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