1 // Locale support (codecvt) -*- C++ -*-
3 // Copyright (C) 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007
4 // Free Software Foundation, Inc.
6 // This file is part of the GNU ISO C++ Library. This library is free
7 // software; you can redistribute it and/or modify it under the
8 // terms of the GNU General Public License as published by the
9 // Free Software Foundation; either version 2, or (at your option)
12 // This library is distributed in the hope that it will be useful,
13 // but WITHOUT ANY WARRANTY; without even the implied warranty of
14 // MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
15 // GNU General Public License for more details.
17 // You should have received a copy of the GNU General Public License along
18 // with this library; see the file COPYING. If not, write to the Free
19 // Software Foundation, 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301,
22 // As a special exception, you may use this file as part of a free software
23 // library without restriction. Specifically, if other files instantiate
24 // templates or use macros or inline functions from this file, or you compile
25 // this file and link it with other files to produce an executable, this
26 // file does not by itself cause the resulting executable to be covered by
27 // the GNU General Public License. This exception does not however
28 // invalidate any other reasons why the executable file might be covered by
29 // the GNU General Public License.
32 // ISO C++ 14882: 22.2.1.5 Template class codecvt
35 // Written by Benjamin Kosnik <bkoz@redhat.com>
37 /** @file ext/codecvt_specializations.h
38 * This file is a GNU extension to the Standard C++ Library.
41 #ifndef _EXT_CODECVT_SPECIALIZATIONS_H
42 #define _EXT_CODECVT_SPECIALIZATIONS_H 1
44 #include <bits/c++config.h>
48 _GLIBCXX_BEGIN_NAMESPACE(__gnu_cxx
)
50 /// Extension to use iconv for dealing with character encodings.
51 // This includes conversions and comparisons between various character
52 // sets. This object encapsulates data that may need to be shared between
53 // char_traits, codecvt and ctype.
58 // NB: A conversion descriptor subsumes and enhances the
59 // functionality of a simple state type such as mbstate_t.
60 typedef iconv_t descriptor_type
;
63 // Name of internal character set encoding.
64 std::string _M_int_enc
;
66 // Name of external character set encoding.
67 std::string _M_ext_enc
;
69 // Conversion descriptor between external encoding to internal encoding.
70 descriptor_type _M_in_desc
;
72 // Conversion descriptor between internal encoding to external encoding.
73 descriptor_type _M_out_desc
;
75 // The byte-order marker for the external encoding, if necessary.
78 // The byte-order marker for the internal encoding, if necessary.
81 // Number of external bytes needed to construct one complete
82 // character in the internal encoding.
83 // NB: -1 indicates variable, or stateful, encodings.
89 : _M_in_desc(0), _M_out_desc(0), _M_ext_bom(0), _M_int_bom(0), _M_bytes(0)
93 encoding_state(const char* __int
, const char* __ext
,
94 int __ibom
= 0, int __ebom
= 0, int __bytes
= 1)
95 : _M_int_enc(__int
), _M_ext_enc(__ext
), _M_in_desc(0), _M_out_desc(0),
96 _M_ext_bom(__ebom
), _M_int_bom(__ibom
), _M_bytes(__bytes
)
99 // 21.1.2 traits typedefs
101 // typedef STATE_T state_type
102 // requires: state_type shall meet the requirements of
103 // CopyConstructible types (20.1.3)
104 // NB: This does not preserve the actual state of the conversion
105 // descriptor member, but it does duplicate the encoding
107 encoding_state(const encoding_state
& __obj
) : _M_in_desc(0), _M_out_desc(0)
108 { construct(__obj
); }
110 // Need assignment operator as well.
