include/wx/strconv.h

   1 ///////////////////////////////////////////////////////////////////////////////
   2 // Name:        strconv.h
   3 // Purpose:     conversion routines for char sets any Unicode
   4 // Author:      Robert Roebling, Ove Kaaven
   5 // Modified by:
   6 // Created:     29/01/98
   7 // RCS-ID:      $Id$
   8 // Copyright:   (c) 1998 Ove Kaaven, Robert Roebling, Vadim Zeitlin
   9 // Licence:     wxWindows licence
  10 ///////////////////////////////////////////////////////////////////////////////
  11
  12 #ifndef _WX_STRCONV_H_
  13 #define _WX_STRCONV_H_
  14
  15 #include "wx/defs.h"
  16 #include "wx/wxchar.h"
  17 #include "wx/buffer.h"
  18
  19 #ifdef __DIGITALMARS__
  20 #include "typeinfo.h"
  21 #endif
  22
  23 #if defined(__VISAGECPP__) && __IBMCPP__ >= 400
  24 #  undef __BSEXCPT__
  25 #endif
  26
  27 #include <stdlib.h>
  28
  29 #if wxUSE_WCHAR_T
  30
  31 // the error value returned by wxMBConv methods
  32 #define wxCONV_FAILED ((size_t)-1)
  33
  34 // ----------------------------------------------------------------------------
  35 // wxMBConv (abstract base class for conversions)
  36 // ----------------------------------------------------------------------------
  37
  38 // When deriving a new class from wxMBConv you must reimplement ToWChar() and
  39 // FromWChar() methods which are not pure virtual only for historical reasons,
  40 // don't let the fact that the existing classes implement MB2WC/WC2MB() instead
  41 // confuse you.
  42 //
  43 // You also have to implement Clone() to allow copying the conversions
  44 // polymorphically.
  45 //
  46 // And you might need to override GetMBNulLen() as well.
  47 class WXDLLIMPEXP_BASE wxMBConv
  48 {
  49 public:
  50     // The functions doing actual conversion from/to narrow to/from wide
  51     // character strings.
  52     //
  53     // On success, the return value is the length (i.e. the number of
  54     // characters, not bytes) of the converted string including any trailing
  55     // L'\0' or (possibly multiple) '\0'(s). If the conversion fails or if
  56     // there is not enough space for everything, including the trailing NUL
  57     // character(s), in the output buffer, (size_t)-1 is returned.
  58     //
  59     // In the special case when dstLen is 0 (outputBuf may be NULL then) the
  60     // return value is the length of the needed buffer but nothing happens
  61     // otherwise. If srcLen is -1, the entire string, up to and including the
  62     // trailing NUL(s), is converted, otherwise exactly srcLen bytes are.
  63     //
  64     // Typical usage:
  65     //
  66     //          size_t dstLen = conv.ToWChar(NULL, 0, src);
  67     //          if ( dstLen != wxCONV_FAILED )
  68     //              ... handle error ...
  69     //          wchar_t *wbuf = new wchar_t[dstLen];
  70     //          conv.ToWChar(wbuf, dstLen, src);
  71     //
  72     virtual size_t ToWChar(wchar_t *dst, size_t dstLen,
  73                            const char *src, size_t srcLen = -1) const;
  74
  75     virtual size_t FromWChar(char *dst, size_t dstLen,
  76                              const wchar_t *src, size_t srcLen = -1) const;
  77
  78
  79     // Convenience functions for translating NUL-terminated strings: returns
  80     // the buffer containing the converted string or NULL pointer if the
  81     // conversion failed.
  82     const wxWCharBuffer cMB2WC(const char *in) const;
  83     const wxCharBuffer cWC2MB(const wchar_t *in) const;
  84
  85     // Convenience functions for converting strings which may contain embedded
  86     // NULs and don't have to be NUL-terminated.
  87     //
  88     // inLen is the length of the buffer including trailing NUL if any: if the
  89     // last 4 bytes of the buffer are all NULs, these functions are more
  90     // efficient as they avoid copying the string, but otherwise a copy is made
  91     // internally which could be quite bad for (very) long strings.
