libstdc++-v3/include/bits/codecvt.h

   1 // Locale support (codecvt) -*- C++ -*-
   2
   3 // Copyright (C) 2000, 2001, 2002, 2003, 2004, 2005
   4 //  Free Software Foundation, Inc.
   5 //
   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)
  10 // any later version.
  11
  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.
  16
  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,
  20 // USA.
  21
  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.
  30
  31 /** @file bits/codecvt.h
  32  *  This is an internal header file, included by other library headers.
  33  *  You should not attempt to use it directly.
  34  */
  35
  36 //
  37 // ISO C++ 14882: 22.2.1.5 Template class codecvt
  38 //
  39
  40 // Written by Benjamin Kosnik <bkoz@redhat.com>
  41
  42 #ifndef _CODECVT_H
  43 #define _CODECVT_H 1
  44
  45 #pragma GCC system_header
  46
  47 _GLIBCXX_BEGIN_NAMESPACE(std)
  48
  49   /// @brief  Empty base class for codecvt facet [22.2.1.5].
  50   class codecvt_base
  51   {
  52   public:
  53     enum result
  54     {
  55       ok,
  56       partial,
  57       error,
  58       noconv
  59     };
  60   };
  61
  62   /**
  63    *  @brief  Common base for codecvt functions.
  64    *
  65    *  This template class provides implementations of the public functions
  66    *  that forward to the protected virtual functions.
  67    *
  68    *  This template also provides abstract stubs for the protected virtual
  69    *  functions.
  70   */
  71   template<typename _InternT, typename _ExternT, typename _StateT>
  72     class __codecvt_abstract_base
  73     : public locale::facet, public codecvt_base
  74     {
  75     public:
  76       // Types:
  77       typedef codecvt_base::result      result;
  78       typedef _InternT                  intern_type;
  79       typedef _ExternT                  extern_type;
  80       typedef _StateT                   state_type;
  81
  82       // 22.2.1.5.1 codecvt members
  83       /**
  84        *  @brief  Convert from internal to external character set.
  85        *
  86        *  Converts input string of intern_type to output string of
  87        *  extern_type.  This is analogous to wcsrtombs.  It does this by
  88        *  calling codecvt::do_out.
  89        *
  90        *  The source and destination character sets are determined by the
  91        *  facet's locale, internal and external types.
  92        *
  93        *  The characters in [from,from_end) are converted and written to
  94        *  [to,to_end).  from_next and to_next are set to point to the
  95        *  character following the last successfully converted character,
  96        *  respectively.  If the result needed no conversion, from_next and
  97        *  to_next are not affected.
  98        *
  99        *  The @a state argument should be intialized if the input is at the
 100        *  beginning and carried from a previous call if continuing
 101        *  conversion.  There are no guarantees about how @a state is used.
 102        *
 103        *  The result returned is a member of codecvt_base::result.  If
 104        *  all the input is converted, returns codecvt_base::ok.  If no
 105        *  conversion is necessary, returns codecvt_base::noconv.  If
 106        *  the input ends early or there is insufficient space in the
 107        *  output, returns codecvt_base::partial.  Otherwise the
 108        *  conversion failed and codecvt_base::error is returned.
 109        *
 110        *  @param  state  Persistent conversion state data.
 111        *  @param  from  Start of input.
 112        *  @param  from_end  End of input.
 113        *  @param  from_next  Returns start of unconverted data.
 114        *  @param  to  Start of output buffer.
 115        *  @param  to_end  End of output buffer.
 116        *  @param  to_next  Returns start of unused output area.
 117        *  @return  codecvt_base::result.
 118       */
 119       result
 120       out(state_type& __state, const intern_type* __from,
 121           const intern_type* __from_end, const intern_type*& __from_next,
 122           extern_type* __to, extern_type* __to_end,
 123           extern_type*& __to_next) const
 124       {
 125         return this->do_out(__state, __from, __from_end, __from_next,
 126                             __to, __to_end, __to_next);
 127       }
 128
 129       /**
 130        *  @brief  Reset conversion state.
