1 // Locale support (codecvt) -*- C++ -*-
2
3 // Copyright (C) 2000-2022 Free Software Foundation, Inc.
4 //
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)
9 // any later version.
10
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.
15
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.
19
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/>.
24
25 //
26 // ISO C++ 14882: 22.2.1.5 Template class codecvt
27 //
28
29 // Written by Benjamin Kosnik <bkoz@redhat.com>
30
31 /** @file ext/codecvt_specializations.h
32 * This file is a GNU extension to the Standard C++ Library.
33 */
34
35 #ifndef _EXT_CODECVT_SPECIALIZATIONS_H
36 #define _EXT_CODECVT_SPECIALIZATIONS_H 1
37
38 #include <bits/c++config.h>
39 #include <locale>
40 #include <iconv.h>
41
_GLIBCXX_VISIBILITY(default)42 namespace __gnu_cxx _GLIBCXX_VISIBILITY(default)
43 {
44 _GLIBCXX_BEGIN_NAMESPACE_VERSION
45 _GLIBCXX_BEGIN_NAMESPACE_CXX11
46
47 /// Extension to use iconv 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.
51 class encoding_state
52 {
53 public:
54 // Types:
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;
58
59 protected:
60 // Name of internal character set encoding.
61 std::string _M_int_enc;
62
63 // Name of external character set encoding.
64 std::string _M_ext_enc;
65
66 // Conversion descriptor between external encoding to internal encoding.
67 descriptor_type _M_in_desc;
68
69 // Conversion descriptor between internal encoding to external encoding.
70 descriptor_type _M_out_desc;
71
72 // The byte-order marker for the external encoding, if necessary.
73 int _M_ext_bom;
74
75 // The byte-order marker for the internal encoding, if necessary.
76 int _M_int_bom;
77
78 // Number of external bytes needed to construct one complete
79 // character in the internal encoding.
80 // NB: -1 indicates variable, or stateful, encodings.
81 int _M_bytes;
82
83 public:
84 explicit
85 encoding_state()
86 : _M_in_desc(0), _M_out_desc(0), _M_ext_bom(0), _M_int_bom(0), _M_bytes(0)
87 { }
88
89 explicit
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)
94 { init(); }
95
96 // 21.1.2 traits typedefs
97 // p4
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 preserve the actual state of the conversion
102 // descriptor member, but it does duplicate the encoding
103 // information.
104 encoding_state(const encoding_state& __obj) : _M_in_desc(0), _M_out_desc(0)
105 { construct(__obj); }
106
107 // Need assignment operator as well.
108 encoding_state&
109 operator=(const encoding_state& __obj)
110 {
111 construct(__obj);
112 return *this;
113 }
114
115 ~encoding_state()
116 { destroy(); }
117
118 bool
119 good() const throw()
120 {
121 const descriptor_type __err = (iconv_t)(-1);
122 bool __test = _M_in_desc && _M_in_desc != __err;
123 __test &= _M_out_desc && _M_out_desc != __err;
124 return __test;
125 }
126
127 int
128 character_ratio() const
129 { return _M_bytes; }
130
131 const std::string
132 internal_encoding() const
133 { return _M_int_enc; }
134
135 int
136 internal_bom() const
137 { return _M_int_bom; }
138
139 const std::string
140 external_encoding() const
141 { return _M_ext_enc; }
142
143 int
144 external_bom() const
145 { return _M_ext_bom; }
146
147 const descriptor_type&
148 in_descriptor() const
149 { return _M_in_desc; }
150
151 const descriptor_type&
152 out_descriptor() const
153 { return _M_out_desc; }
154
155 protected:
156 void
157 init()
158 {
159 const descriptor_type __err = (iconv_t)(-1);
160 const bool __have_encodings = _M_int_enc.size() && _M_ext_enc.size();
161 if (!_M_in_desc && __have_encodings)
162 {
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"));
167 }
168 if (!_M_out_desc && __have_encodings)
169 {
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"));
174 }
175 }
176
177 void
178 construct(const encoding_state& __obj)
179 {
180 destroy();
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;
186 init();
187 }
188
189 void
190 destroy() throw()
191 {
192 const descriptor_type __err = (iconv_t)(-1);
193 if (_M_in_desc && _M_in_desc != __err)
194 {
195 iconv_close(_M_in_desc);
196 _M_in_desc = 0;
197 }
198 if (_M_out_desc && _M_out_desc != __err)
199 {
200 iconv_close(_M_out_desc);
201 _M_out_desc = 0;
202 }
203 }
204 };
205
206 /// 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
212 : public std::char_traits<_CharT>
213 {
214 typedef encoding_state state_type;
215 typedef typename std::fpos<state_type> pos_type;
216 };
217
218 _GLIBCXX_END_NAMESPACE_CXX11
219 _GLIBCXX_END_NAMESPACE_VERSION
220 } // namespace
221
222
_GLIBCXX_VISIBILITY(default)223 namespace std _GLIBCXX_VISIBILITY(default)
224 {
225 _GLIBCXX_BEGIN_NAMESPACE_VERSION
226
227 using __gnu_cxx::encoding_state;
228
229 /// codecvt<InternT, _ExternT, encoding_state> specialization.
