1 // wstring_convert implementation -*- C++ -*-
3 // Copyright (C) 2012 Free Software Foundation, Inc.
5 // This file is part of the GNU ISO C++ Library. This library is free
6 // software; you can redistribute it and/or modify it under the
7 // terms of the GNU General Public License as published by the
8 // Free Software Foundation; either version 3, or (at your option)
11 // This library is distributed in the hope that it will be useful,
12 // but WITHOUT ANY WARRANTY; without even the implied warranty of
13 // MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14 // GNU General Public License for more details.
16 // Under Section 7 of GPL version 3, you are granted additional
17 // permissions described in the GCC Runtime Library Exception, version
18 // 3.1, as published by the Free Software Foundation.
20 // You should have received a copy of the GNU General Public License and
21 // a copy of the GCC Runtime Library Exception along with this program;
22 // see the files COPYING3 and COPYING.RUNTIME respectively. If not, see
23 // <http://www.gnu.org/licenses/>.
25 /** @file bits/locale_conv.h
26 * This is an internal header file, included by other library headers.
27 * Do not attempt to use it directly. @headername{locale}
30 #ifndef _LOCALE_CONV_H
31 #define _LOCALE_CONV_H 1
33 #if __cplusplus < 201103L
34 # include <bits/c++0x_warning.h>
38 #include "stringfwd.h"
39 #include "allocator.h"
41 #include "unique_ptr.h"
43 namespace std _GLIBCXX_VISIBILITY(default)
45 _GLIBCXX_BEGIN_NAMESPACE_VERSION
47 #ifdef _GLIBCXX_USE_WCHAR_T
54 /// String conversions
55 template<typename _Codecvt, typename _Elem = wchar_t,
56 typename _Wide_alloc = allocator<_Elem>,
57 typename _Byte_alloc = allocator<char>>
61 typedef basic_string<char, char_traits<char>, _Byte_alloc> byte_string;
62 typedef basic_string<_Elem, char_traits<_Elem>, _Wide_alloc> wide_string;
63 typedef typename _Codecvt::state_type state_type;
64 typedef typename wide_string::traits_type::int_type int_type;
66 /** Default constructor.
68 * @param __pcvt The facet to use for conversions.
70 * Takes ownership of @p __pcvt and will delete it in the destructor.
73 wstring_convert(_Codecvt* __pcvt = new _Codecvt()) : _M_cvt(__pcvt)
76 __throw_logic_error("wstring_convert");
79 /** Construct with an initial converstion state.
81 * @param __pcvt The facet to use for conversions.
82 * @param __state Initial conversion state.
84 * Takes ownership of @p __pcvt and will delete it in the destructor.
85 * The object's conversion state will persist between conversions.
87 wstring_convert(_Codecvt* __pcvt, state_type __state)
88 : _M_cvt(__pcvt), _M_state(__state), _M_with_cvtstate(true)
91 __throw_logic_error("wstring_convert");
94 /** Construct with error strings.
96 * @param __byte_err A string to return on failed conversions.
97 * @param __wide_err A wide string to return on failed conversions.
100 wstring_convert(const byte_string& __byte_err,
101 const wide_string& __wide_err = wide_string())
102 : _M_cvt(new _Codecvt),
103 _M_byte_err_string(__byte_err), _M_wide_err_string(__wide_err),
104 _M_with_strings(true)
107 __throw_logic_error("wstring_convert");
110 ~wstring_convert() = default;
112 // _GLIBCXX_RESOLVE_LIB_DEFECTS
113 // 2176. Special members for wstring_convert and wbuffer_convert
114 wstring_convert(const wstring_convert&) = delete;
115 wstring_convert& operator=(const wstring_convert&) = delete;
117 /// @{ Convert from bytes.
119 from_bytes(char __byte)
121 char __bytes[2] = { __byte };
122 return from_bytes(__bytes, __bytes+1);
126 from_bytes(const char* __ptr)
127 { return from_bytes(__ptr, __ptr+char_traits<char>::length(__ptr)); }
130 from_bytes(const byte_string& __str)
132 auto __ptr = __str.data();
133 return from_bytes(__ptr, __ptr + __str.size());
137 from_bytes(const char* __first, const char* __last)
139 auto __errstr = _M_with_strings ? &_M_wide_err_string : nullptr;
140 _ConvFn<char, _Elem> __fn = &_Codecvt::in;
141 return _M_conv(__first, __last, __errstr, __fn);
145 /// @{ Convert to bytes.