 131        *
 132        *  Writes characters to output that would restore @a state to initial
 133        *  conditions.  The idea is that if a partial conversion occurs, then
 134        *  the converting the characters written by this function would leave
 135        *  the state in initial conditions, rather than partial conversion
 136        *  state.  It does this by calling codecvt::do_unshift().
 137        *
 138        *  For example, if 4 external characters always converted to 1 internal
 139        *  character, and input to in() had 6 external characters with state
 140        *  saved, this function would write two characters to the output and
 141        *  set the state to initialized conditions.
 142        *
 143        *  The source and destination character sets are determined by the
 144        *  facet's locale, internal and external types.
 145        *
 146        *  The result returned is a member of codecvt_base::result.  If the
 147        *  state could be reset and data written, returns codecvt_base::ok.  If
 148        *  no conversion is necessary, returns codecvt_base::noconv.  If the
 149        *  output has insufficient space, returns codecvt_base::partial.
 150        *  Otherwise the reset failed and codecvt_base::error is returned.
 151        *
 152        *  @param  state  Persistent conversion state data.
 153        *  @param  to  Start of output buffer.
 154        *  @param  to_end  End of output buffer.
 155        *  @param  to_next  Returns start of unused output area.
 156        *  @return  codecvt_base::result.
 157       */
 158       result
 159       unshift(state_type& __state, extern_type* __to, extern_type* __to_end,
 160               extern_type*& __to_next) const
 161       { return this->do_unshift(__state, __to,__to_end,__to_next); }
 162
 163       /**
 164        *  @brief  Convert from external to internal character set.
 165        *
 166        *  Converts input string of extern_type to output string of
 167        *  intern_type.  This is analogous to mbsrtowcs.  It does this by
 168        *  calling codecvt::do_in.
 169        *
 170        *  The source and destination character sets are determined by the
 171        *  facet's locale, internal and external types.
 172        *
 173        *  The characters in [from,from_end) are converted and written to
 174        *  [to,to_end).  from_next and to_next are set to point to the
 175        *  character following the last successfully converted character,
 176        *  respectively.  If the result needed no conversion, from_next and
 177        *  to_next are not affected.
 178        *
 179        *  The @a state argument should be intialized if the input is at the
 180        *  beginning and carried from a previous call if continuing
 181        *  conversion.  There are no guarantees about how @a state is used.
 182        *
 183        *  The result returned is a member of codecvt_base::result.  If
 184        *  all the input is converted, returns codecvt_base::ok.  If no
 185        *  conversion is necessary, returns codecvt_base::noconv.  If
 186        *  the input ends early or there is insufficient space in the
 187        *  output, returns codecvt_base::partial.  Otherwise the
 188        *  conversion failed and codecvt_base::error is returned.
 189        *
 190        *  @param  state  Persistent conversion state data.
 191        *  @param  from  Start of input.
 192        *  @param  from_end  End of input.
 193        *  @param  from_next  Returns start of unconverted data.
 194        *  @param  to  Start of output buffer.
 195        *  @param  to_end  End of output buffer.
 196        *  @param  to_next  Returns start of unused output area.
 197        *  @return  codecvt_base::result.
 198       */
 199       result
 200       in(state_type& __state, const extern_type* __from,
 201          const extern_type* __from_end, const extern_type*& __from_next,
 202          intern_type* __to, intern_type* __to_end,
 203          intern_type*& __to_next) const
 204       {
 205         return this->do_in(__state, __from, __from_end, __from_next,
 206                            __to, __to_end, __to_next);
 207       }
 208
 209       int
 210       encoding() const throw()
 211       { return this->do_encoding(); }
 212
 213       bool
 214       always_noconv() const throw()
 215       { return this->do_always_noconv(); }
 216
 217       int
 218       length(state_type& __state, const extern_type* __from,
 219              const extern_type* __end, size_t __max) const
 220       { return this->do_length(__state, __from, __end, __max); }
 221
 222       int
 223       max_length() const throw()
 224       { return this->do_max_length(); }
 225
 226     protected:
 227       explicit
 228       __codecvt_abstract_base(size_t __refs = 0) : locale::facet(__refs) { }
 229
 230       virtual
 231       ~__codecvt_abstract_base() { }
 232
 233       /**
 234        *  @brief  Convert from internal to external character set.
