]> git.saurik.com Git - wxWidgets.git/blame - include/wx/strconv.h
Rebake from clean wx tree.
[wxWidgets.git] / include / wx / strconv.h
CommitLineData
6001e347
RR
1///////////////////////////////////////////////////////////////////////////////
2// Name: strconv.h
3// Purpose: conversion routines for char sets any Unicode
467e0479 4// Author: Ove Kaaven, Robert Roebling, Vadim Zeitlin
6001e347
RR
5// Modified by:
6// Created: 29/01/98
7// RCS-ID: $Id$
467e0479
VZ
8// Copyright: (c) 1998 Ove Kaaven, Robert Roebling
9// (c) 1998-2006 Vadim Zeitlin
65571936 10// Licence: wxWindows licence
6001e347
RR
11///////////////////////////////////////////////////////////////////////////////
12
d36c9347
VZ
13#ifndef _WX_STRCONV_H_
14#define _WX_STRCONV_H_
6001e347 15
6001e347 16#include "wx/defs.h"
e3f6cbd9 17#include "wx/chartype.h"
6001e347
RR
18#include "wx/buffer.h"
19
7db39dd6
CE
20#ifdef __DIGITALMARS__
21#include "typeinfo.h"
22#endif
23
9dea36ef
DW
24#if defined(__VISAGECPP__) && __IBMCPP__ >= 400
25# undef __BSEXCPT__
26#endif
dccce9ea 27
9b4da627 28#ifndef __WXPALMOS5__
6001e347 29#include <stdlib.h>
9b4da627 30#endif // ! __WXPALMOS5__
6001e347
RR
31
32#if wxUSE_WCHAR_T
33
b5dbe15d 34class WXDLLIMPEXP_FWD_BASE wxString;
86501081 35
483b0434
VZ
36// the error value returned by wxMBConv methods
37#define wxCONV_FAILED ((size_t)-1)
38
e90c1d2a 39// ----------------------------------------------------------------------------
bde4baac 40// wxMBConv (abstract base class for conversions)
e90c1d2a 41// ----------------------------------------------------------------------------
6001e347 42
509da451
VZ
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
46// confuse you.
47//
d36c9347
VZ
48// You also have to implement Clone() to allow copying the conversions
49// polymorphically.
50//
509da451 51// And you might need to override GetMBNulLen() as well.
bddd7a8d 52class WXDLLIMPEXP_BASE wxMBConv
6001e347
RR
53{
54public:
483b0434
VZ
55 // The functions doing actual conversion from/to narrow to/from wide
56 // character strings.
bde4baac 57 //
483b0434
VZ
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
467e0479 62 // character(s), in the output buffer, wxCONV_FAILED is returned.
483b0434 63 //
96132605
VZ
64 // In the special case when dst is NULL (the value of dstLen is ignored
65 // then) the return value is the length of the needed buffer but nothing
66 // happens otherwise. If srcLen is wxNO_LEN, the entire string, up to and
467e0479
VZ
67 // including the trailing NUL(s), is converted, otherwise exactly srcLen
68 // bytes are.
483b0434
VZ
69 //
70 // Typical usage:
71 //
72 // size_t dstLen = conv.ToWChar(NULL, 0, src);
96132605 73 // if ( dstLen == wxCONV_FAILED )
483b0434
VZ
74 // ... handle error ...
75 // wchar_t *wbuf = new wchar_t[dstLen];
76 // conv.ToWChar(wbuf, dstLen, src);
96132605
VZ
77 // ... work with wbuf ...