147 to_bytes(_Elem __wchar)
149 _Elem __wchars[2] = { __wchar };
150 return to_bytes(__wchars, __wchars+1);
154 to_bytes(const _Elem* __ptr)
156 return to_bytes(__ptr, __ptr+wide_string::traits_type::length(__ptr));
160 to_bytes(const wide_string& __wstr)
162 auto __ptr = __wstr.data();
163 return to_bytes(__ptr, __ptr + __wstr.size());
167 to_bytes(const _Elem* __first, const _Elem* __last)
169 auto __errstr = _M_with_strings ? &_M_byte_err_string : nullptr;
170 _ConvFn<_Elem, char> __fn = &_Codecvt::out;
171 return _M_conv(__first, __last, __errstr, __fn);
175 // _GLIBCXX_RESOLVE_LIB_DEFECTS
176 // 2174. wstring_convert::converted() should be noexcept
177 /// The number of elements successfully converted in the last conversion.
178 size_t converted() const noexcept { return _M_count; }
180 /// The final conversion state of the last conversion.
181 state_type state() const { return _M_state; }
184 template<typename _InC, typename _OutC>
186 = codecvt_base::result
187 (_Codecvt::*)(state_type&, const _InC*, const _InC*, const _InC*&,
188 _OutC*, _OutC*, _OutC*&) const;
190 template<typename _InChar, typename _OutStr, typename _MemFn>
192 _M_conv(const _InChar* __first, const _InChar* __last,
193 const _OutStr* __err, _MemFn __memfn)
195 if (!_M_with_cvtstate)
196 _M_state = state_type();
198 auto __outstr = __err ? _OutStr(__err->get_allocator()) : _OutStr();
199 size_t __outchars = 0;
200 auto __next = __first;
201 const auto __maxlen = _M_cvt->max_length();
203 codecvt_base::result __result;
206 __outstr.resize(__outstr.size() + (__last - __next) + __maxlen);
207 auto __outnext = &__outstr.front() + __outchars;
208 auto const __outlast = &__outstr.back() + 1;
209 __result = ((*_M_cvt).*__memfn)(_M_state, __next, __last, __next,
210 __outnext, __outlast, __outnext);
211 __outchars = __outnext - &__outstr.front();
213 while (__result == codecvt_base::partial && __next != __last
214 && (__outstr.size() - __outchars) < __maxlen);
216 if (__result == codecvt_base::noconv)
218 __outstr.assign(__first, __last);
219 _M_count = __outstr.size();
223 __outstr.resize(__outchars);
224 _M_count = __next - __first;
226 if (__result != codecvt_base::error)
231 __throw_range_error("wstring_convert");
234 unique_ptr<_Codecvt> _M_cvt;
235 byte_string _M_byte_err_string;
236 wide_string _M_wide_err_string;
237 state_type _M_state = state_type();
239 bool _M_with_cvtstate = false;
240 bool _M_with_strings = false;
243 /// Buffer conversions
244 template<typename _Codecvt, typename _Elem = wchar_t,
245 typename _Tr = char_traits<_Elem>>
246 class wbuffer_convert : public basic_streambuf<_Elem, _Tr>
248 typedef basic_streambuf<_Elem, _Tr> _Wide_streambuf;
251 typedef typename _Codecvt::state_type state_type;
253 /** Default constructor.
255 * @param __bytebuf The underlying byte stream buffer.
256 * @param __pcvt The facet to use for conversions.
257 * @param __state Initial conversion state.
259 * Takes ownership of @p __pcvt and will delete it in the destructor.
262 wbuffer_convert(streambuf* __bytebuf = 0, _Codecvt* __pcvt = new _Codecvt,
263 state_type __state = state_type())
264 : _M_buf(__bytebuf), _M_cvt(__pcvt), _M_state(__state)
267 __throw_logic_error("wstring_convert");
269 _M_always_noconv = _M_cvt->always_noconv();
273 this->setp(_M_put_area, _M_put_area + _S_buffer_length);
274 this->setg(_M_get_area + _S_putback_length,
275 _M_get_area + _S_putback_length,
276 _M_get_area + _S_putback_length);
280 ~wbuffer_convert() = default;
282 // _GLIBCXX_RESOLVE_LIB_DEFECTS
283 // 2176. Special members for wstring_convert and wbuffer_convert
284 wbuffer_convert(const wbuffer_convert&) = delete;
285 wbuffer_convert& operator=(const wbuffer_convert&) = delete;
287 streambuf* rdbuf() const noexcept { return _M_buf; }
290 rdbuf(streambuf *__bytebuf) noexcept
292 auto __prev = _M_buf;
297 /// The conversion state following the last conversion.
