1 // wstring_convert implementation -*- C++ -*-
3 // Copyright (C) 2015-2017 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
52 template<typename _OutStr
, typename _InChar
, typename _Codecvt
,
53 typename _State
, typename _Fn
>
55 __do_str_codecvt(const _InChar
* __first
, const _InChar
* __last
,
56 _OutStr
& __outstr
, const _Codecvt
& __cvt
, _State
& __state
,
57 size_t& __count
, _Fn __fn
)
59 if (__first
== __last
)
66 size_t __outchars
= 0;
67 auto __next
= __first
;
68 const auto __maxlen
= __cvt
.max_length() + 1;
70 codecvt_base::result __result
;
73 __outstr
.resize(__outstr
.size() + (__last
- __next
) * __maxlen
);
74 auto __outnext
= &__outstr
.front() + __outchars
;
75 auto const __outlast
= &__outstr
.back() + 1;
76 __result
= (__cvt
.*__fn
)(__state
, __next
, __last
, __next
,
77 __outnext
, __outlast
, __outnext
);
78 __outchars
= __outnext
- &__outstr
.front();
80 while (__result
== codecvt_base::partial
&& __next
!= __last
81 && (__outstr
.size() - __outchars
) < __maxlen
);
83 if (__result
== codecvt_base::error
)
85 __count
= __next
- __first
;
89 if (__result
== codecvt_base::noconv
)
91 __outstr
.assign(__first
, __last
);
92 __count
= __last
- __first
;
96 __outstr
.resize(__outchars
);
97 __count
= __next
- __first
;
103 // Convert narrow character string to wide.
104 template<typename _CharT
, typename _Traits
, typename _Alloc
, typename _State
>
106 __str_codecvt_in(const char* __first
, const char* __last
,
107 basic_string
<_CharT
, _Traits
, _Alloc
>& __outstr
,
108 const codecvt
<_CharT
, char, _State
>& __cvt
,
109 _State
& __state
, size_t& __count
)
111 using _Codecvt
= codecvt
<_CharT
, char, _State
>;
113 = codecvt_base::result
114 (_Codecvt::*)(_State
&, const char*, const char*, const char*&,
115 _CharT
*, _CharT
*, _CharT
*&) const;
116 _ConvFn __fn
= &codecvt
<_CharT
, char, _State
>::in
;
117 return __do_str_codecvt(__first
, __last
, __outstr
, __cvt
, __state
,
121 template<typename _CharT
, typename _Traits
, typename _Alloc
, typename _State
>
123 __str_codecvt_in(const char* __first
, const char* __last
,
124 basic_string
<_CharT
, _Traits
, _Alloc
>& __outstr
,
125 const codecvt
<_CharT
, char, _State
>& __cvt
)
129 return __str_codecvt_in(__first
, __last
, __outstr
, __cvt
, __state
, __n
);
132 // Convert wide character string to narrow.
133 template<typename _CharT
, typename _Traits
, typename _Alloc
, typename _State
>
135 __str_codecvt_out(const _CharT
* __first
, const _CharT
* __last
,
136 basic_string
<char, _Traits
, _Alloc
>& __outstr
,
137 const codecvt
<_CharT
, char, _State
>& __cvt
,
138 _State
& __state
, size_t& __count
)
140 using _Codecvt
= codecvt
<_CharT
, char, _State
>;
142 = codecvt_base::result
143 (_Codecvt::*)(_State
&, const _CharT
*, const _CharT
*, const _CharT
*&,
144 char*, char*, char*&) const;
145 _ConvFn __fn
= &codecvt
<_CharT
, char, _State
>::out
;
146 return __do_str_codecvt(__first
, __last
, __outstr
, __cvt
, __state
,
150 template<typename _CharT
, typename _Traits
, typename _Alloc
, typename _State
>
152 __str_codecvt_out(const _CharT
* __first
, const _CharT
* __last
,
153 basic_string
<char, _Traits
, _Alloc
>& __outstr
,
154 const codecvt
<_CharT
, char, _State
>& __cvt
)
158 return __str_codecvt_out(__first
, __last
, __outstr
, __cvt
, __state
, __n
);
161 #ifdef _GLIBCXX_USE_WCHAR_T
163 _GLIBCXX_BEGIN_NAMESPACE_CXX11
165 /// String conversions
166 template<typename _Codecvt
, typename _Elem
= wchar_t,
167 typename _Wide_alloc
= allocator
<_Elem
>,
168 typename _Byte_alloc
= allocator
<char>>
169 class wstring_convert
172 typedef basic_string
<char, char_traits
<char>, _Byte_alloc
> byte_string
;
173 typedef basic_string
<_Elem
, char_traits
<_Elem
>, _Wide_alloc
> wide_string
;
174 typedef typename
_Codecvt::state_type state_type
;
175 typedef typename
wide_string::traits_type::int_type int_type
;
177 /** Default constructor.