78 // delete [] wbuf;
483b0434
VZ
79 //
80 virtual size_t ToWChar(wchar_t *dst, size_t dstLen,
467e0479 81 const char *src, size_t srcLen = wxNO_LEN) const;
483b0434
VZ
82
83 virtual size_t FromWChar(char *dst, size_t dstLen,
467e0479 84 const wchar_t *src, size_t srcLen = wxNO_LEN) const;
483b0434 85
e90c1d2a 86
483b0434
VZ
87 // Convenience functions for translating NUL-terminated strings: returns
88 // the buffer containing the converted string or NULL pointer if the
89 // conversion failed.
eec47cc6
VZ
90 const wxWCharBuffer cMB2WC(const char *in) const;
91 const wxCharBuffer cWC2MB(const wchar_t *in) const;
6001e347 92
483b0434
VZ
93 // Convenience functions for converting strings which may contain embedded
94 // NULs and don't have to be NUL-terminated.
f5fb6871 95 //
f6a02087
VZ
96 // inLen is the length of the buffer including trailing NUL if any or
97 // wxNO_LEN if the input is NUL-terminated.
eec47cc6
VZ
98 //
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
f6a02087
VZ
102 // this is done for backwards compatibility). Notice that the rules for
103 // whether outLen accounts or not for the last NUL are the same as for
104 // To/FromWChar() above: if inLen is specified, outLen is exactly the
105 // number of characters converted, whether the last one of them was NUL or
106 // not. But if inLen == wxNO_LEN then outLen doesn't account for the last
107 // NUL even though it is present.
eec47cc6
VZ
108 const wxWCharBuffer
109 cMB2WC(const char *in, size_t inLen, size_t *outLen) const;
110 const wxCharBuffer
111 cWC2MB(const wchar_t *in, size_t inLen, size_t *outLen) const;
f5fb6871 112
40ac5040
VZ
113 // And yet more convenience functions for converting the entire buffers:
114 // these are the simplest and least error-prone as you never need to bother
115 // with lengths/sizes directly.
116 const wxWCharBuffer cMB2WC(const wxScopedCharBuffer& in) const;
117 const wxCharBuffer cWC2MB(const wxScopedWCharBuffer& in) const;
118
bde4baac 119 // convenience functions for converting MB or WC to/from wxWin default
6001e347 120#if wxUSE_UNICODE
e90c1d2a
VZ
121 const wxWCharBuffer cMB2WX(const char *psz) const { return cMB2WC(psz); }
122 const wxCharBuffer cWX2MB(const wchar_t *psz) const { return cWC2MB(psz); }
123 const wchar_t* cWC2WX(const wchar_t *psz) const { return psz; }
f6bcfd97 124 const wchar_t* cWX2WC(const wchar_t *psz) const { return psz; }
e90c1d2a
VZ
125#else // ANSI
126 const char* cMB2WX(const char *psz) const { return psz; }
127 const char* cWX2MB(const char *psz) const { return psz; }
128 const wxCharBuffer cWC2WX(const wchar_t *psz) const { return cWC2MB(psz); }
129 const wxWCharBuffer cWX2WC(const char *psz) const { return cMB2WC(psz); }
130#endif // Unicode/ANSI
2b5f62a0 131
c1464d9d
VZ
132 // this function is used in the implementation of cMB2WC() to distinguish
133 // between the following cases:
eec47cc6 134 //
c1464d9d
VZ
135 // a) var width encoding with strings terminated by a single NUL
136 // (usual multibyte encodings): return 1 in this case
137 // b) fixed width encoding with 2 bytes/char and so terminated by
138 // 2 NULs (UTF-16/UCS-2 and variants): return 2 in this case
139 // c) fixed width encoding with 4 bytes/char and so terminated by
140 // 4 NULs (UTF-32/UCS-4 and variants): return 4 in this case
141 //
142 // anything else is not supported currently and -1 should be returned
7ef3ab50
VZ
143 virtual size_t GetMBNulLen() const { return 1; }
144
483b0434
VZ
145 // return the maximal value currently returned by GetMBNulLen() for any
146 // encoding
147 static size_t GetMaxMBNulLen() { return 4 /* for UTF-32 */; }
148
111d9948
VS
149#if wxUSE_UNICODE_UTF8
150 // return true if the converter's charset is UTF-8, i.e. char* strings
151 // decoded using this object can be directly copied to wxString's internal
152 // storage without converting to WC and than back to UTF-8 MB string
153 virtual bool IsUTF8() const { return false; }
154#endif
483b0434
VZ
155
156 // The old conversion functions. The existing classes currently mostly
157 // implement these ones but we're in transition to using To/FromWChar()
158 // instead and any new classes should implement just the new functions.