112 operator=(const encoding_state
& __obj
)
124 const descriptor_type __err
= (iconv_t
)(-1);
125 bool __test
= _M_in_desc
&& _M_in_desc
!= __err
;
126 __test
&= _M_out_desc
&& _M_out_desc
!= __err
;
131 character_ratio() const
135 internal_encoding() const
136 { return _M_int_enc
; }
140 { return _M_int_bom
; }
143 external_encoding() const
144 { return _M_ext_enc
; }
148 { return _M_ext_bom
; }
150 const descriptor_type
&
151 in_descriptor() const
152 { return _M_in_desc
; }
154 const descriptor_type
&
155 out_descriptor() const
156 { return _M_out_desc
; }
162 const descriptor_type __err
= (iconv_t
)(-1);
163 const bool __have_encodings
= _M_int_enc
.size() && _M_ext_enc
.size();
164 if (!_M_in_desc
&& __have_encodings
)
166 _M_in_desc
= iconv_open(_M_int_enc
.c_str(), _M_ext_enc
.c_str());
167 if (_M_in_desc
== __err
)
168 std::__throw_runtime_error(__N("encoding_state::_M_init "
169 "creating iconv input descriptor failed"));
171 if (!_M_out_desc
&& __have_encodings
)
173 _M_out_desc
= iconv_open(_M_ext_enc
.c_str(), _M_int_enc
.c_str());
174 if (_M_out_desc
== __err
)
175 std::__throw_runtime_error(__N("encoding_state::_M_init "
176 "creating iconv output descriptor failed"));
181 construct(const encoding_state
& __obj
)
184 _M_int_enc
= __obj
._M_int_enc
;
185 _M_ext_enc
= __obj
._M_ext_enc
;
186 _M_ext_bom
= __obj
._M_ext_bom
;
187 _M_int_bom
= __obj
._M_int_bom
;
188 _M_bytes
= __obj
._M_bytes
;
195 const descriptor_type __err
= (iconv_t
)(-1);
196 if (_M_in_desc
&& _M_in_desc
!= __err
)
198 iconv_close(_M_in_desc
);
201 if (_M_out_desc
&& _M_out_desc
!= __err
)
203 iconv_close(_M_out_desc
);
209 /// encoding_char_traits
210 // Custom traits type with encoding_state for the state type, and the
211 // associated fpos<encoding_state> for the position type, all other
212 // bits equivalent to the required char_traits instantiations.
213 template<typename _CharT
>
214 struct encoding_char_traits
: public std::char_traits
<_CharT
>
216 typedef encoding_state state_type
;
217 typedef typename
std::fpos
<state_type
> pos_type
;
220 _GLIBCXX_END_NAMESPACE
223 _GLIBCXX_BEGIN_NAMESPACE(std
)
225 using __gnu_cxx::encoding_state
;
227 /// codecvt<InternT, _ExternT, encoding_state> specialization.
228 // This partial specialization takes advantage of iconv to provide
229 // code conversions between a large number of character encodings.
230 template<typename _InternT
, typename _ExternT
>
231 class codecvt
<_InternT
, _ExternT
, encoding_state
>
232 : public __codecvt_abstract_base
<_InternT
, _ExternT
, encoding_state
>
236 typedef codecvt_base::result result
;
237 typedef _InternT intern_type
;
238 typedef _ExternT extern_type
;
239 typedef __gnu_cxx::encoding_state state_type
;
240 typedef state_type::descriptor_type descriptor_type
;
243 static locale::id id
;
246 codecvt(size_t __refs
= 0)
247 : __codecvt_abstract_base
<intern_type
, extern_type
, state_type
>(__refs
)
251 codecvt(state_type
& __enc
, size_t __refs
= 0)
252 : __codecvt_abstract_base
<intern_type
, extern_type
, state_type
>(__refs
)
260 do_out(state_type
& __state
, const intern_type
* __from
,
261 const intern_type
* __from_end
, const intern_type
*& __from_next
,
262 extern_type
* __to
, extern_type
* __to_end
,
263 extern_type
*& __to_next
) const;
266 do_unshift(state_type
& __state
, extern_type
* __to
,
267 extern_type
* __to_end
, extern_type
*& __to_next
) const;
270 do_in(state_type
& __state
, const extern_type
* __from
,
271 const extern_type
* __from_end
, const extern_type
*& __from_next
,
272 intern_type
* __to
, intern_type
* __to_end
,
273 intern_type
*& __to_next
) const;
276 do_encoding() const throw();
279 do_always_noconv() const throw();
282 do_length(state_type
&, const extern_type
* __from
,
283 const extern_type
* __end
, size_t __max
) const;
286 do_max_length() const throw();
289 template<typename _InternT
, typename _ExternT
>
291 codecvt
<_InternT
, _ExternT
, encoding_state
>::id
;
293 // This adaptor works around the signature problems of the second
294 // argument to iconv(): SUSv2 and others use 'const char**', but glibc 2.2
295 // uses 'char**', which matches the POSIX 1003.1-2001 standard.