179 * @param __pcvt The facet to use for conversions.
181 * Takes ownership of @p __pcvt and will delete it in the destructor.
184 wstring_convert(_Codecvt
* __pcvt
= new _Codecvt()) : _M_cvt(__pcvt
)
187 __throw_logic_error("wstring_convert");
190 /** Construct with an initial converstion state.
192 * @param __pcvt The facet to use for conversions.
193 * @param __state Initial conversion state.
195 * Takes ownership of @p __pcvt and will delete it in the destructor.
196 * The object's conversion state will persist between conversions.
198 wstring_convert(_Codecvt
* __pcvt
, state_type __state
)
199 : _M_cvt(__pcvt
), _M_state(__state
), _M_with_cvtstate(true)
202 __throw_logic_error("wstring_convert");
205 /** Construct with error strings.
207 * @param __byte_err A string to return on failed conversions.
208 * @param __wide_err A wide string to return on failed conversions.
211 wstring_convert(const byte_string
& __byte_err
,
212 const wide_string
& __wide_err
= wide_string())
213 : _M_cvt(new _Codecvt
),
214 _M_byte_err_string(__byte_err
), _M_wide_err_string(__wide_err
),
215 _M_with_strings(true)
218 __throw_logic_error("wstring_convert");
221 ~wstring_convert() = default;
223 // _GLIBCXX_RESOLVE_LIB_DEFECTS
224 // 2176. Special members for wstring_convert and wbuffer_convert
225 wstring_convert(const wstring_convert
&) = delete;
226 wstring_convert
& operator=(const wstring_convert
&) = delete;
228 /// @{ Convert from bytes.
230 from_bytes(char __byte
)
232 char __bytes
[2] = { __byte
};
233 return from_bytes(__bytes
, __bytes
+1);
237 from_bytes(const char* __ptr
)
238 { return from_bytes(__ptr
, __ptr
+char_traits
<char>::length(__ptr
)); }
241 from_bytes(const byte_string
& __str
)
243 auto __ptr
= __str
.data();
244 return from_bytes(__ptr
, __ptr
+ __str
.size());
248 from_bytes(const char* __first
, const char* __last
)
250 if (!_M_with_cvtstate
)
251 _M_state
= state_type();
252 wide_string __out
{ _M_wide_err_string
.get_allocator() };
253 if (__str_codecvt_in(__first
, __last
, __out
, *_M_cvt
, _M_state
,
257 return _M_wide_err_string
;
258 __throw_range_error("wstring_convert::from_bytes");
262 /// @{ Convert to bytes.
264 to_bytes(_Elem __wchar
)
266 _Elem __wchars
[2] = { __wchar
};
267 return to_bytes(__wchars
, __wchars
+1);
271 to_bytes(const _Elem
* __ptr
)
273 return to_bytes(__ptr
, __ptr
+wide_string::traits_type::length(__ptr
));
277 to_bytes(const wide_string
& __wstr
)
279 auto __ptr
= __wstr
.data();
280 return to_bytes(__ptr
, __ptr
+ __wstr
.size());
284 to_bytes(const _Elem
* __first
, const _Elem
* __last
)
286 if (!_M_with_cvtstate
)
287 _M_state
= state_type();
288 byte_string __out
{ _M_byte_err_string
.get_allocator() };
289 if (__str_codecvt_out(__first
, __last
, __out
, *_M_cvt
, _M_state
,
293 return _M_byte_err_string
;
294 __throw_range_error("wstring_convert::to_bytes");
298 // _GLIBCXX_RESOLVE_LIB_DEFECTS
299 // 2174. wstring_convert::converted() should be noexcept
300 /// The number of elements successfully converted in the last conversion.