159 // For now, however, we provide default implementation of To/FromWChar() in
160 // this base class in terms of MB2WC/WC2MB() to avoid having to rewrite all
161 // the conversions at once.
162 //
163 // On success, the return value is the length (i.e. the number of
164 // characters, not bytes) not counting the trailing NUL(s) of the converted
165 // string. On failure, (size_t)-1 is returned. In the special case when
166 // outputBuf is NULL the return value is the same one but nothing is
167 // written to the buffer.
168 //
169 // Note that outLen is the length of the output buffer, not the length of
170 // the input (which is always supposed to be terminated by one or more
171 // NULs, as appropriate for the encoding)!
509da451
VZ
172 virtual size_t MB2WC(wchar_t *out, const char *in, size_t outLen) const;
173 virtual size_t WC2MB(char *out, const wchar_t *in, size_t outLen) const;
483b0434
VZ
174
175
d36c9347
VZ
176 // make a heap-allocated copy of this object
177 virtual wxMBConv *Clone() const = 0;
178
7ef3ab50
VZ
179 // virtual dtor for any base class
180 virtual ~wxMBConv();
6001e347
RR
181};
182
bde4baac
VZ
183// ----------------------------------------------------------------------------
184// wxMBConvLibc uses standard mbstowcs() and wcstombs() functions for
185// conversion (hence it depends on the current locale)
186// ----------------------------------------------------------------------------
187
188class WXDLLIMPEXP_BASE wxMBConvLibc : public wxMBConv
189{
190public:
75736a9c
DS
191 virtual size_t MB2WC(wchar_t *outputBuf, const char *psz, size_t outputSize) const;
192 virtual size_t WC2MB(char *outputBuf, const wchar_t *psz, size_t outputSize) const;
d36c9347
VZ
193
194 virtual wxMBConv *Clone() const { return new wxMBConvLibc; }
111d9948
VS
195
196#if wxUSE_UNICODE_UTF8
197 virtual bool IsUTF8() const { return wxLocaleIsUtf8; }
198#endif
bde4baac
VZ
199};
200
5576edf8
RR
201#ifdef __UNIX__
202
203// ----------------------------------------------------------------------------
204// wxConvBrokenFileNames is made for Unix in Unicode mode when
205// files are accidentally written in an encoding which is not
206// the system encoding. Typically, the system encoding will be
207// UTF8 but there might be files stored in ISO8859-1 on disk.
