1 ///////////////////////////////////////////////////////////////////////////////
3 // Purpose: conversion routines for char sets any Unicode
4 // Author: Ove Kaaven, Robert Roebling, Vadim Zeitlin
8 // Copyright: (c) 1998 Ove Kaaven, Robert Roebling
9 // (c) 1998-2006 Vadim Zeitlin
10 // Licence: wxWindows licence
11 ///////////////////////////////////////////////////////////////////////////////
13 #ifndef _WX_STRCONV_H_
14 #define _WX_STRCONV_H_
17 #include "wx/wxchar.h"
18 #include "wx/buffer.h"
20 #ifdef __DIGITALMARS__
24 #if defined(__VISAGECPP__) && __IBMCPP__ >= 400
32 // the error value returned by wxMBConv methods
33 #define wxCONV_FAILED ((size_t)-1)
35 // the default value for some length parameters meaning that the string is
37 #define wxNO_LEN ((size_t)-1)
39 // ----------------------------------------------------------------------------
40 // wxMBConv (abstract base class for conversions)
41 // ----------------------------------------------------------------------------
43 // When deriving a new class from wxMBConv you must reimplement ToWChar() and
44 // FromWChar() methods which are not pure virtual only for historical reasons,
45 // don't let the fact that the existing classes implement MB2WC/WC2MB() instead
48 // You also have to implement Clone() to allow copying the conversions
51 // And you might need to override GetMBNulLen() as well.
52 class WXDLLIMPEXP_BASE wxMBConv
55 // The functions doing actual conversion from/to narrow to/from wide
58 // On success, the return value is the length (i.e. the number of
59 // characters, not bytes) of the converted string including any trailing
60 // L'\0' or (possibly multiple) '\0'(s). If the conversion fails or if
61 // there is not enough space for everything, including the trailing NUL
62 // character(s), in the output buffer, wxCONV_FAILED is returned.
64 // In the special case when dstLen is 0 (outputBuf may be NULL then) the
65 // return value is the length of the needed buffer but nothing happens
66 // otherwise. If srcLen is wxNO_LEN, the entire string, up to and
67 // including the trailing NUL(s), is converted, otherwise exactly srcLen
72 // size_t dstLen = conv.ToWChar(NULL, 0, src);
73 // if ( dstLen != wxCONV_FAILED )
74 // ... handle error ...
75 // wchar_t *wbuf = new wchar_t[dstLen];
76 // conv.ToWChar(wbuf, dstLen, src);
78 virtual size_t ToWChar(wchar_t *dst
, size_t dstLen
,
79 const char *src
, size_t srcLen
= wxNO_LEN
) const;
81 virtual size_t FromWChar(char *dst
, size_t dstLen
,
82 const wchar_t *src
, size_t srcLen
= wxNO_LEN
) const;
85 // Convenience functions for translating NUL-terminated strings: returns
86 // the buffer containing the converted string or NULL pointer if the
88 const wxWCharBuffer
cMB2WC(const char *in
) const;
89 const wxCharBuffer
cWC2MB(const wchar_t *in
) const;
91 // Convenience functions for converting strings which may contain embedded
92 // NULs and don't have to be NUL-terminated.
94 // inLen is the length of the buffer including trailing NUL if any: if the
95 // last 4 bytes of the buffer are all NULs, these functions are more
96 // efficient as they avoid copying the string, but otherwise a copy is made
97 // internally which could be quite bad for (very) long strings.