301 size_t converted() const noexcept
{ return _M_count
; }
303 /// The final conversion state of the last conversion.
304 state_type
state() const { return _M_state
; }
307 unique_ptr
<_Codecvt
> _M_cvt
;
308 byte_string _M_byte_err_string
;
309 wide_string _M_wide_err_string
;
310 state_type _M_state
= state_type();
312 bool _M_with_cvtstate
= false;
313 bool _M_with_strings
= false;
316 _GLIBCXX_END_NAMESPACE_CXX11
318 /// Buffer conversions
319 template<typename _Codecvt
, typename _Elem
= wchar_t,
320 typename _Tr
= char_traits
<_Elem
>>
321 class wbuffer_convert
: public basic_streambuf
<_Elem
, _Tr
>
323 typedef basic_streambuf
<_Elem
, _Tr
> _Wide_streambuf
;
326 typedef typename
_Codecvt::state_type state_type
;
328 /** Default constructor.
330 * @param __bytebuf The underlying byte stream buffer.
331 * @param __pcvt The facet to use for conversions.
332 * @param __state Initial conversion state.
334 * Takes ownership of @p __pcvt and will delete it in the destructor.
337 wbuffer_convert(streambuf
* __bytebuf
= 0, _Codecvt
* __pcvt
= new _Codecvt
,
338 state_type __state
= state_type())
339 : _M_buf(__bytebuf
), _M_cvt(__pcvt
), _M_state(__state
)
342 __throw_logic_error("wbuffer_convert");
344 _M_always_noconv
= _M_cvt
->always_noconv();
348 this->setp(_M_put_area
, _M_put_area
+ _S_buffer_length
);
349 this->setg(_M_get_area
+ _S_putback_length
,
350 _M_get_area
+ _S_putback_length
,
351 _M_get_area
+ _S_putback_length
);
355 ~wbuffer_convert() = default;
357 // _GLIBCXX_RESOLVE_LIB_DEFECTS
358 // 2176. Special members for wstring_convert and wbuffer_convert
359 wbuffer_convert(const wbuffer_convert
&) = delete;
360 wbuffer_convert
& operator=(const wbuffer_convert
&) = delete;
362 streambuf
* rdbuf() const noexcept
{ return _M_buf
; }
365 rdbuf(streambuf
*__bytebuf
) noexcept
367 auto __prev
= _M_buf
;
372 /// The conversion state following the last conversion.