208// ----------------------------------------------------------------------------
209
210class WXDLLIMPEXP_BASE wxConvBrokenFileNames : public wxMBConv
211{
212public:
86501081 213 wxConvBrokenFileNames(const wxString& charset);
d36c9347 214 wxConvBrokenFileNames(const wxConvBrokenFileNames& conv)
2e2cf78d
VZ
215 : wxMBConv(),
216 m_conv(conv.m_conv ? conv.m_conv->Clone() : NULL)
d36c9347
VZ
217 {
218 }
5576edf8
RR
219 virtual ~wxConvBrokenFileNames() { delete m_conv; }
220
eec47cc6
VZ
221 virtual size_t MB2WC(wchar_t *out, const char *in, size_t outLen) const
222 {
223 return m_conv->MB2WC(out, in, outLen);
224 }
225
226 virtual size_t WC2MB(char *out, const wchar_t *in, size_t outLen) const
227 {
228 return m_conv->WC2MB(out, in, outLen);
229 }
5576edf8 230
7ef3ab50 231 virtual size_t GetMBNulLen() const
eec47cc6 232 {
22886fb3 233 // cast needed to call a private function
7ef3ab50 234 return m_conv->GetMBNulLen();
eec47cc6
VZ
235 }
236
ba98e032
VS
237#if wxUSE_UNICODE_UTF8
238 virtual bool IsUTF8() const { return m_conv->IsUTF8(); }
239#endif
240
d36c9347
VZ
241 virtual wxMBConv *Clone() const { return new wxConvBrokenFileNames(*this); }
242
7ef3ab50 243private:
5576edf8
RR
244 // the conversion object we forward to
245 wxMBConv *m_conv;
d36c9347 246
c0c133e1 247 wxDECLARE_NO_ASSIGN_CLASS(wxConvBrokenFileNames);
5576edf8
RR
248};
249
eec47cc6 250#endif // __UNIX__
5576edf8 251
e90c1d2a 252// ----------------------------------------------------------------------------
6001e347 253// wxMBConvUTF7 (for conversion using UTF7 encoding)
e90c1d2a 254// ----------------------------------------------------------------------------
6001e347 255
bddd7a8d 256class WXDLLIMPEXP_BASE wxMBConvUTF7 : public wxMBConv
6001e347
RR
257{
258public:
9d653e81
VZ
259 wxMBConvUTF7() { }
260
261 // compiler-generated copy ctor, assignment operator and dtor are ok
262 // (assuming it's ok to copy the shift state -- not really sure about it)
263
264 virtual size_t ToWChar(wchar_t *dst, size_t dstLen,
265 const char *src, size_t srcLen = wxNO_LEN) const;
266 virtual size_t FromWChar(char *dst, size_t dstLen,
267 const wchar_t *src, size_t srcLen = wxNO_LEN) const;
d36c9347
VZ
268
269 virtual wxMBConv *Clone() const { return new wxMBConvUTF7; }
9d653e81
VZ
270
271private:
272 // UTF-7 decoder/encoder may be in direct mode or in shifted mode after a
273 // '+' (and until the '-' or any other non-base64 character)
1bc82105 274 struct StateMode
9d653e81 275 {
1bc82105
VZ
276 enum Mode
277 {
278 Direct, // pass through state
279 Shifted // after a '+' (and before '-')
280 };
9d653e81
VZ
281 };
282
283 // the current decoder state: this is only used by ToWChar() if srcLen
284 // parameter is not wxNO_LEN, when working on the entire NUL-terminated
285 // strings we neither update nor use the state
5c69ef61 286 class DecoderState : private StateMode
9d653e81
VZ
287 {
288 private:
289 // current state: this one is private as we want to enforce the use of
290 // ToDirect/ToShifted() methods below
291 Mode mode;
292
293 public:
294 // the initial state is direct
295 DecoderState() { mode = Direct; }
296
297 // switch to/from shifted mode
298 void ToDirect() { mode = Direct; }
299 void ToShifted() { mode = Shifted; accum = bit = 0; isLSB = false; }
300
301 bool IsDirect() const { return mode == Direct; }
302 bool IsShifted() const { return mode == Shifted; }
303
304
305 // these variables are only used in shifted mode