99 // outLen receives, if not NULL, the length of the converted string or 0 if
100 // the conversion failed (returning 0 and not -1 in this case makes it
101 // difficult to distinguish between failed conversion and empty input but
102 // this is done for backwards compatibility)
104 cMB2WC(const char *in
, size_t inLen
, size_t *outLen
) const;
106 cWC2MB(const wchar_t *in
, size_t inLen
, size_t *outLen
) const;
108 // convenience functions for converting MB or WC to/from wxWin default
110 const wxWCharBuffer
cMB2WX(const char *psz
) const { return cMB2WC(psz
); }
111 const wxCharBuffer
cWX2MB(const wchar_t *psz
) const { return cWC2MB(psz
); }
112 const wchar_t* cWC2WX(const wchar_t *psz
) const { return psz
; }
113 const wchar_t* cWX2WC(const wchar_t *psz
) const { return psz
; }
115 const char* cMB2WX(const char *psz
) const { return psz
; }
116 const char* cWX2MB(const char *psz
) const { return psz
; }
117 const wxCharBuffer
cWC2WX(const wchar_t *psz
) const { return cWC2MB(psz
); }
118 const wxWCharBuffer
cWX2WC(const char *psz
) const { return cMB2WC(psz
); }
119 #endif // Unicode/ANSI
121 // this function is used in the implementation of cMB2WC() to distinguish
122 // between the following cases:
124 // a) var width encoding with strings terminated by a single NUL
125 // (usual multibyte encodings): return 1 in this case
126 // b) fixed width encoding with 2 bytes/char and so terminated by
127 // 2 NULs (UTF-16/UCS-2 and variants): return 2 in this case
128 // c) fixed width encoding with 4 bytes/char and so terminated by
129 // 4 NULs (UTF-32/UCS-4 and variants): return 4 in this case
131 // anything else is not supported currently and -1 should be returned
132 virtual size_t GetMBNulLen() const { return 1; }
134 // return the maximal value currently returned by GetMBNulLen() for any
136 static size_t GetMaxMBNulLen() { return 4 /* for UTF-32 */; }
139 // The old conversion functions. The existing classes currently mostly
140 // implement these ones but we're in transition to using To/FromWChar()
141 // instead and any new classes should implement just the new functions.
142 // For now, however, we provide default implementation of To/FromWChar() in
143 // this base class in terms of MB2WC/WC2MB() to avoid having to rewrite all
144 // the conversions at once.
146 // On success, the return value is the length (i.e. the number of
147 // characters, not bytes) not counting the trailing NUL(s) of the converted
148 // string. On failure, (size_t)-1 is returned. In the special case when
149 // outputBuf is NULL the return value is the same one but nothing is
150 // written to the buffer.
152 // Note that outLen is the length of the output buffer, not the length of
153 // the input (which is always supposed to be terminated by one or more
154 // NULs, as appropriate for the encoding)!
155 virtual size_t MB2WC(wchar_t *out
, const char *in
, size_t outLen
) const;
156 virtual size_t WC2MB(char *out
, const wchar_t *in
, size_t outLen
) const;
159 // make a heap-allocated copy of this object
160 virtual wxMBConv
*Clone() const = 0;
162 // virtual dtor for any base class
166 // ----------------------------------------------------------------------------
167 // wxMBConvLibc uses standard mbstowcs() and wcstombs() functions for
168 // conversion (hence it depends on the current locale)
169 // ----------------------------------------------------------------------------
171 class WXDLLIMPEXP_BASE wxMBConvLibc
: public wxMBConv
174 virtual size_t MB2WC(wchar_t *outputBuf
, const char *psz
, size_t outputSize
) const;
175 virtual size_t WC2MB(char *outputBuf
, const wchar_t *psz
, size_t outputSize
) const;
177 virtual wxMBConv
*Clone() const { return new wxMBConvLibc
; }
182 // ----------------------------------------------------------------------------
183 // wxConvBrokenFileNames is made for Unix in Unicode mode when
184 // files are accidentally written in an encoding which is not
185 // the system encoding. Typically, the system encoding will be
186 // UTF8 but there might be files stored in ISO8859-1 on disk.