  92     //
  93     // outLen receives, if not NULL, the length of the converted string or 0 if
  94     // the conversion failed (returning 0 and not -1 in this case makes it
  95     // difficult to distinguish between failed conversion and empty input but
  96     // this is done for backwards compatibility)
  97     const wxWCharBuffer
  98         cMB2WC(const char *in, size_t inLen, size_t *outLen) const;
  99     const wxCharBuffer
 100         cWC2MB(const wchar_t *in, size_t inLen, size_t *outLen) const;
 101
 102     // convenience functions for converting MB or WC to/from wxWin default
 103 #if wxUSE_UNICODE
 104     const wxWCharBuffer cMB2WX(const char *psz) const { return cMB2WC(psz); }
 105     const wxCharBuffer cWX2MB(const wchar_t *psz) const { return cWC2MB(psz); }
 106     const wchar_t* cWC2WX(const wchar_t *psz) const { return psz; }
 107     const wchar_t* cWX2WC(const wchar_t *psz) const { return psz; }
 108 #else // ANSI
 109     const char* cMB2WX(const char *psz) const { return psz; }
 110     const char* cWX2MB(const char *psz) const { return psz; }
 111     const wxCharBuffer cWC2WX(const wchar_t *psz) const { return cWC2MB(psz); }
 112     const wxWCharBuffer cWX2WC(const char *psz) const { return cMB2WC(psz); }
 113 #endif // Unicode/ANSI
 114
 115     // this function is used in the implementation of cMB2WC() to distinguish
 116     // between the following cases:
 117     //
 118     //      a) var width encoding with strings terminated by a single NUL
 119     //         (usual multibyte encodings): return 1 in this case
 120     //      b) fixed width encoding with 2 bytes/char and so terminated by
 121     //         2 NULs (UTF-16/UCS-2 and variants): return 2 in this case
 122     //      c) fixed width encoding with 4 bytes/char and so terminated by
 123     //         4 NULs (UTF-32/UCS-4 and variants): return 4 in this case
 124     //
 125     // anything else is not supported currently and -1 should be returned
 126     virtual size_t GetMBNulLen() const { return 1; }
 127
 128     // return the maximal value currently returned by GetMBNulLen() for any
 129     // encoding
 130     static size_t GetMaxMBNulLen() { return 4 /* for UTF-32 */; }
 131
 132
 133     // The old conversion functions. The existing classes currently mostly
 134     // implement these ones but we're in transition to using To/FromWChar()
 135     // instead and any new classes should implement just the new functions.
 136     // For now, however, we provide default implementation of To/FromWChar() in
 137     // this base class in terms of MB2WC/WC2MB() to avoid having to rewrite all
 138     // the conversions at once.
 139     //
 140     // On success, the return value is the length (i.e. the number of
 141     // characters, not bytes) not counting the trailing NUL(s) of the converted
 142     // string. On failure, (size_t)-1 is returned. In the special case when
 143     // outputBuf is NULL the return value is the same one but nothing is
 144     // written to the buffer.
 145     //
 146     // Note that outLen is the length of the output buffer, not the length of
 147     // the input (which is always supposed to be terminated by one or more
 148     // NULs, as appropriate for the encoding)!
 149     virtual size_t MB2WC(wchar_t *out, const char *in, size_t outLen) const;
 150     virtual size_t WC2MB(char *out, const wchar_t *in, size_t outLen) const;
 151
 152
 153     // make a heap-allocated copy of this object
 154     virtual wxMBConv *Clone() const = 0;
 155
 156     // virtual dtor for any base class
 157     virtual ~wxMBConv();
 158 };
 159
 160 // ----------------------------------------------------------------------------
 161 // wxMBConvLibc uses standard mbstowcs() and wcstombs() functions for
 162 //              conversion (hence it depends on the current locale)
 163 // ----------------------------------------------------------------------------
 164
 165 class WXDLLIMPEXP_BASE wxMBConvLibc : public wxMBConv
 166 {
 167 public:
 168     virtual size_t MB2WC(wchar_t *outputBuf, const char *psz, size_t outputSize) const;
 169     virtual size_t WC2MB(char *outputBuf, const wchar_t *psz, size_t outputSize) const;
 170
 171     virtual wxMBConv *Clone() const { return new wxMBConvLibc; }
 172 };
 173
 174 #ifdef __UNIX__
 175
 176 // ----------------------------------------------------------------------------
 177 // wxConvBrokenFileNames is made for Unix in Unicode mode when
 178 // files are accidentally written in an encoding which is not
 179 // the system encoding. Typically, the system encoding will be
 180 // UTF8 but there might be files stored in ISO8859-1 on disk.