230 // This partial specialization takes advantage of iconv to provide
231 // code conversions between a large number of character encodings.
232 template<typename _InternT, typename _ExternT>
233 class codecvt<_InternT, _ExternT, encoding_state>
234 : public __codecvt_abstract_base<_InternT, _ExternT, encoding_state>
235 {
236 public:
237 // Types:
238 typedef codecvt_base::result result;
239 typedef _InternT intern_type;
240 typedef _ExternT extern_type;
241 typedef __gnu_cxx::encoding_state state_type;
242 typedef state_type::descriptor_type descriptor_type;
243
244 // Data Members:
245 static locale::id id;
246
247 explicit
248 codecvt(size_t __refs = 0)
249 : __codecvt_abstract_base<intern_type, extern_type, state_type>(__refs)
250 { }
251
252 explicit
253 codecvt(state_type& __enc, size_t __refs = 0)
254 : __codecvt_abstract_base<intern_type, extern_type, state_type>(__refs)
255 { }
256
257 protected:
258 virtual
259 ~codecvt() { }
260
261 virtual result
262 do_out(state_type& __state, const intern_type* __from,
263 const intern_type* __from_end, const intern_type*& __from_next,
264 extern_type* __to, extern_type* __to_end,
265 extern_type*& __to_next) const;
266
267 virtual result
268 do_unshift(state_type& __state, extern_type* __to,
269 extern_type* __to_end, extern_type*& __to_next) const;
270
271 virtual result
272 do_in(state_type& __state, const extern_type* __from,
273 const extern_type* __from_end, const extern_type*& __from_next,
274 intern_type* __to, intern_type* __to_end,
275 intern_type*& __to_next) const;
276
277 virtual int
278 do_encoding() const throw();
279
280 virtual bool
281 do_always_noconv() const throw();
282
283 virtual int
284 do_length(state_type&, const extern_type* __from,
285 const extern_type* __end, size_t __max) const;
286
287 virtual int
288 do_max_length() const throw();
289 };
290
291 template<typename _InternT, typename _ExternT>
292 locale::id
293 codecvt<_InternT, _ExternT, encoding_state>::id;
294
295 // This adaptor works around the signature problems of the second
296 // argument to iconv(): SUSv2 and others use 'const char**', but glibc 2.2
297 // uses 'char**', which matches the POSIX 1003.1-2001 standard.
298 // Using this adaptor, g++ will do the work for us.