306
307 unsigned int accum; // accumulator of the bit we've already got
308 unsigned int bit; // the number of bits consumed mod 8
309 unsigned char msb; // the high byte of UTF-16 word
310 bool isLSB; // whether we're decoding LSB or MSB of UTF-16 word
311 };
312
313 DecoderState m_stateDecoder;
314
315
316 // encoder state is simpler as we always receive entire Unicode characters
317 // on input
5c69ef61 318 class EncoderState : private StateMode
9d653e81
VZ
319 {
320 private:
321 Mode mode;
322
323 public:
324 EncoderState() { mode = Direct; }
325
326 void ToDirect() { mode = Direct; }
327 void ToShifted() { mode = Shifted; accum = bit = 0; }
328
329 bool IsDirect() const { return mode == Direct; }
330 bool IsShifted() const { return mode == Shifted; }
331
332 unsigned int accum;
333 unsigned int bit;
334 };
335
336 EncoderState m_stateEncoder;
6001e347
RR
337};
338
e90c1d2a 339// ----------------------------------------------------------------------------
6001e347 340// wxMBConvUTF8 (for conversion using UTF8 encoding)
e90c1d2a 341// ----------------------------------------------------------------------------
6001e347 342
0286d08d
VZ
343// this is the real UTF-8 conversion class, it has to be called "strict UTF-8"
344// for compatibility reasons: the wxMBConvUTF8 class below also supports lossy
345// conversions if it is created with non default options
346class WXDLLIMPEXP_BASE wxMBConvStrictUTF8 : public wxMBConv
347{
348public:
349 // compiler-generated default ctor and other methods are ok
350
351 virtual size_t ToWChar(wchar_t *dst, size_t dstLen,
352 const char *src, size_t srcLen = wxNO_LEN) const;
353 virtual size_t FromWChar(char *dst, size_t dstLen,
354 const wchar_t *src, size_t srcLen = wxNO_LEN) const;
355
356 virtual wxMBConv *Clone() const { return new wxMBConvStrictUTF8(); }
357
358#if wxUSE_UNICODE_UTF8
359 // NB: other mapping modes are not, strictly speaking, UTF-8, so we can't
360 // take the shortcut in that case
361 virtual bool IsUTF8() const { return true; }
362#endif
363};
364
365class WXDLLIMPEXP_BASE wxMBConvUTF8 : public wxMBConvStrictUTF8
6001e347
RR
366{
367public:
d36c9347
VZ
368 enum
369 {
ea8ce907
RR
370 MAP_INVALID_UTF8_NOT = 0,
371 MAP_INVALID_UTF8_TO_PUA = 1,
372 MAP_INVALID_UTF8_TO_OCTAL = 2
373 };
374
375 wxMBConvUTF8(int options = MAP_INVALID_UTF8_NOT) : m_options(options) { }
d16d0917
VZ
376
377 virtual size_t ToWChar(wchar_t *dst, size_t dstLen,
378 const char *src, size_t srcLen = wxNO_LEN) const;
379 virtual size_t FromWChar(char *dst, size_t dstLen,
380 const wchar_t *src, size_t srcLen = wxNO_LEN) const;
eec47cc6 381
d36c9347
VZ
382 virtual wxMBConv *Clone() const { return new wxMBConvUTF8(m_options); }
383
111d9948
VS
384#if wxUSE_UNICODE_UTF8
385 // NB: other mapping modes are not, strictly speaking, UTF-8, so we can't
386 // take the shortcut in that case
387 virtual bool IsUTF8() const { return m_options == MAP_INVALID_UTF8_NOT; }
388#endif
389
ea8ce907
RR
390private:
391 int m_options;
6001e347
RR
392};
393
eec47cc6
VZ
394// ----------------------------------------------------------------------------
395// wxMBConvUTF16Base: for both LE and BE variants
396// ----------------------------------------------------------------------------
397
398class WXDLLIMPEXP_BASE wxMBConvUTF16Base : public wxMBConv
399{
7ef3ab50 400public:
467e0479
VZ
401 enum { BYTES_PER_CHAR = 2 };
402
403 virtual size_t GetMBNulLen() const { return BYTES_PER_CHAR; }