298 state_type state() const noexcept { return _M_state; }
303 { return _M_buf && _M_conv_put() && _M_buf->pubsync() ? 0 : -1; }
305 typename _Wide_streambuf::int_type
306 overflow(typename _Wide_streambuf::int_type __out)
308 if (!_M_buf || !_M_conv_put())
310 else if (!_Tr::eq_int_type(__out, _Tr::eof()))
311 return this->sputc(__out);
312 return _Tr::not_eof(__out);
315 typename _Wide_streambuf::int_type
321 if (this->gptr() < this->egptr() || (_M_buf && _M_conv_get()))
322 return _Tr::to_int_type(*this->gptr());
328 xsputn(const typename _Wide_streambuf::char_type* __s, streamsize __n)
330 if (!_M_buf || __n == 0)
332 streamsize __done = 0;
335 auto __nn = std::min<streamsize>(this->epptr() - this->pptr(),
337 _Tr::copy(this->pptr(), __s + __done, __nn);
340 } while (__done < __n && _M_conv_put());
345 // fill the get area from converted contents of the byte stream buffer
349 const streamsize __pb1 = this->gptr() - this->eback();
350 const streamsize __pb2 = _S_putback_length;
351 const streamsize __npb = std::min(__pb1, __pb2);
353 _Tr::move(_M_get_area + _S_putback_length - __npb,
354 this->gptr() - __npb, __npb);
356 streamsize __nbytes = sizeof(_M_get_buf) - _M_unconv;
357 __nbytes = std::min(__nbytes, _M_buf->in_avail());
360 __nbytes = _M_buf->sgetn(_M_get_buf + _M_unconv, __nbytes);
363 __nbytes += _M_unconv;
365 // convert _M_get_buf into _M_get_area
367 _Elem* __outbuf = _M_get_area + _S_putback_length;
368 _Elem* __outnext = __outbuf;
369 const char* __bnext = _M_get_buf;
371 codecvt_base::result __result;
372 if (_M_always_noconv)
373 __result = codecvt_base::noconv;
376 _Elem* __outend = _M_get_area + _S_buffer_length;
378 __result = _M_cvt->in(_M_state,
379 __bnext, __bnext + __nbytes, __bnext,
380 __outbuf, __outend, __outnext);
383 if (__result == codecvt_base::noconv)
385 // cast is safe because noconv means _Elem is same type as char
386 auto __get_buf = reinterpret_cast<const _Elem*>(_M_get_buf);
387 _Tr::copy(__outbuf, __get_buf, __nbytes);
392 if ((_M_unconv = _M_get_buf + __nbytes - __bnext))
393 char_traits<char>::move(_M_get_buf, __bnext, _M_unconv);
395 this->setg(__outbuf, __outbuf, __outnext);
397 return __result != codecvt_base::error;
406 _M_put(const char* __p, streamsize __n)
408 if (_M_buf->sputn(__p, __n) < __n)
412 // convert the put area and write to the byte stream buffer
416 _Elem* const __first = this->pbase();
417 const _Elem* const __last = this->pptr();
418 const streamsize __pending = __last - __first;
420 if (_M_always_noconv)
421 return _M_put(__first, __pending);
423 char __outbuf[2 * _S_buffer_length];
425 const _Elem* __next = __first;
426 const _Elem* __start;
430 char* __outnext = __outbuf;
431 char* const __outlast = __outbuf + sizeof(__outbuf);
432 auto __result = _M_cvt->out(_M_state, __next, __last, __next,
433 __outnext, __outlast, __outnext);
434 if (__result == codecvt_base::error)
436 else if (__result == codecvt_base::noconv)
437 return _M_put(__next, __pending);
439 if (!_M_put(__outbuf, __outnext - __outbuf))
442 while (__next != __last && __next != __start);
444 if (__next != __last)
445 _Tr::move(__first, __next, __last - __next);
447 this->pbump(__first - __next);
448 return __next != __first;
452 unique_ptr<_Codecvt> _M_cvt;
455 static const streamsize _S_buffer_length = 32;
456 static const streamsize _S_putback_length = 3;
457 _Elem _M_put_area[_S_buffer_length];
458 _Elem _M_get_area[_S_buffer_length];
459 streamsize _M_unconv = 0;
460 char _M_get_buf[_S_buffer_length-_S_putback_length];
461 bool _M_always_noconv;
466 #endif // _GLIBCXX_USE_WCHAR_T
468 _GLIBCXX_END_NAMESPACE_VERSION
471 #endif // __cplusplus
473 #endif /* _LOCALE_CONV_H */