 181 // ----------------------------------------------------------------------------
 182
 183 class WXDLLIMPEXP_BASE wxConvBrokenFileNames : public wxMBConv
 184 {
 185 public:
 186     wxConvBrokenFileNames(const wxChar *charset);
 187     wxConvBrokenFileNames(const wxConvBrokenFileNames& conv)
 188         : m_conv(conv.m_conv ? conv.m_conv->Clone() : NULL)
 189     {
 190     }
 191     virtual ~wxConvBrokenFileNames() { delete m_conv; }
 192
 193     virtual size_t MB2WC(wchar_t *out, const char *in, size_t outLen) const
 194     {
 195         return m_conv->MB2WC(out, in, outLen);
 196     }
 197
 198     virtual size_t WC2MB(char *out, const wchar_t *in, size_t outLen) const
 199     {
 200         return m_conv->WC2MB(out, in, outLen);
 201     }
 202
 203     virtual size_t GetMBNulLen() const
 204     {
 205         // cast needed to call a private function
 206         return m_conv->GetMBNulLen();
 207     }
 208
 209     virtual wxMBConv *Clone() const { return new wxConvBrokenFileNames(*this); }
 210
 211 private:
 212     // the conversion object we forward to
 213     wxMBConv *m_conv;
 214
 215     DECLARE_NO_ASSIGN_CLASS(wxConvBrokenFileNames)
 216 };
 217
 218 #endif // __UNIX__
 219
 220 // ----------------------------------------------------------------------------
 221 // wxMBConvUTF7 (for conversion using UTF7 encoding)
 222 // ----------------------------------------------------------------------------
 223
 224 class WXDLLIMPEXP_BASE wxMBConvUTF7 : public wxMBConv
 225 {
 226 public:
 227     virtual size_t MB2WC(wchar_t *outputBuf, const char *psz, size_t outputSize) const;
 228     virtual size_t WC2MB(char *outputBuf, const wchar_t *psz, size_t outputSize) const;
 229
 230     virtual wxMBConv *Clone() const { return new wxMBConvUTF7; }
 231 };
 232
 233 // ----------------------------------------------------------------------------
 234 // wxMBConvUTF8 (for conversion using UTF8 encoding)
 235 // ----------------------------------------------------------------------------
 236
 237 class WXDLLIMPEXP_BASE wxMBConvUTF8 : public wxMBConv
 238 {
 239 public:
 240     enum
 241     {
 242         MAP_INVALID_UTF8_NOT = 0,
 243         MAP_INVALID_UTF8_TO_PUA = 1,
 244         MAP_INVALID_UTF8_TO_OCTAL = 2
 245     };
 246
 247     wxMBConvUTF8(int options = MAP_INVALID_UTF8_NOT) : m_options(options) { }
 248     virtual size_t MB2WC(wchar_t *outputBuf, const char *psz, size_t outputSize) const;
 249     virtual size_t WC2MB(char *outputBuf, const wchar_t *psz, size_t outputSize) const;
 250
 251     virtual wxMBConv *Clone() const { return new wxMBConvUTF8(m_options); }
 252
 253 private:
 254     int m_options;
 255 };
 256
 257 // ----------------------------------------------------------------------------
 258 // wxMBConvUTF16Base: for both LE and BE variants
 259 // ----------------------------------------------------------------------------
 260
 261 class WXDLLIMPEXP_BASE wxMBConvUTF16Base : public wxMBConv
 262 {
 263 public:
 264     virtual size_t GetMBNulLen() const { return 2; }
 265 };
 266
 267 // ----------------------------------------------------------------------------
 268 // wxMBConvUTF16LE (for conversion using UTF16 Little Endian encoding)
 269 // ----------------------------------------------------------------------------
 270
 271 class WXDLLIMPEXP_BASE wxMBConvUTF16LE : public wxMBConvUTF16Base
 272 {
 273 public:
 274     virtual size_t MB2WC(wchar_t *outputBuf, const char *psz, size_t outputSize) const;
 275     virtual size_t WC2MB(char *outputBuf, const wchar_t *psz, size_t outputSize) const;
 276     virtual wxMBConv *Clone() const { return new wxMBConvUTF16LE; }
 277 };
 278
 279 // ----------------------------------------------------------------------------
 280 // wxMBConvUTF16BE (for conversion using UTF16 Big Endian encoding)
 281 // ----------------------------------------------------------------------------
 282
 283 class WXDLLIMPEXP_BASE wxMBConvUTF16BE : public wxMBConvUTF16Base
 284 {
 285 public:
