1 // Locale support (codecvt) -*- C++ -*-
3 // Copyright (C) 2000, 2001, 2002, 2003, 2004, 2005
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 bits/codecvt_specializations.h
38 * This file is a GNU extension to the Standard C++ Library.
42 // Define this here so codecvt.cc can have _S_max_size definition.
43 #define _GLIBCXX_USE_ENCODING_STATE 1
47 /// @brief Extension to use icov for dealing with character encodings.
48 // This includes conversions and comparisons between various character
49 // sets. This object encapsulates data that may need to be shared between
50 // char_traits, codecvt and ctype.
55 // NB: A conversion descriptor subsumes and enhances the
56 // functionality of a simple state type such as mbstate_t.
57 typedef iconv_t descriptor_type
;
60 // Name of internal character set encoding.
61 std::string _M_int_enc
;
63 // Name of external character set encoding.
64 std::string _M_ext_enc
;
66 // Conversion descriptor between external encoding to internal encoding.
67 descriptor_type _M_in_desc
;
69 // Conversion descriptor between internal encoding to external encoding.
70 descriptor_type _M_out_desc
;
72 // The byte-order marker for the external encoding, if necessary.
75 // The byte-order marker for the internal encoding, if necessary.
78 // Number of external bytes needed to construct one complete
79 // character in the internal encoding.
80 // NB: -1 indicates variable, or stateful, encodings.
86 : _M_in_desc(0), _M_out_desc(0), _M_ext_bom(0), _M_int_bom(0), _M_bytes(0)
90 encoding_state(const char* __int
, const char* __ext
,
91 int __ibom
= 0, int __ebom
= 0, int __bytes
= 1)
92 : _M_int_enc(__int
), _M_ext_enc(__ext
), _M_in_desc(0), _M_out_desc(0),
93 _M_ext_bom(__ebom
), _M_int_bom(__ibom
), _M_bytes(__bytes
)
96 // 21.1.2 traits typedefs
98 // typedef STATE_T state_type
99 // requires: state_type shall meet the requirements of
100 // CopyConstructible types (20.1.3)
101 // NB: This does not preseve the actual state of the conversion
102 // descriptor member, but it does duplicate the encoding
104 encoding_state(const encoding_state
& __obj
) : _M_in_desc(0), _M_out_desc(0)
105 { construct(__obj
); }
107 // Need assignment operator as well.
109 operator=(const encoding_state
& __obj
)
121 const descriptor_type __err
= reinterpret_cast<iconv_t
>(-1);
122 bool __test
= _M_in_desc
&& _M_in_desc
!= __err
;
123 __test
&= _M_out_desc
&& _M_out_desc
!= __err
;
128 character_ratio() const
132 internal_encoding() const
133 { return _M_int_enc
; }
137 { return _M_int_bom
; }
140 external_encoding() const
141 { return _M_ext_enc
; }
145 { return _M_ext_bom
; }
147 const descriptor_type
&
148 in_descriptor() const
149 { return _M_in_desc
; }
151 const descriptor_type
&
152 out_descriptor() const
153 { return _M_out_desc
; }
159 const descriptor_type __err
= reinterpret_cast<iconv_t
>(-1);
160 const bool __have_encodings
= _M_int_enc
.size() && _M_ext_enc
.size();
161 if (!_M_in_desc
&& __have_encodings
)
163 _M_in_desc
= iconv_open(_M_int_enc
.c_str(), _M_ext_enc
.c_str());
164 if (_M_in_desc
== __err
)
165 std::__throw_runtime_error(__N("encoding_state::_M_init "
166 "creating iconv input descriptor failed"));
168 if (!_M_out_desc
&& __have_encodings
)
170 _M_out_desc
= iconv_open(_M_ext_enc
.c_str(), _M_int_enc
.c_str());
171 if (_M_out_desc
== __err
)
172 std::__throw_runtime_error(__N("encoding_state::_M_init "
173 "creating iconv output descriptor failed"));
178 construct(const encoding_state
& __obj
)
181 _M_int_enc
= __obj
._M_int_enc
;
182 _M_ext_enc
= __obj
._M_ext_enc
;
183 _M_ext_bom
= __obj
._M_ext_bom
;
184 _M_int_bom
= __obj
._M_int_bom
;
185 _M_bytes
= __obj
._M_bytes
;
192 const descriptor_type __err
= reinterpret_cast<iconv_t
>(-1);
193 if (_M_in_desc
&& _M_in_desc
!= __err
)
195 iconv_close(_M_in_desc
);
198 if (_M_out_desc
&& _M_out_desc
!= __err
)
200 iconv_close(_M_out_desc
);
206 /// @brief encoding_char_traits.