187 // ----------------------------------------------------------------------------
189 class WXDLLIMPEXP_BASE wxConvBrokenFileNames
: public wxMBConv
192 wxConvBrokenFileNames(const wxChar
*charset
);
193 wxConvBrokenFileNames(const wxConvBrokenFileNames
& conv
)
195 m_conv(conv
.m_conv
? conv
.m_conv
->Clone() : NULL
)
198 virtual ~wxConvBrokenFileNames() { delete m_conv
; }
200 virtual size_t MB2WC(wchar_t *out
, const char *in
, size_t outLen
) const
202 return m_conv
->MB2WC(out
, in
, outLen
);
205 virtual size_t WC2MB(char *out
, const wchar_t *in
, size_t outLen
) const
207 return m_conv
->WC2MB(out
, in
, outLen
);
210 virtual size_t GetMBNulLen() const
212 // cast needed to call a private function
213 return m_conv
->GetMBNulLen();
216 virtual wxMBConv
*Clone() const { return new wxConvBrokenFileNames(*this); }
219 // the conversion object we forward to
222 DECLARE_NO_ASSIGN_CLASS(wxConvBrokenFileNames
)
227 // ----------------------------------------------------------------------------
228 // wxMBConvUTF7 (for conversion using UTF7 encoding)
229 // ----------------------------------------------------------------------------
231 class WXDLLIMPEXP_BASE wxMBConvUTF7
: public wxMBConv
234 virtual size_t MB2WC(wchar_t *outputBuf
, const char *psz
, size_t outputSize
) const;
235 virtual size_t WC2MB(char *outputBuf
, const wchar_t *psz
, size_t outputSize
) const;
237 virtual wxMBConv
*Clone() const { return new wxMBConvUTF7
; }
240 // ----------------------------------------------------------------------------
241 // wxMBConvUTF8 (for conversion using UTF8 encoding)
242 // ----------------------------------------------------------------------------
244 class WXDLLIMPEXP_BASE wxMBConvUTF8
: public wxMBConv
249 MAP_INVALID_UTF8_NOT
= 0,
250 MAP_INVALID_UTF8_TO_PUA
= 1,
251 MAP_INVALID_UTF8_TO_OCTAL
= 2
254 wxMBConvUTF8(int options
= MAP_INVALID_UTF8_NOT
) : m_options(options
) { }
255 virtual size_t MB2WC(wchar_t *outputBuf
, const char *psz
, size_t outputSize
) const;
256 virtual size_t WC2MB(char *outputBuf
, const wchar_t *psz
, size_t outputSize
) const;
258 virtual wxMBConv
*Clone() const { return new wxMBConvUTF8(m_options
); }
264 // ----------------------------------------------------------------------------
265 // wxMBConvUTF16Base: for both LE and BE variants
266 // ----------------------------------------------------------------------------
268 class WXDLLIMPEXP_BASE wxMBConvUTF16Base
: public wxMBConv
271 enum { BYTES_PER_CHAR
= 2 };
273 virtual size_t GetMBNulLen() const { return BYTES_PER_CHAR
; }
276 // return the length of the buffer using srcLen if it's not wxNO_LEN and
277 // computing the length ourselves if it is; also checks that the length is
278 // even if specified as we need an entire number of UTF-16 characters and
279 // returns wxNO_LEN which indicates error if it is odd
280 static size_t GetLength(const char *src
, size_t srcLen
);
283 // ----------------------------------------------------------------------------
284 // wxMBConvUTF16LE (for conversion using UTF16 Little Endian encoding)
285 // ----------------------------------------------------------------------------
287 class WXDLLIMPEXP_BASE wxMBConvUTF16LE
: public wxMBConvUTF16Base
290 virtual size_t ToWChar(wchar_t *dst
, size_t dstLen
,
291 const char *src
, size_t srcLen
= wxNO_LEN
) const;
292 virtual size_t FromWChar(char *dst
, size_t dstLen
,
293 const wchar_t *src
, size_t srcLen
= wxNO_LEN
) const;
294 virtual wxMBConv
*Clone() const { return new wxMBConvUTF16LE
; }
297 // ----------------------------------------------------------------------------
298 // wxMBConvUTF16BE (for conversion using UTF16 Big Endian encoding)
299 // ----------------------------------------------------------------------------
301 class WXDLLIMPEXP_BASE wxMBConvUTF16BE
: public wxMBConvUTF16Base
304 virtual size_t ToWChar(wchar_t *dst
, size_t dstLen
,
305 const char *src
, size_t srcLen
= wxNO_LEN
) const;
306 virtual size_t FromWChar(char *dst
, size_t dstLen
,
307 const wchar_t *src