296 // Using this adaptor, g++ will do the work for us.
297 template<typename _Tp
>
299 __iconv_adaptor(size_t(*__func
)(iconv_t
, _Tp
, size_t*, char**, size_t*),
300 iconv_t __cd
, char** __inbuf
, size_t* __inbytes
,
301 char** __outbuf
, size_t* __outbytes
)
302 { return __func(__cd
, (_Tp
)__inbuf
, __inbytes
, __outbuf
, __outbytes
); }
304 template<typename _InternT
, typename _ExternT
>
306 codecvt
<_InternT
, _ExternT
, encoding_state
>::
307 do_out(state_type
& __state
, const intern_type
* __from
,
308 const intern_type
* __from_end
, const intern_type
*& __from_next
,
309 extern_type
* __to
, extern_type
* __to_end
,
310 extern_type
*& __to_next
) const
312 result __ret
= codecvt_base::error
;
315 const descriptor_type
& __desc
= __state
.out_descriptor();
316 const size_t __fmultiple
= sizeof(intern_type
);
317 size_t __fbytes
= __fmultiple
* (__from_end
- __from
);
318 const size_t __tmultiple
= sizeof(extern_type
);
319 size_t __tbytes
= __tmultiple
* (__to_end
- __to
);
321 // Argument list for iconv specifies a byte sequence. Thus,
322 // all to/from arrays must be brutally casted to char*.
323 char* __cto
= reinterpret_cast<char*>(__to
);
327 // Some encodings need a byte order marker as the first item
328 // in the byte stream, to designate endian-ness. The default
329 // value for the byte order marker is NULL, so if this is
330 // the case, it's not necessary and we can just go on our
332 int __int_bom
= __state
.internal_bom();
335 size_t __size
= __from_end
- __from
;
336 intern_type
* __cfixed
= static_cast<intern_type
*>
337 (__builtin_alloca(sizeof(intern_type
) * (__size
+ 1)));
338 __cfixed
[0] = static_cast<intern_type
>(__int_bom
);
339 char_traits
<intern_type
>::copy(__cfixed
+ 1, __from
, __size
);
340 __cfrom
= reinterpret_cast<char*>(__cfixed
);
341 __conv
= __iconv_adaptor(iconv
, __desc
, &__cfrom
,
342 &__fbytes
, &__cto
, &__tbytes
);
346 intern_type
* __cfixed
= const_cast<intern_type
*>(__from
);
347 __cfrom
= reinterpret_cast<char*>(__cfixed
);
348 __conv
= __iconv_adaptor(iconv
, __desc
, &__cfrom
, &__fbytes
,
352 if (__conv
!= size_t(-1))
354 __from_next
= reinterpret_cast<const intern_type
*>(__cfrom
);
355 __to_next
= reinterpret_cast<extern_type
*>(__cto
);
356 __ret
= codecvt_base::ok
;
360 if (__fbytes
< __fmultiple
* (__from_end
- __from
))
362 __from_next
= reinterpret_cast<const intern_type
*>(__cfrom
);
363 __to_next
= reinterpret_cast<extern_type
*>(__cto
);
364 __ret
= codecvt_base::partial
;
367 __ret
= codecvt_base::error
;
373 template<typename _InternT
, typename _ExternT
>
375 codecvt
<_InternT
, _ExternT
, encoding_state
>::
376 do_unshift(state_type
& __state
, extern_type
* __to
,
377 extern_type
* __to_end
, extern_type
*& __to_next
) const
379 result __ret
= codecvt_base::error
;
382 const descriptor_type
& __desc
= __state
.in_descriptor();
383 const size_t __tmultiple
= sizeof(intern_type
);
384 size_t __tlen
= __tmultiple
* (__to_end
- __to
);
386 // Argument list for iconv specifies a byte sequence. Thus,
387 // all to/from arrays must be brutally casted to char*.