 286     virtual size_t MB2WC(wchar_t *outputBuf, const char *psz, size_t outputSize) const;
 287     virtual size_t WC2MB(char *outputBuf, const wchar_t *psz, size_t outputSize) const;
 288     virtual wxMBConv *Clone() const { return new wxMBConvUTF16BE; }
 289 };
 290
 291 // ----------------------------------------------------------------------------
 292 // wxMBConvUTF32Base: base class for both LE and BE variants
 293 // ----------------------------------------------------------------------------
 294
 295 class WXDLLIMPEXP_BASE wxMBConvUTF32Base : public wxMBConv
 296 {
 297 public:
 298     virtual size_t GetMBNulLen() const { return 4; }
 299 };
 300
 301 // ----------------------------------------------------------------------------
 302 // wxMBConvUTF32LE (for conversion using UTF32 Little Endian encoding)
 303 // ----------------------------------------------------------------------------
 304
 305 class WXDLLIMPEXP_BASE wxMBConvUTF32LE : public wxMBConvUTF32Base
 306 {
 307 public:
 308     virtual size_t MB2WC(wchar_t *outputBuf, const char *psz, size_t outputSize) const;
 309     virtual size_t WC2MB(char *outputBuf, const wchar_t *psz, size_t outputSize) const;
 310     virtual wxMBConv *Clone() const { return new wxMBConvUTF32LE; }
 311 };
 312
 313 // ----------------------------------------------------------------------------
 314 // wxMBConvUTF32BE (for conversion using UTF32 Big Endian encoding)
 315 // ----------------------------------------------------------------------------
 316
 317 class WXDLLIMPEXP_BASE wxMBConvUTF32BE : public wxMBConvUTF32Base
 318 {
 319 public:
 320     virtual size_t MB2WC(wchar_t *outputBuf, const char *psz, size_t outputSize) const;
 321     virtual size_t WC2MB(char *outputBuf, const wchar_t *psz, size_t outputSize) const;
 322     virtual wxMBConv *Clone() const { return new wxMBConvUTF32BE; }
 323 };
 324
 325 // ----------------------------------------------------------------------------
 326 // wxCSConv (for conversion based on loadable char sets)
 327 // ----------------------------------------------------------------------------
 328
 329 #include "wx/fontenc.h"
 330
 331 class WXDLLIMPEXP_BASE wxCSConv : public wxMBConv
 332 {
 333 public:
 334     // we can be created either from charset name or from an encoding constant
 335     // but we can't have both at once
 336     wxCSConv(const wxChar *charset);
 337     wxCSConv(wxFontEncoding encoding);
 338
 339     wxCSConv(const wxCSConv& conv);
 340     virtual ~wxCSConv();
 341
 342     wxCSConv& operator=(const wxCSConv& conv);
 343
 344     virtual size_t MB2WC(wchar_t *outputBuf, const char *psz, size_t outputSize) const;
 345     virtual size_t WC2MB(char *outputBuf, const wchar_t *psz, size_t outputSize) const;
 346     virtual size_t GetMBNulLen() const;
 347     virtual wxMBConv *Clone() const { return new wxCSConv(*this); }
 348
 349     void Clear();
 350
 351 private:
 352     // common part of all ctors
 353     void Init();
 354
 355     // creates m_convReal if necessary
 356     void CreateConvIfNeeded() const;
 357
 358     // do create m_convReal (unconditionally)
 359     wxMBConv *DoCreate() const;
 360
 361     // set the name (may be only called when m_name == NULL), makes copy of
 362     // charset string
 363     void SetName(const wxChar *charset);
 364
 365
 366     // note that we can't use wxString here because of compilation
 367     // dependencies: we're included from wx/string.h
 368     wxChar *m_name;
 369     wxFontEncoding m_encoding;
 370
 371     // use CreateConvIfNeeded() before accessing m_convReal!
 372     wxMBConv *m_convReal;
 373     bool m_deferred;
 374 };
 375
 376
 377 // ----------------------------------------------------------------------------
 378 // declare predefined conversion objects
 379 // ----------------------------------------------------------------------------
 380
 381 // conversion to be used with all standard functions affected by locale, e.g.
 382 // strtol(), strftime(), ...