, size_t srcLen
= wxNO_LEN
) const;
308 virtual wxMBConv
*Clone() const { return new wxMBConvUTF16BE
; }
311 // ----------------------------------------------------------------------------
312 // wxMBConvUTF32Base: base class for both LE and BE variants
313 // ----------------------------------------------------------------------------
315 class WXDLLIMPEXP_BASE wxMBConvUTF32Base
: public wxMBConv
318 enum { BYTES_PER_CHAR
= 4 };
320 virtual size_t GetMBNulLen() const { return BYTES_PER_CHAR
; }
323 // this is similar to wxMBConvUTF16Base method with the same name except
324 // that, of course, it verifies that length is divisible by 4 if given and
326 static size_t GetLength(const char *src
, size_t srcLen
);
329 // ----------------------------------------------------------------------------
330 // wxMBConvUTF32LE (for conversion using UTF32 Little Endian encoding)
331 // ----------------------------------------------------------------------------
333 class WXDLLIMPEXP_BASE wxMBConvUTF32LE
: public wxMBConvUTF32Base
336 virtual size_t ToWChar(wchar_t *dst
, size_t dstLen
,
337 const char *src
, size_t srcLen
= wxNO_LEN
) const;
338 virtual size_t FromWChar(char *dst
, size_t dstLen
,
339 const wchar_t *src
, size_t srcLen
= wxNO_LEN
) const;
340 virtual wxMBConv
*Clone() const { return new wxMBConvUTF32LE
; }
343 // ----------------------------------------------------------------------------
344 // wxMBConvUTF32BE (for conversion using UTF32 Big Endian encoding)
345 // ----------------------------------------------------------------------------
347 class WXDLLIMPEXP_BASE wxMBConvUTF32BE
: public wxMBConvUTF32Base
350 virtual size_t ToWChar(wchar_t *dst
, size_t dstLen
,
351 const char *src
, size_t srcLen
= wxNO_LEN
) const;
352 virtual size_t FromWChar(char *dst
, size_t dstLen
,
353 const wchar_t *src
, size_t srcLen
= wxNO_LEN
) const;
354 virtual wxMBConv
*Clone() const { return new wxMBConvUTF32BE
; }
357 // ----------------------------------------------------------------------------
358 // wxCSConv (for conversion based on loadable char sets)
359 // ----------------------------------------------------------------------------
361 #include "wx/fontenc.h"
363 class WXDLLIMPEXP_BASE wxCSConv
: public wxMBConv
366 // we can be created either from charset name or from an encoding constant
367 // but we can't have both at once
368 wxCSConv(const wxChar
*charset
);
369 wxCSConv(wxFontEncoding encoding
);
371 wxCSConv(const wxCSConv
& conv
);
374 wxCSConv
& operator=(const wxCSConv
& conv
);
376 virtual size_t ToWChar(wchar_t *dst
, size_t dstLen
,
377 const char *src
, size_t srcLen
= wxNO_LEN
) const;
378 virtual size_t FromWChar(char *dst
, size_t dstLen
,
379 const wchar_t *src
, size_t srcLen
= wxNO_LEN
) const;
380 virtual size_t MB2WC(wchar_t *outputBuf
, const char *psz
, size_t outputSize
) const;
381 virtual size_t WC2MB(char *outputBuf
, const wchar_t *psz
, size_t outputSize
) const;
382 virtual size_t GetMBNulLen() const;
384 virtual wxMBConv
*Clone() const { return new wxCSConv(*this); }
388 #if wxABI_VERSION >= 20802
389 // return true if the conversion could be initilized successfully
394 // common part of all ctors
397 // creates m_convReal if necessary
398 void CreateConvIfNeeded() const;
400 // do create m_convReal (unconditionally)
401 wxMBConv
*DoCreate() const;
403 // set the name (may be only called when m_name == NULL), makes copy of
405 void SetName(const wxChar
*charset
);
408 // note that we can't use wxString here because of compilation
409 // dependencies: we're included from wx/string.h
411 wxFontEncoding m_encoding
;
413 // use CreateConvIfNeeded() before accessing m_convReal!
414 wxMBConv
*m_convReal
;
419 // ----------------------------------------------------------------------------
420 // declare predefined conversion objects
421 // ----------------------------------------------------------------------------
423 // conversion to be used with all standard functions affected by locale, e.g.
424 // strtol(), strftime(), ...