388 char* __cto
= reinterpret_cast<char*>(__to
);
389 size_t __conv
= __iconv_adaptor(iconv
,__desc
, NULL
, NULL
,
392 if (__conv
!= size_t(-1))
394 __to_next
= reinterpret_cast<extern_type
*>(__cto
);
395 if (__tlen
== __tmultiple
* (__to_end
- __to
))
396 __ret
= codecvt_base::noconv
;
397 else if (__tlen
== 0)
398 __ret
= codecvt_base::ok
;
400 __ret
= codecvt_base::partial
;
403 __ret
= codecvt_base::error
;
408 template<typename _InternT
, typename _ExternT
>
410 codecvt
<_InternT
, _ExternT
, encoding_state
>::
411 do_in(state_type
& __state
, const extern_type
* __from
,
412 const extern_type
* __from_end
, const extern_type
*& __from_next
,
413 intern_type
* __to
, intern_type
* __to_end
,
414 intern_type
*& __to_next
) const
416 result __ret
= codecvt_base::error
;
419 const descriptor_type
& __desc
= __state
.in_descriptor();
420 const size_t __fmultiple
= sizeof(extern_type
);
421 size_t __flen
= __fmultiple
* (__from_end
- __from
);
422 const size_t __tmultiple
= sizeof(intern_type
);
423 size_t __tlen
= __tmultiple
* (__to_end
- __to
);
425 // Argument list for iconv specifies a byte sequence. Thus,
426 // all to/from arrays must be brutally casted to char*.
427 char* __cto
= reinterpret_cast<char*>(__to
);
431 // Some encodings need a byte order marker as the first item
432 // in the byte stream, to designate endian-ness. The default
433 // value for the byte order marker is NULL, so if this is
434 // the case, it's not necessary and we can just go on our
436 int __ext_bom
= __state
.external_bom();
439 size_t __size
= __from_end
- __from
;
440 extern_type
* __cfixed
= static_cast<extern_type
*>
441 (__builtin_alloca(sizeof(extern_type
) * (__size
+ 1)));
442 __cfixed
[0] = static_cast<extern_type
>(__ext_bom
);
443 char_traits
<extern_type
>::copy(__cfixed
+ 1, __from
, __size
);
444 __cfrom
= reinterpret_cast<char*>(__cfixed
);
445 __conv
= __iconv_adaptor(iconv
, __desc
, &__cfrom
,
446 &__flen
, &__cto
, &__tlen
);
450 extern_type
* __cfixed
= const_cast<extern_type
*>(__from
);
451 __cfrom
= reinterpret_cast<char*>(__cfixed
);
452 __conv
= __iconv_adaptor(iconv
, __desc
, &__cfrom
,
453 &__flen
, &__cto
, &__tlen
);
457 if (__conv
!= size_t(-1))
459 __from_next
= reinterpret_cast<const extern_type
*>(__cfrom
);
460 __to_next
= reinterpret_cast<intern_type
*>(__cto
);
461 __ret
= codecvt_base::ok
;
465 if (__flen
< static_cast<size_t>(__from_end
- __from
))
467 __from_next
= reinterpret_cast<const extern_type
*>(__cfrom
);
468 __to_next
= reinterpret_cast<intern_type
*>(__cto
);
469 __ret
= codecvt_base::partial
;
472 __ret
= codecvt_base::error
;
478 template<typename _InternT
, typename _ExternT
>
480 codecvt
<_InternT
, _ExternT
, encoding_state
>::
481 do_encoding() const throw()
484 if (sizeof(_ExternT
) <= sizeof(_InternT
))
485 __ret
= sizeof(_InternT
) / sizeof(_ExternT
);
489 template<typename _InternT
, typename _ExternT
>
491 codecvt
<_InternT
, _ExternT
, encoding_state
>::
492 do_always_noconv() const throw()
495 template<typename _InternT
, typename _ExternT
>
497 codecvt
<_InternT
, _ExternT
, encoding_state
>::
498 do_length(state_type
&, const extern_type
* __from
,
499 const extern_type
* __end
, size_t __max
) const
500 { return std::min(__max
, static_cast<size_t>(__end
- __from
)); }
502 // _GLIBCXX_RESOLVE_LIB_DEFECTS
503 // 74. Garbled text for codecvt::do_max_length
504 template<typename _InternT
, typename _ExternT
>
506 codecvt
<_InternT
, _ExternT
, encoding_state
>::
507 do_max_length() const throw()
510 _GLIBCXX_END_NAMESPACE