207 // Custom traits type with encoding_state for the state type, and the
208 // associated fpos<encoding_state> for the position type, all other
209 // bits equivalent to the required char_traits instantiations.
210 template<typename _CharT
>
211 struct encoding_char_traits
: public std::char_traits
<_CharT
>
213 typedef encoding_state state_type
;
214 typedef typename
std::fpos
<state_type
> pos_type
;
216 } // namespace __gnu_cxx
220 using __gnu_cxx::encoding_state
;
222 /// @brief codecvt<InternT, _ExternT, encoding_state> specialization.
223 // This partial specialization takes advantage of iconv to provide
224 // code conversions between a large number of character encodings.
225 template<typename _InternT
, typename _ExternT
>
226 class codecvt
<_InternT
, _ExternT
, encoding_state
>
227 : public __codecvt_abstract_base
<_InternT
, _ExternT
, encoding_state
>
231 typedef codecvt_base::result result
;
232 typedef _InternT intern_type
;
233 typedef _ExternT extern_type
;
234 typedef __gnu_cxx::encoding_state state_type
;
235 typedef state_type::descriptor_type descriptor_type
;
238 static locale::id id
;
241 codecvt(size_t __refs
= 0)
242 : __codecvt_abstract_base
<intern_type
, extern_type
, state_type
>(__refs
)
246 codecvt(state_type
& __enc
, size_t __refs
= 0)
247 : __codecvt_abstract_base
<intern_type
, extern_type
, state_type
>(__refs
)
255 do_out(state_type
& __state
, const intern_type
* __from
,
256 const intern_type
* __from_end
, const intern_type
*& __from_next
,
257 extern_type
* __to
, extern_type
* __to_end
,
258 extern_type
*& __to_next
) const;
261 do_unshift(state_type
& __state
, extern_type
* __to
,
262 extern_type
* __to_end
, extern_type
*& __to_next
) const;
265 do_in(state_type
& __state
, const extern_type
* __from
,
266 const extern_type
* __from_end
, const extern_type
*& __from_next
,
267 intern_type
* __to
, intern_type
* __to_end
,
268 intern_type
*& __to_next
) const;
271 do_encoding() const throw();
274 do_always_noconv() const throw();
277 do_length(state_type
&, const extern_type
* __from
,
278 const extern_type
* __end
, size_t __max
) const;
281 do_max_length() const throw();
284 template<typename _InternT
, typename _ExternT
>
286 codecvt
<_InternT
, _ExternT
, encoding_state
>::id
;
288 // This adaptor works around the signature problems of the second
289 // argument to iconv(): SUSv2 and others use 'const char**', but glibc 2.2
290 // uses 'char**', which matches the POSIX 1003.1-2001 standard.
291 // Using this adaptor, g++ will do the work for us.