 235        *
 236        *  Converts input string of intern_type to output string of
 237        *  extern_type.  This function is a hook for derived classes to change
 238        *  the value returned.  @see out for more information.
 239       */
 240       virtual result
 241       do_out(state_type& __state, const intern_type* __from,
 242              const intern_type* __from_end, const intern_type*& __from_next,
 243              extern_type* __to, extern_type* __to_end,
 244              extern_type*& __to_next) const = 0;
 245
 246       virtual result
 247       do_unshift(state_type& __state, extern_type* __to,
 248                  extern_type* __to_end, extern_type*& __to_next) const = 0;
 249
 250       virtual result
 251       do_in(state_type& __state, const extern_type* __from,
 252             const extern_type* __from_end, const extern_type*& __from_next,
 253             intern_type* __to, intern_type* __to_end,
 254             intern_type*& __to_next) const = 0;
 255
 256       virtual int
 257       do_encoding() const throw() = 0;
 258
 259       virtual bool
 260       do_always_noconv() const throw() = 0;
 261
 262       virtual int
 263       do_length(state_type&, const extern_type* __from,
 264                 const extern_type* __end, size_t __max) const = 0;
 265
 266       virtual int
 267       do_max_length() const throw() = 0;
 268     };
 269
 270   /// @brief class codecvt [22.2.1.5].
 271   /// NB: Generic, mostly useless implementation.
 272   template<typename _InternT, typename _ExternT, typename _StateT>
 273     class codecvt
 274     : public __codecvt_abstract_base<_InternT, _ExternT, _StateT>
 275     {
 276     public:
 277       // Types:
 278       typedef codecvt_base::result      result;
 279       typedef _InternT                  intern_type;
 280       typedef _ExternT                  extern_type;
 281       typedef _StateT                   state_type;
 282
 283     protected:
 284       __c_locale                        _M_c_locale_codecvt;
 285
 286     public:
 287       static locale::id                 id;
 288
 289       explicit
 290       codecvt(size_t __refs = 0)
 291       : __codecvt_abstract_base<_InternT, _ExternT, _StateT> (__refs) { }
 292
 293       explicit
 294       codecvt(__c_locale __cloc, size_t __refs = 0);
 295
 296     protected:
 297       virtual
 298       ~codecvt() { }
 299
 300       virtual result
 301       do_out(state_type& __state, const intern_type* __from,
 302              const intern_type* __from_end, const intern_type*& __from_next,
 303              extern_type* __to, extern_type* __to_end,
 304              extern_type*& __to_next) const;
 305
 306       virtual result
 307       do_unshift(state_type& __state, extern_type* __to,
 308                  extern_type* __to_end, extern_type*& __to_next) const;
 309
 310       virtual result
 311       do_in(state_type& __state, const extern_type* __from,
 312             const extern_type* __from_end, const extern_type*& __from_next,
 313             intern_type* __to, intern_type* __to_end,
 314             intern_type*& __to_next) const;
 315
 316       virtual int
 317       do_encoding() const throw();
 318
 319       virtual bool
 320       do_always_noconv() const throw();
 321
 322       virtual int
 323       do_length(state_type&, const extern_type* __from,
 324                 const extern_type* __end, size_t __max) const;
 325
 326       virtual int
 327       do_max_length() const throw();
 328     };
 329
 330   template<typename _InternT, typename _ExternT, typename _StateT>
 331     locale::id codecvt<_InternT, _ExternT, _StateT>::id;
 332
 333   /// @brief class codecvt<char, char, mbstate_t> specialization.