404
405protected:
406 // return the length of the buffer using srcLen if it's not wxNO_LEN and
407 // computing the length ourselves if it is; also checks that the length is
408 // even if specified as we need an entire number of UTF-16 characters and
409 // returns wxNO_LEN which indicates error if it is odd
410 static size_t GetLength(const char *src, size_t srcLen);
eec47cc6
VZ
411};
412
e90c1d2a 413// ----------------------------------------------------------------------------
c91830cb
VZ
414// wxMBConvUTF16LE (for conversion using UTF16 Little Endian encoding)
415// ----------------------------------------------------------------------------
416
eec47cc6 417class WXDLLIMPEXP_BASE wxMBConvUTF16LE : public wxMBConvUTF16Base
c91830cb
VZ
418{
419public:
467e0479
VZ
420 virtual size_t ToWChar(wchar_t *dst, size_t dstLen,
421 const char *src, size_t srcLen = wxNO_LEN) const;
422 virtual size_t FromWChar(char *dst, size_t dstLen,
423 const wchar_t *src, size_t srcLen = wxNO_LEN) const;
d36c9347 424 virtual wxMBConv *Clone() const { return new wxMBConvUTF16LE; }
c91830cb
VZ
425};
426
427// ----------------------------------------------------------------------------
428// wxMBConvUTF16BE (for conversion using UTF16 Big Endian encoding)
429// ----------------------------------------------------------------------------
430
eec47cc6 431class WXDLLIMPEXP_BASE wxMBConvUTF16BE : public wxMBConvUTF16Base
c91830cb
VZ
432{
433public:
467e0479
VZ
434 virtual size_t ToWChar(wchar_t *dst, size_t dstLen,
435 const char *src, size_t srcLen = wxNO_LEN) const;
436 virtual size_t FromWChar(char *dst, size_t dstLen,
437 const wchar_t *src, size_t srcLen = wxNO_LEN) const;
d36c9347 438 virtual wxMBConv *Clone() const { return new wxMBConvUTF16BE; }
c91830cb
VZ
439};
440
eec47cc6
VZ
441// ----------------------------------------------------------------------------
442// wxMBConvUTF32Base: base class for both LE and BE variants
443// ----------------------------------------------------------------------------
444
445class WXDLLIMPEXP_BASE wxMBConvUTF32Base : public wxMBConv
446{
7ef3ab50 447public:
467e0479
VZ
448 enum { BYTES_PER_CHAR = 4 };
449
450 virtual size_t GetMBNulLen() const { return BYTES_PER_CHAR; }
451
452protected:
453 // this is similar to wxMBConvUTF16Base method with the same name except
454 // that, of course, it verifies that length is divisible by 4 if given and
455 // not by 2
456 static size_t GetLength(const char *src, size_t srcLen);
eec47cc6
VZ
457};
458
c91830cb 459// ----------------------------------------------------------------------------
8b9e1f43 460// wxMBConvUTF32LE (for conversion using UTF32 Little Endian encoding)
c91830cb
VZ
461// ----------------------------------------------------------------------------
462
eec47cc6 463class WXDLLIMPEXP_BASE wxMBConvUTF32LE : public wxMBConvUTF32Base
c91830cb
VZ
464{
465public:
467e0479
VZ
466 virtual size_t ToWChar(wchar_t *dst, size_t dstLen,
467 const char *src, size_t srcLen = wxNO_LEN) const;
468 virtual size_t FromWChar(char *dst, size_t dstLen,
469 const wchar_t *src, size_t srcLen = wxNO_LEN) const;
d36c9347 470 virtual wxMBConv *Clone() const { return new wxMBConvUTF32LE; }
c91830cb
VZ
471};
472
473// ----------------------------------------------------------------------------
8b9e1f43 474// wxMBConvUTF32BE (for conversion using UTF32 Big Endian encoding)
c91830cb
VZ
475// ----------------------------------------------------------------------------