299 template<typename _Tp>
300 inline size_t
301 __iconv_adaptor(size_t(*__func)(iconv_t, _Tp, size_t*, char**, size_t*),
302 iconv_t __cd, char** __inbuf, size_t* __inbytes,
303 char** __outbuf, size_t* __outbytes)
304 { return __func(__cd, (_Tp)__inbuf, __inbytes, __outbuf, __outbytes); }
305
306 template<typename _InternT, typename _ExternT>
307 codecvt_base::result
308 codecvt<_InternT, _ExternT, encoding_state>::
309 do_out(state_type& __state, const intern_type* __from,
310 const intern_type* __from_end, const intern_type*& __from_next,
311 extern_type* __to, extern_type* __to_end,
312 extern_type*& __to_next) const
313 {
314 result __ret = codecvt_base::error;
315 if (__state.good())
316 {
317 const descriptor_type& __desc = __state.out_descriptor();
318 const size_t __fmultiple = sizeof(intern_type);
319 size_t __fbytes = __fmultiple * (__from_end - __from);
320 const size_t __tmultiple = sizeof(extern_type);
321 size_t __tbytes = __tmultiple * (__to_end - __to);
322
323 // Argument list for iconv specifies a byte sequence. Thus,
324 // all to/from arrays must be brutally casted to char*.
325 char* __cto = reinterpret_cast<char*>(__to);
326 char* __cfrom;
327 size_t __conv;
328
329 // Some encodings need a byte order marker as the first item
330 // in the byte stream, to designate endian-ness. The default
331 // value for the byte order marker is NULL, so if this is
332 // the case, it's not necessary and we can just go on our
333 // merry way.
334 int __int_bom = __state.internal_bom();
335 if (__int_bom)
336 {
337 size_t __size = __from_end - __from;
338 intern_type* __cfixed = static_cast<intern_type*>
339 (__builtin_alloca(sizeof(intern_type) * (__size + 1)));
340 __cfixed[0] = static_cast<intern_type>(__int_bom);
341 char_traits<intern_type>::copy(__cfixed + 1, __from, __size);
342 __cfrom = reinterpret_cast<char*>(__cfixed);
343 __conv = __iconv_adaptor(iconv, __desc, &__cfrom,
344 &__fbytes, &__cto, &__tbytes);
345 }
346 else
347 {
348 intern_type* __cfixed = const_cast<intern_type*>(__from);
349 __cfrom = reinterpret_cast<char*>(__cfixed);
350 __conv = __iconv_adaptor(iconv, __desc, &__cfrom, &__fbytes,
351 &__cto, &__tbytes);
352 }
353
354 if (__conv != size_t(-1))
355 {
356 __from_next = reinterpret_cast<const intern_type*>(__cfrom);
357 __to_next = reinterpret_cast<extern_type*>(__cto);
358 __ret = codecvt_base::ok;
359 }
360 else
361 {
362 if (__fbytes < __fmultiple * (__from_end - __from))
363 {
364 __from_next = reinterpret_cast<const intern_type*>(__cfrom);
365 __to_next = reinterpret_cast<extern_type*>(__cto);
366 __ret = codecvt_base::partial;
367 }
368 else
369 __ret = codecvt_base::error;
370 }
371 }
372 return __ret;
373 }
374
375 template<typename _InternT, typename _ExternT>
376 codecvt_base::result
377 codecvt<_InternT, _ExternT, encoding_state>::
378 do_unshift(state_type& __state, extern_type* __to,
379 extern_type* __to_end, extern_type*& __to_next) const
380 {
381 result __ret = codecvt_base::error;
382 if (__state.good())
383 {
384 const descriptor_type& __desc = __state.in_descriptor();
385 const size_t __tmultiple = sizeof(intern_type);
386 size_t __tlen = __tmultiple * (__to_end - __to);
387
388 // Argument list for iconv specifies a byte sequence. Thus,
389 // all to/from arrays must be brutally casted to char*.