292 template<typename _T
>
294 __iconv_adaptor(size_t(*__func
)(iconv_t
, _T
, size_t*, char**, size_t*),
295 iconv_t __cd
, char** __inbuf
, size_t* __inbytes
,
296 char** __outbuf
, size_t* __outbytes
)
297 { return __func(__cd
, (_T
)__inbuf
, __inbytes
, __outbuf
, __outbytes
); }
299 template<typename _InternT
, typename _ExternT
>
301 codecvt
<_InternT
, _ExternT
, encoding_state
>::
302 do_out(state_type
& __state
, const intern_type
* __from
,
303 const intern_type
* __from_end
, const intern_type
*& __from_next
,
304 extern_type
* __to
, extern_type
* __to_end
,
305 extern_type
*& __to_next
) const
307 result __ret
= codecvt_base::error
;
310 const descriptor_type
& __desc
= __state
.out_descriptor();
311 const size_t __fmultiple
= sizeof(intern_type
);
312 size_t __fbytes
= __fmultiple
* (__from_end
- __from
);
313 const size_t __tmultiple
= sizeof(extern_type
);
314 size_t __tbytes
= __tmultiple
* (__to_end
- __to
);
316 // Argument list for iconv specifies a byte sequence. Thus,
317 // all to/from arrays must be brutally casted to char*.
318 char* __cto
= reinterpret_cast<char*>(__to
);
322 // Some encodings need a byte order marker as the first item
323 // in the byte stream, to designate endian-ness. The default
324 // value for the byte order marker is NULL, so if this is
325 // the case, it's not necessary and we can just go on our
327 int __int_bom
= __state
.internal_bom();
330 size_t __size
= __from_end
- __from
;
331 intern_type
* __cfixed
= static_cast<intern_type
*>
332 (__builtin_alloca(sizeof(intern_type
) * (__size
+ 1)));
333 __cfixed
[0] = static_cast<intern_type
>(__int_bom
);
334 char_traits
<intern_type
>::copy(__cfixed
+ 1, __from
, __size
);
335 __cfrom
= reinterpret_cast<char*>(__cfixed
);
336 __conv
= __iconv_adaptor(iconv
, __desc
, &__cfrom
,
337 &__fbytes
, &__cto
, &__tbytes
);
341 intern_type
* __cfixed
= const_cast<intern_type
*>(__from
);
342 __cfrom
= reinterpret_cast<char*>(__cfixed
);
343 __conv
= __iconv_adaptor(iconv
, __desc
, &__cfrom
, &__fbytes
,
347 if (__conv
!= size_t(-1))
349 __from_next
= reinterpret_cast<const intern_type
*>(__cfrom
);
350 __to_next
= reinterpret_cast<extern_type
*>(__cto
);
351 __ret
= codecvt_base::ok
;
355 if (__fbytes
< __fmultiple
* (__from_end
- __from
))
357 __from_next
= reinterpret_cast<const intern_type
*>(__cfrom
);
358 __to_next
= reinterpret_cast<extern_type
*>(__cto
);
359 __ret
= codecvt_base::partial
;
362 __ret
= codecvt_base::error
;
368 template<typename _InternT
, typename _ExternT
>
370 codecvt
<_InternT
, _ExternT
, encoding_state
>::
371 do_unshift(state_type
& __state
, extern_type
* __to
,
372 extern_type
* __to_end
, extern_type
*& __to_next
) const
374 result __ret
= codecvt_base::error
;
377 const descriptor_type
& __desc
= __state
.in_descriptor();
378 const size_t __tmultiple
= sizeof(intern_type
);
379 size_t __tlen
= __tmultiple
* (__to_end
- __to
);
381 // Argument list for iconv specifies a byte sequence. Thus,
382 // all to/from arrays must be brutally casted to char*.