 334   template<>
 335     class codecvt<char, char, mbstate_t>
 336     : public __codecvt_abstract_base<char, char, mbstate_t>
 337     {
 338     public:
 339       // Types:
 340       typedef char                      intern_type;
 341       typedef char                      extern_type;
 342       typedef mbstate_t                 state_type;
 343
 344     protected:
 345       __c_locale                        _M_c_locale_codecvt;
 346
 347     public:
 348       static locale::id id;
 349
 350       explicit
 351       codecvt(size_t __refs = 0);
 352
 353       explicit
 354       codecvt(__c_locale __cloc, size_t __refs = 0);
 355
 356     protected:
 357       virtual
 358       ~codecvt();
 359
 360       virtual result
 361       do_out(state_type& __state, const intern_type* __from,
 362              const intern_type* __from_end, const intern_type*& __from_next,
 363              extern_type* __to, extern_type* __to_end,
 364              extern_type*& __to_next) const;
 365
 366       virtual result
 367       do_unshift(state_type& __state, extern_type* __to,
 368                  extern_type* __to_end, extern_type*& __to_next) const;
 369
 370       virtual result
 371       do_in(state_type& __state, const extern_type* __from,
 372             const extern_type* __from_end, const extern_type*& __from_next,
 373             intern_type* __to, intern_type* __to_end,
 374             intern_type*& __to_next) const;
 375
 376       virtual int
 377       do_encoding() const throw();
 378
 379       virtual bool
 380       do_always_noconv() const throw();
 381
 382       virtual int
 383       do_length(state_type&, const extern_type* __from,
 384                 const extern_type* __end, size_t __max) const;
 385
 386       virtual int
 387       do_max_length() const throw();
 388   };
 389
 390 #ifdef _GLIBCXX_USE_WCHAR_T
 391   /// @brief  class codecvt<wchar_t, char, mbstate_t> specialization.
 392   template<>
 393     class codecvt<wchar_t, char, mbstate_t>
 394     : public __codecvt_abstract_base<wchar_t, char, mbstate_t>
 395     {
 396     public:
 397       // Types:
 398       typedef wchar_t                   intern_type;
 399       typedef char                      extern_type;
 400       typedef mbstate_t                 state_type;
 401
 402     protected:
 403       __c_locale                        _M_c_locale_codecvt;
 404
 405     public:
 406       static locale::id                 id;
 407
 408       explicit
 409       codecvt(size_t __refs = 0);
 410
 411       explicit
 412       codecvt(__c_locale __cloc, size_t __refs = 0);
 413
 414     protected:
 415       virtual
 416       ~codecvt();
 417
 418       virtual result
 419       do_out(state_type& __state, const intern_type* __from,
 420              const intern_type* __from_end, const intern_type*& __from_next,
 421              extern_type* __to, extern_type* __to_end,
 422              extern_type*& __to_next) const;
 423
 424       virtual result
 425       do_unshift(state_type& __state,
 426                  extern_type* __to, extern_type* __to_end,
 427                  extern_type*& __to_next) const;
 428
 429       virtual result
 430       do_in(state_type& __state,
 431              const extern_type* __from, const extern_type* __from_end,
 432              const extern_type*& __from_next,
 433              intern_type* __to, intern_type* __to_end,
 434              intern_type*& __to_next) const;
 435
 436       virtual
 437       int do_encoding() const throw();
 438
 439       virtual
 440       bool do_always_noconv() const throw();
 441
 442       virtual
 443       int do_length(state_type&, const extern_type* __from,
 444                     const extern_type* __end, size_t __max) const;
 445
 446       virtual int
 447       do_max_length() const throw();
 448     };
 449 #endif //_GLIBCXX_USE_WCHAR_T
 450
 451   /// @brief class codecvt_byname [22.2.1.6].
 452   template<typename _InternT, typename _ExternT, typename _StateT>
 453     class codecvt_byname : public codecvt<_InternT, _ExternT, _StateT>
 454     {
 455     public:
 456       explicit
 457       codecvt_byname(const char* __s, size_t __refs = 0)
 458       : codecvt<_InternT, _ExternT, _StateT>(__refs)
 459       {
 460         if (std::strcmp(__s, "C") != 0 && std::strcmp(__s, "POSIX") != 0)
 461           {
 462             this->_S_destroy_c_locale(this->_M_c_locale_codecvt);
 463             this->_S_create_c_locale(this->_M_c_locale_codecvt, __s);
 464           }
 465       }
 466
 467     protected:
 468       virtual
 469       ~codecvt_byname() { }
 470     };
 471
 472 _GLIBCXX_END_NAMESPACE
 473
 474 #endif // _CODECVT_H