390 char* __cto = reinterpret_cast<char*>(__to);
391 size_t __conv = __iconv_adaptor(iconv,__desc, 0, 0,
392 &__cto, &__tlen);
393
394 if (__conv != size_t(-1))
395 {
396 __to_next = reinterpret_cast<extern_type*>(__cto);
397 if (__tlen == __tmultiple * (__to_end - __to))
398 __ret = codecvt_base::noconv;
399 else if (__tlen == 0)
400 __ret = codecvt_base::ok;
401 else
402 __ret = codecvt_base::partial;
403 }
404 else
405 __ret = codecvt_base::error;
406 }
407 return __ret;
408 }
409
410 template<typename _InternT, typename _ExternT>
411 codecvt_base::result
412 codecvt<_InternT, _ExternT, encoding_state>::
413 do_in(state_type& __state, const extern_type* __from,
414 const extern_type* __from_end, const extern_type*& __from_next,
415 intern_type* __to, intern_type* __to_end,
416 intern_type*& __to_next) const
417 {
418 result __ret = codecvt_base::error;
419 if (__state.good())
420 {
421 const descriptor_type& __desc = __state.in_descriptor();
422 const size_t __fmultiple = sizeof(extern_type);
423 size_t __flen = __fmultiple * (__from_end - __from);
424 const size_t __tmultiple = sizeof(intern_type);
425 size_t __tlen = __tmultiple * (__to_end - __to);
426
427 // Argument list for iconv specifies a byte sequence. Thus,
428 // all to/from arrays must be brutally casted to char*.
429 char* __cto = reinterpret_cast<char*>(__to);
430 char* __cfrom;
431 size_t __conv;
432
433 // Some encodings need a byte order marker as the first item
434 // in the byte stream, to designate endian-ness. The default
435 // value for the byte order marker is NULL, so if this is
436 // the case, it's not necessary and we can just go on our
437 // merry way.
438 int __ext_bom = __state.external_bom();
439 if (__ext_bom)
440 {
441 size_t __size = __from_end - __from;
442 extern_type* __cfixed = static_cast<extern_type*>
443 (__builtin_alloca(sizeof(extern_type) * (__size + 1)));
444 __cfixed[0] = static_cast<extern_type>(__ext_bom);
445 char_traits<extern_type>::copy(__cfixed + 1, __from, __size);
446 __cfrom = reinterpret_cast<char*>(__cfixed);
447 __conv = __iconv_adaptor(iconv, __desc, &__cfrom,
448 &__flen, &__cto, &__tlen);
449 }
450 else
451 {
452 extern_type* __cfixed = const_cast<extern_type*>(__from);
453 __cfrom = reinterpret_cast<char*>(__cfixed);
454 __conv = __iconv_adaptor(iconv, __desc, &__cfrom,
455 &__flen, &__cto, &__tlen);
456 }
457
458
459 if (__conv != size_t(-1))
460 {
461 __from_next = reinterpret_cast<const extern_type*>(__cfrom);
462 __to_next = reinterpret_cast<intern_type*>(__cto);
463 __ret = codecvt_base::ok;
464 }
465 else
466 {
467 if (__flen < static_cast<size_t>(__from_end - __from))
468 {
469 __from_next = reinterpret_cast<const extern_type*>(__cfrom);
470 __to_next = reinterpret_cast<intern_type*>(__cto);
471 __ret = codecvt_base::partial;
472 }
473 else
474 __ret = codecvt_base::error;
475 }
476 }
477 return __ret;
478 }
479
480 template<typename _InternT, typename _ExternT>
481 int
482 codecvt<_InternT, _ExternT, encoding_state>::
483 do_encoding() const throw()
484 {
485 int __ret = 0;
486 if (sizeof(_ExternT) <= sizeof(_InternT))
487 __ret = sizeof(_InternT) / sizeof(_ExternT);
488 return __ret;
489 }
490
491 template<typename _InternT, typename _ExternT>
492 bool
493 codecvt<_InternT, _ExternT, encoding_state>::
494 do_always_noconv() const throw()
495 { return false; }
496
497 template<typename _InternT, typename _ExternT>
498 int
499 codecvt<_InternT, _ExternT, encoding_state>::
500 do_length(state_type&, const extern_type* __from,
501 const extern_type* __end, size_t __max) const
502 { return std::min(__max, static_cast<size_t>(__end - __from)); }
503
504 // _GLIBCXX_RESOLVE_LIB_DEFECTS
505 // 74. Garbled text for codecvt::do_max_length
506 template<typename _InternT, typename _ExternT>
507 int
508 codecvt<_InternT, _ExternT, encoding_state>::
509 do_max_length() const throw()
510 { return 1; }
511
512 _GLIBCXX_END_NAMESPACE_VERSION
513 } // namespace
514
515 #endif
516