383 char* __cto
= reinterpret_cast<char*>(__to
);
384 size_t __conv
= __iconv_adaptor(iconv
,__desc
, NULL
, NULL
,
387 if (__conv
!= size_t(-1))
389 __to_next
= reinterpret_cast<extern_type
*>(__cto
);
390 if (__tlen
== __tmultiple
* (__to_end
- __to
))
391 __ret
= codecvt_base::noconv
;
392 else if (__tlen
== 0)
393 __ret
= codecvt_base::ok
;
395 __ret
= codecvt_base::partial
;
398 __ret
= codecvt_base::error
;
403 template<typename _InternT
, typename _ExternT
>
405 codecvt
<_InternT
, _ExternT
, encoding_state
>::
406 do_in(state_type
& __state
, const extern_type
* __from
,
407 const extern_type
* __from_end
, const extern_type
*& __from_next
,
408 intern_type
* __to
, intern_type
* __to_end
,
409 intern_type
*& __to_next
) const
411 result __ret
= codecvt_base::error
;
414 const descriptor_type
& __desc
= __state
.in_descriptor();
415 const size_t __fmultiple
= sizeof(extern_type
);
416 size_t __flen
= __fmultiple
* (__from_end
- __from
);
417 const size_t __tmultiple
= sizeof(intern_type
);
418 size_t __tlen
= __tmultiple
* (__to_end
- __to
);
420 // Argument list for iconv specifies a byte sequence. Thus,
421 // all to/from arrays must be brutally casted to char*.
422 char* __cto
= reinterpret_cast<char*>(__to
);
426 // Some encodings need a byte order marker as the first item
427 // in the byte stream, to designate endian-ness. The default
428 // value for the byte order marker is NULL, so if this is
429 // the case, it's not necessary and we can just go on our
431 int __ext_bom
= __state
.external_bom();
434 size_t __size
= __from_end
- __from
;
435 extern_type
* __cfixed
= static_cast<extern_type
*>
436 (__builtin_alloca(sizeof(extern_type
) * (__size
+ 1)));
437 __cfixed
[0] = static_cast<extern_type
>(__ext_bom
);
438 char_traits
<extern_type
>::copy(__cfixed
+ 1, __from
, __size
);
439 __cfrom
= reinterpret_cast<char*>(__cfixed
);
440 __conv
= __iconv_adaptor(iconv
, __desc
, &__cfrom
,
441 &__flen
, &__cto
, &__tlen
);
445 extern_type
* __cfixed
= const_cast<extern_type
*>(__from
);
446 __cfrom
= reinterpret_cast<char*>(__cfixed
);
447 __conv
= __iconv_adaptor(iconv
, __desc
, &__cfrom
,
448 &__flen
, &__cto
, &__tlen
);
452 if (__conv
!= size_t(-1))
454 __from_next
= reinterpret_cast<const extern_type
*>(__cfrom
);
455 __to_next
= reinterpret_cast<intern_type
*>(__cto
);
456 __ret
= codecvt_base::ok
;
460 if (__flen
< static_cast<size_t>(__from_end
- __from
))
462 __from_next
= reinterpret_cast<const extern_type
*>(__cfrom
);
463 __to_next
= reinterpret_cast<intern_type
*>(__cto
);
464 __ret
= codecvt_base::partial
;
467 __ret
= codecvt_base::error
;
473 template<typename _InternT
, typename _ExternT
>
475 codecvt
<_InternT
, _ExternT
, encoding_state
>::
476 do_encoding() const throw()
479 if (sizeof(_ExternT
) <= sizeof(_InternT
))
480 __ret
= sizeof(_InternT
) / sizeof(_ExternT
);
484 template<typename _InternT
, typename _ExternT
>
486 codecvt
<_InternT
, _ExternT
, encoding_state
>::
487 do_always_noconv() const throw()
490 template<typename _InternT
, typename _ExternT
>
492 codecvt
<_InternT
, _ExternT
, encoding_state
>::
493 do_length(state_type
&, const extern_type
* __from
,
494 const extern_type
* __end
, size_t __max
) const
495 { return std::min(__max
, static_cast<size_t>(__end
- __from
)); }
497 // _GLIBCXX_RESOLVE_LIB_DEFECTS
498 // 74. Garbled text for codecvt::do_max_length
499 template<typename _InternT
, typename _ExternT
>
501 codecvt
<_InternT
, _ExternT
, encoding_state
>::
502 do_max_length() const throw()