425 extern WXDLLIMPEXP_DATA_BASE(wxMBConv
&) wxConvLibc
;
427 // conversion ISO-8859-1/UTF-7/UTF-8 <-> wchar_t
428 extern WXDLLIMPEXP_DATA_BASE(wxCSConv
&) wxConvISO8859_1
;
429 extern WXDLLIMPEXP_DATA_BASE(wxMBConvUTF7
&) wxConvUTF7
;
430 extern WXDLLIMPEXP_DATA_BASE(wxMBConvUTF8
&) wxConvUTF8
;
432 // conversion used for the file names on the systems where they're not Unicode
433 // (basically anything except Windows)
435 // this is used by all file functions, can be changed by the application
437 // by default UTF-8 under Mac OS X and wxConvLibc elsewhere (but it's not used
438 // under Windows normally)
439 extern WXDLLIMPEXP_DATA_BASE(wxMBConv
*) wxConvFileName
;
441 // backwards compatible define
442 #define wxConvFile (*wxConvFileName)
444 // the current conversion object, may be set to any conversion, is used by
445 // default in a couple of places inside wx (initially same as wxConvLibc)
446 extern WXDLLIMPEXP_DATA_BASE(wxMBConv
*) wxConvCurrent
;
448 // the conversion corresponding to the current locale
449 extern WXDLLIMPEXP_DATA_BASE(wxCSConv
&) wxConvLocal
;
451 // the conversion corresponding to the encoding of the standard UI elements
453 // by default this is the same as wxConvLocal but may be changed if the program
454 // needs to use a fixed encoding
455 extern WXDLLIMPEXP_DATA_BASE(wxMBConv
*) wxConvUI
;
457 // ----------------------------------------------------------------------------
458 // endianness-dependent conversions
459 // ----------------------------------------------------------------------------
461 #ifdef WORDS_BIGENDIAN
462 typedef wxMBConvUTF16BE wxMBConvUTF16
;
463 typedef wxMBConvUTF32BE wxMBConvUTF32
;
465 typedef wxMBConvUTF16LE wxMBConvUTF16
;
466 typedef wxMBConvUTF32LE wxMBConvUTF32
;
469 // ----------------------------------------------------------------------------
470 // filename conversion macros
471 // ----------------------------------------------------------------------------
473 // filenames are multibyte on Unix and widechar on Windows
474 #if defined(__UNIX__) || defined(__WXMAC__)
480 #if wxMBFILES && wxUSE_UNICODE
481 #define wxFNCONV(name) wxConvFileName->cWX2MB(name)
482 #define wxFNSTRINGCAST wxMBSTRINGCAST
484 #if defined( __WXOSX__ ) && wxMBFILES
485 #define wxFNCONV(name) wxConvFileName->cWC2MB( wxConvLocal.cWX2WC(name) )
487 #define wxFNCONV(name) name
489 #define wxFNSTRINGCAST WXSTRINGCAST
492 #else // !wxUSE_WCHAR_T
494 // ----------------------------------------------------------------------------
495 // stand-ins in absence of wchar_t
496 // ----------------------------------------------------------------------------
498 class WXDLLIMPEXP_BASE wxMBConv
501 const char* cMB2WX(const char *psz
) const { return psz
; }
502 const char* cWX2MB(const char *psz
) const { return psz
; }
505 #define wxConvFile wxConvLocal
507 extern WXDLLIMPEXP_DATA_BASE(wxMBConv
) wxConvLibc
,
511 extern WXDLLIMPEXP_DATA_BASE(wxMBConv
*) wxConvCurrent
;
513 #define wxFNCONV(name) name
514 #define wxFNSTRINGCAST WXSTRINGCAST
519 // ----------------------------------------------------------------------------
520 // macros for the most common conversions
521 // ----------------------------------------------------------------------------
524 #define wxConvertWX2MB(s) wxConvCurrent->cWX2MB(s)
525 #define wxConvertMB2WX(s) wxConvCurrent->cMB2WX(s)
527 // no conversions to do
528 #define wxConvertWX2MB(s) (s)
529 #define wxConvertMB2WX(s) (s)
530 #endif // Unicode/ANSI
532 #endif // _WX_STRCONV_H_