest); } } template void rope<_CharT, _Alloc>:: _S_add_leaf_to_forest(_RopeRep* __r, _RopeRep** __forest) { _RopeRep* __insertee; // included in refcount _RopeRep* __too_tiny = 0; // included in refcount int __i; // forest[0..__i-1] is empty size_t __s = __r->_M_size; for (__i = 0; __s >= _S_min_len[__i+1]/* not this bucket */; ++__i) { if (0 != __forest[__i]) { #ifndef __GC _Self_destruct_ptr __old(__too_tiny); #endif __too_tiny = _S_concat_and_set_balanced(__forest[__i], __too_tiny); __forest[__i]->_M_unref_nonnil(); __forest[__i] = 0; } } { #ifndef __GC _Self_destruct_ptr __old(__too_tiny); #endif __insertee = _S_concat_and_set_balanced(__too_tiny, __r); } // Too_tiny dead, and no longer included in refcount. // Insertee is live and included. for (;; ++__i) { if (0 != __forest[__i]) { #ifndef __GC _Self_destruct_ptr __old(__insertee); #endif __insertee = _S_concat_and_set_balanced(__forest[__i], __insertee); __forest[__i]->_M_unref_nonnil(); __forest[__i] = 0; } if (__i == int(__detail::_S_max_rope_depth) || __insertee->_M_size < _S_min_len[__i+1]) { __forest[__i] = __insertee; // refcount is OK since __insertee is now dead. return; } } } template _CharT rope<_CharT, _Alloc>:: _S_fetch(_RopeRep* __r, size_type __i) { __GC_CONST _CharT* __cstr = __r->_M_c_string; if (0 != __cstr) return __cstr[__i]; for(;;) { switch(__r->_M_tag) { case __detail::_S_concat: { _RopeConcatenation* __c = (_RopeConcatenation*)__r; _RopeRep* __left = __c->_M_left; size_t __left_len = __left->_M_size; if (__i >= __left_len) { __i -= __left_len; __r = __c->_M_right; } else __r = __left; } break; case __detail::_S_leaf: { _RopeLeaf* __l = (_RopeLeaf*)__r; return __l->_M_data[__i]; } case __detail::_S_function: case __detail::_S_substringfn: { _RopeFunction* __f = (_RopeFunction*)__r; _CharT __result; (*(__f->_M_fn))(__i, 1, &__result); return __result; } } } } #ifndef __GC // Return a uniquely referenced character slot for the given // position, or 0 if that's not possible. template _CharT* rope<_CharT, _Alloc>:: _S_fetch_ptr(_RopeRep* __r, size_type __i) { _RopeRep* __clrstack[__detail::_S_max_rope_depth]; size_t __csptr = 0; for(;;) { if (__r->_M_ref_count > 1) return 0; switch(__r->_M_tag) { case __detail::_S_concat: { _RopeConcatenation* __c = (_RopeConcatenation*)__r; _RopeRep* __left = __c->_M_left; size_t __left_len = __left->_M_size; if (__c->_M_c_string != 0) __clrstack[__csptr++] = __c; if (__i >= __left_len) { __i -= __left_len; __r = __c->_M_right; } else __r = __left; } break; case __detail::_S_leaf: { _RopeLeaf* __l = (_RopeLeaf*)__r; if (__l->_M_c_string != __l->_M_data && __l->_M_c_string != 0) __clrstack[__csptr++] = __l; while (__csptr > 0) { -- __csptr; _RopeRep* __d = __clrstack[__csptr]; __d->_M_free_c_string(); __d->_M_c_string = 0; } return __l->_M_data + __i; } case __detail::_S_function: case __detail::_S_substringfn: return 0; } } } #endif /* __GC */ // The following could be implemented trivially using // lexicographical_compare_3way. // We do a little more work to avoid dealing with rope iterators for // flat strings. template int rope<_CharT, _Alloc>:: _S_compare (const _RopeRep* __left, const _RopeRep* __right) { size_t __left_len; size_t __right_len; if (0 == __right) return 0 != __left; if (0 == __left) return -1; __left_len = __left->_M_size; __right_len = __right->_M_size; if (__detail::_S_leaf == __left->_M_tag) { _RopeLeaf* __l = (_RopeLeaf*) __left; if (__detail::_S_leaf == __right->_M_tag) { _RopeLeaf* __r = (_RopeLeaf*) __right; return lexicographical_compare_3way(__l->_M_data, __l->_M_data + __left_len, __r->_M_data, __r->_M_data + __right_len); } else { const_iterator __rstart(__right, 0); const_iterator __rend(__right, __right_len); return lexicographical_compare_3way(__l->_M_data, __l->_M_data + __left_len, __rstart, __rend); } } else { const_iterator __lstart(__left, 0); const_iterator __lend(__left, __left_len); if (__detail::_S_leaf == __right->_M_tag) { _RopeLeaf* __r = (_RopeLeaf*) __right; return lexicographical_compare_3way(__lstart, __lend, __r->_M_data, __r->_M_data + __right_len); } else { const_iterator __rstart(__right, 0); const_iterator __rend(__right, __right_len); return lexicographical_compare_3way(__lstart, __lend, __rstart, __rend); } } } // Assignment to reference proxies. template _Rope_char_ref_proxy<_CharT, _Alloc>& _Rope_char_ref_proxy<_CharT, _Alloc>:: operator=(_CharT __c) { _RopeRep* __old = _M_root->_M_tree_ptr; #ifndef __GC // First check for the case in which everything is uniquely // referenced. In that case we can do this destructively. _CharT* __ptr = _My_rope::_S_fetch_ptr(__old, _M_pos); if (0 != __ptr) { *__ptr = __c; return *this;