476
eec47cc6 477class WXDLLIMPEXP_BASE wxMBConvUTF32BE : public wxMBConvUTF32Base
c91830cb
VZ
478{
479public:
467e0479
VZ
480 virtual size_t ToWChar(wchar_t *dst, size_t dstLen,
481 const char *src, size_t srcLen = wxNO_LEN) const;
482 virtual size_t FromWChar(char *dst, size_t dstLen,
483 const wchar_t *src, size_t srcLen = wxNO_LEN) const;
d36c9347 484 virtual wxMBConv *Clone() const { return new wxMBConvUTF32BE; }
c91830cb
VZ
485};
486
487// ----------------------------------------------------------------------------
e90c1d2a
VZ
488// wxCSConv (for conversion based on loadable char sets)
489// ----------------------------------------------------------------------------
6001e347 490
8b04d4c4
VZ
491#include "wx/fontenc.h"
492
bddd7a8d 493class WXDLLIMPEXP_BASE wxCSConv : public wxMBConv
6001e347 494{
6001e347 495public:
e95354ec
VZ
496 // we can be created either from charset name or from an encoding constant
497 // but we can't have both at once
86501081 498 wxCSConv(const wxString& charset);
8b04d4c4 499 wxCSConv(wxFontEncoding encoding);
e95354ec 500
54380f29 501 wxCSConv(const wxCSConv& conv);
e90c1d2a
VZ
502 virtual ~wxCSConv();
503
54380f29 504 wxCSConv& operator=(const wxCSConv& conv);
2b5f62a0 505
1c714a5d
VZ
506 virtual size_t ToWChar(wchar_t *dst, size_t dstLen,
507 const char *src, size_t srcLen = wxNO_LEN) const;
508 virtual size_t FromWChar(char *dst, size_t dstLen,
509 const wchar_t *src, size_t srcLen = wxNO_LEN) const;
7ef3ab50 510 virtual size_t GetMBNulLen() const;
1c714a5d 511
ba98e032
VS
512#if wxUSE_UNICODE_UTF8
513 virtual bool IsUTF8() const;
514#endif
515
d36c9347 516 virtual wxMBConv *Clone() const { return new wxCSConv(*this); }
e90c1d2a 517
d36c9347 518 void Clear();
65e50848 519
a08a37d0 520 // return true if the conversion could be initialized successfully
0f0298b1 521 bool IsOk() const;
0f0298b1 522
e90c1d2a 523private:
8b04d4c4
VZ
524 // common part of all ctors
525 void Init();
526
e95354ec
VZ
527 // creates m_convReal if necessary
528 void CreateConvIfNeeded() const;
529
530 // do create m_convReal (unconditionally)
531 wxMBConv *DoCreate() const;
532
bda3d86a 533 // set the name (may be only called when m_name == NULL), makes copy of
a08a37d0 534 // the charset string
86501081 535 void SetName(const char *charset);
e90c1d2a 536
e95354ec 537
a08a37d0
VZ
538 // m_name may be NULL in which case m_encoding should be used
539 //
dccce9ea
VZ
540 // note that we can't use wxString here because of compilation
541 // dependencies: we're included from wx/string.h
86501081 542 char *m_name;
a08a37d0
VZ
543
544 // may be wxFONTENCODING_SYSTEM in which case m_name is used
545 //
546 // if m_name is NULL, then we should use the default system encoding
8b04d4c4 547 wxFontEncoding m_encoding;
e95354ec
VZ
548
549 // use CreateConvIfNeeded() before accessing m_convReal!
550 wxMBConv *m_convReal;
e90c1d2a 551 bool m_deferred;
6001e347
RR
552};
553
c3c1a9a9 554
f5a1953b
VZ
555// ----------------------------------------------------------------------------
556// declare predefined conversion objects
557// ----------------------------------------------------------------------------
d5c8817c 558
1e50d914
VS
559// Note: this macro is an implementation detail (see the comment in
560// strconv.cpp). The wxGet_XXX() and wxGet_XXXPtr() functions shouldn't be
561// used by user code and neither should XXXPtr, use the wxConvXXX macro
562// instead.