 383 extern WXDLLIMPEXP_DATA_BASE(wxMBConv&) wxConvLibc;
 384
 385 // conversion ISO-8859-1/UTF-7/UTF-8 <-> wchar_t
 386 extern WXDLLIMPEXP_DATA_BASE(wxCSConv&) wxConvISO8859_1;
 387 extern WXDLLIMPEXP_DATA_BASE(wxMBConvUTF7&) wxConvUTF7;
 388 extern WXDLLIMPEXP_DATA_BASE(wxMBConvUTF8&) wxConvUTF8;
 389
 390 // conversion used for the file names on the systems where they're not Unicode
 391 // (basically anything except Windows)
 392 //
 393 // this is used by all file functions, can be changed by the application
 394 //
 395 // by default UTF-8 under Mac OS X and wxConvLibc elsewhere (but it's not used
 396 // under Windows normally)
 397 extern WXDLLIMPEXP_DATA_BASE(wxMBConv *) wxConvFileName;
 398
 399 // backwards compatible define
 400 #define wxConvFile (*wxConvFileName)
 401
 402 // the current conversion object, may be set to any conversion, is used by
 403 // default in a couple of places inside wx (initially same as wxConvLibc)
 404 extern WXDLLIMPEXP_DATA_BASE(wxMBConv *) wxConvCurrent;
 405
 406 // ???
 407 extern WXDLLIMPEXP_DATA_BASE(wxCSConv&) wxConvLocal;
 408
 409
 410 // ----------------------------------------------------------------------------
 411 // endianness-dependent conversions
 412 // ----------------------------------------------------------------------------
 413
 414 #ifdef WORDS_BIGENDIAN
 415     typedef wxMBConvUTF16BE wxMBConvUTF16;
 416     typedef wxMBConvUTF32BE wxMBConvUTF32;
 417 #else
 418     typedef wxMBConvUTF16LE wxMBConvUTF16;
 419     typedef wxMBConvUTF32LE wxMBConvUTF32;
 420 #endif
 421
 422 // ----------------------------------------------------------------------------
 423 // filename conversion macros
 424 // ----------------------------------------------------------------------------
 425
 426 // filenames are multibyte on Unix and probably widechar on Windows?
 427 #if defined(__UNIX__) || defined(__BORLANDC__) || defined(__WXMAC__ )
 428     #define wxMBFILES 1
 429 #else
 430     #define wxMBFILES 0
 431 #endif
 432
 433 #if wxMBFILES && wxUSE_UNICODE
 434     #define wxFNCONV(name) wxConvFileName->cWX2MB(name)
 435     #define wxFNSTRINGCAST wxMBSTRINGCAST
 436 #else
 437 #if defined( __WXOSX__ ) && wxMBFILES
 438     #define wxFNCONV(name) wxConvFileName->cWC2MB( wxConvLocal.cWX2WC(name) )
 439 #else
 440     #define wxFNCONV(name) name
 441 #endif
 442     #define wxFNSTRINGCAST WXSTRINGCAST
 443 #endif
 444
 445 #else // !wxUSE_WCHAR_T
 446
 447 // ----------------------------------------------------------------------------
 448 // stand-ins in absence of wchar_t
 449 // ----------------------------------------------------------------------------
 450
 451 class WXDLLIMPEXP_BASE wxMBConv
 452 {
 453 public:
 454     const char* cMB2WX(const char *psz) const { return psz; }
 455     const char* cWX2MB(const char *psz) const { return psz; }
 456 };
 457
 458 #define wxConvFile wxConvLocal
 459
 460 extern WXDLLIMPEXP_DATA_BASE(wxMBConv) wxConvLibc,
 461                                        wxConvLocal,
 462                                        wxConvISO8859_1,
 463                                        wxConvUTF8;
 464 extern WXDLLIMPEXP_DATA_BASE(wxMBConv *) wxConvCurrent;
 465
 466 #define wxFNCONV(name) name
 467 #define wxFNSTRINGCAST WXSTRINGCAST
 468
 469 #endif
 470   // wxUSE_WCHAR_T
 471
 472 // ----------------------------------------------------------------------------
 473 // macros for the most common conversions
 474 // ----------------------------------------------------------------------------
 475
 476 #if wxUSE_UNICODE
 477     #define wxConvertWX2MB(s)   wxConvCurrent->cWX2MB(s)
 478     #define wxConvertMB2WX(s)   wxConvCurrent->cMB2WX(s)
 479 #else // ANSI
 480     // no conversions to do
 481     #define wxConvertWX2MB(s)   (s)
 482     #define wxConvertMB2WX(s)   (s)
 483 #endif // Unicode/ANSI
 484
 485 #endif // _WX_STRCONV_H_
 486