373 state_type
state() const noexcept
{ return _M_state
; }
378 { return _M_buf
&& _M_conv_put() && !_M_buf
->pubsync() ? 0 : -1; }
380 typename
_Wide_streambuf::int_type
381 overflow(typename
_Wide_streambuf::int_type __out
)
383 if (!_M_buf
|| !_M_conv_put())
385 else if (!_Tr::eq_int_type(__out
, _Tr::eof()))
386 return this->sputc(__out
);
387 return _Tr::not_eof(__out
);
390 typename
_Wide_streambuf::int_type
396 if (this->gptr() < this->egptr() || (_M_buf
&& _M_conv_get()))
397 return _Tr::to_int_type(*this->gptr());
403 xsputn(const typename
_Wide_streambuf::char_type
* __s
, streamsize __n
)
405 if (!_M_buf
|| __n
== 0)
407 streamsize __done
= 0;
410 auto __nn
= std::min
<streamsize
>(this->epptr() - this->pptr(),
412 _Tr::copy(this->pptr(), __s
+ __done
, __nn
);
415 } while (__done
< __n
&& _M_conv_put());
420 // fill the get area from converted contents of the byte stream buffer
424 const streamsize __pb1
= this->gptr() - this->eback();
425 const streamsize __pb2
= _S_putback_length
;
426 const streamsize __npb
= std::min(__pb1
, __pb2
);
428 _Tr::move(_M_get_area
+ _S_putback_length
- __npb
,
429 this->gptr() - __npb
, __npb
);
431 streamsize __nbytes
= sizeof(_M_get_buf
) - _M_unconv
;
432 __nbytes
= std::min(__nbytes
, _M_buf
->in_avail());
435 __nbytes
= _M_buf
->sgetn(_M_get_buf
+ _M_unconv
, __nbytes
);
438 __nbytes
+= _M_unconv
;
440 // convert _M_get_buf into _M_get_area
442 _Elem
* __outbuf
= _M_get_area
+ _S_putback_length
;
443 _Elem
* __outnext
= __outbuf
;
444 const char* __bnext
= _M_get_buf
;
446 codecvt_base::result __result
;
447 if (_M_always_noconv
)
448 __result
= codecvt_base::noconv
;
451 _Elem
* __outend
= _M_get_area
+ _S_buffer_length
;
453 __result
= _M_cvt
->in(_M_state
,
454 __bnext
, __bnext
+ __nbytes
, __bnext
,
455 __outbuf
, __outend
, __outnext
);
458 if (__result
== codecvt_base::noconv
)
460 // cast is safe because noconv means _Elem is same type as char
461 auto __get_buf
= reinterpret_cast<const _Elem
*>(_M_get_buf
);
462 _Tr::copy(__outbuf
, __get_buf
, __nbytes
);
467 if ((_M_unconv
= _M_get_buf
+ __nbytes
- __bnext
))
468 char_traits
<char>::move(_M_get_buf
, __bnext
, _M_unconv
);
470 this->setg(__outbuf
, __outbuf
, __outnext
);
472 return __result
!= codecvt_base::error
;
481 _M_put(const char* __p
, streamsize __n
)
483 if (_M_buf
->sputn(__p
, __n
) < __n
)
488 // convert the put area and write to the byte stream buffer
492 _Elem
* const __first
= this->pbase();
493 const _Elem
* const __last
= this->pptr();
494 const streamsize __pending
= __last
- __first
;
496 if (_M_always_noconv
)
497 return _M_put(__first
, __pending
);
499 char __outbuf
[2 * _S_buffer_length
];
501 const _Elem
* __next
= __first
;
502 const _Elem
* __start
;
506 char* __outnext
= __outbuf
;
507 char* const __outlast
= __outbuf
+ sizeof(__outbuf
);
508 auto __result
= _M_cvt
->out(_M_state
, __next
, __last
, __next
,
509 __outnext
, __outlast
, __outnext
);
510 if (__result
== codecvt_base::error
)
512 else if (__result
== codecvt_base::noconv
)
513 return _M_put(__next
, __pending
);
515 if (!_M_put(__outbuf
, __outnext
- __outbuf
))
518 while (__next
!= __last
&& __next
!= __start
);
520 if (__next
!= __last
)
521 _Tr::move(__first
, __next
, __last
- __next
);
523 this->pbump(__first
- __next
);
524 return __next
!= __first
;
528 unique_ptr
<_Codecvt
> _M_cvt
;
531 static const streamsize _S_buffer_length
= 32;
532 static const streamsize _S_putback_length
= 3;
533 _Elem _M_put_area
[_S_buffer_length
];
534 _Elem _M_get_area
[_S_buffer_length
];
535 streamsize _M_unconv
= 0;
536 char _M_get_buf
[_S_buffer_length
-_S_putback_length
];
537 bool _M_always_noconv
;
540 #endif // _GLIBCXX_USE_WCHAR_T
544 _GLIBCXX_END_NAMESPACE_VERSION
547 #endif // __cplusplus
549 #endif /* _LOCALE_CONV_H */