563#define WX_DECLARE_GLOBAL_CONV(klass, name) \
564 extern WXDLLIMPEXP_DATA_BASE(klass*) name##Ptr; \
092ee46f 565 extern WXDLLIMPEXP_BASE klass* wxGet_##name##Ptr(); \
1e50d914
VS
566 inline klass& wxGet_##name() \
567 { \
568 if ( !name##Ptr ) \
569 name##Ptr = wxGet_##name##Ptr(); \
570 return *name##Ptr; \
571 }
572
573
f5a1953b
VZ
574// conversion to be used with all standard functions affected by locale, e.g.
575// strtol(), strftime(), ...
1e50d914
VS
576WX_DECLARE_GLOBAL_CONV(wxMBConv, wxConvLibc)
577#define wxConvLibc wxGet_wxConvLibc()
f5a1953b
VZ
578
579// conversion ISO-8859-1/UTF-7/UTF-8 <-> wchar_t
1e50d914
VS
580WX_DECLARE_GLOBAL_CONV(wxCSConv, wxConvISO8859_1)
581#define wxConvISO8859_1 wxGet_wxConvISO8859_1()
582
0286d08d 583WX_DECLARE_GLOBAL_CONV(wxMBConvStrictUTF8, wxConvUTF8)
1e50d914
VS
584#define wxConvUTF8 wxGet_wxConvUTF8()
585
586WX_DECLARE_GLOBAL_CONV(wxMBConvUTF7, wxConvUTF7)
587#define wxConvUTF7 wxGet_wxConvUTF7()
f5a1953b
VZ
588
589// conversion used for the file names on the systems where they're not Unicode
590// (basically anything except Windows)
591//
592// this is used by all file functions, can be changed by the application
593//
594// by default UTF-8 under Mac OS X and wxConvLibc elsewhere (but it's not used
595// under Windows normally)
596extern WXDLLIMPEXP_DATA_BASE(wxMBConv *) wxConvFileName;
597
598// backwards compatible define
599#define wxConvFile (*wxConvFileName)
600
601// the current conversion object, may be set to any conversion, is used by
602// default in a couple of places inside wx (initially same as wxConvLibc)
16cba29d 603extern WXDLLIMPEXP_DATA_BASE(wxMBConv *) wxConvCurrent;
6001e347 604
0e052272 605// the conversion corresponding to the current locale
1e50d914
VS
606WX_DECLARE_GLOBAL_CONV(wxCSConv, wxConvLocal)
607#define wxConvLocal wxGet_wxConvLocal()
f5a1953b 608
d5bef0a3
VZ
609// the conversion corresponding to the encoding of the standard UI elements
610//
611// by default this is the same as wxConvLocal but may be changed if the program
612// needs to use a fixed encoding
613extern WXDLLIMPEXP_DATA_BASE(wxMBConv *) wxConvUI;
614
1e50d914
VS
615#undef WX_DECLARE_GLOBAL_CONV
616
e95354ec
VZ
617// ----------------------------------------------------------------------------
618// endianness-dependent conversions
619// ----------------------------------------------------------------------------
620
621#ifdef WORDS_BIGENDIAN
622 typedef wxMBConvUTF16BE wxMBConvUTF16;
623 typedef wxMBConvUTF32BE wxMBConvUTF32;
624#else
625 typedef wxMBConvUTF16LE wxMBConvUTF16;
626 typedef wxMBConvUTF32LE wxMBConvUTF32;
627#endif
628
e90c1d2a 629// ----------------------------------------------------------------------------
6001e347 630// filename conversion macros
e90c1d2a 631// ----------------------------------------------------------------------------
6001e347 632
bc4b4779 633// filenames are multibyte on Unix and widechar on Windows
80df4d31 634#if wxMBFILES && wxUSE_UNICODE
f5a1953b 635 #define wxFNCONV(name) wxConvFileName->cWX2MB(name)
e90c1d2a 636 #define wxFNSTRINGCAST wxMBSTRINGCAST
d5c8817c 637#else
0b6a49c2 638#if defined( __WXOSX_OR_COCOA__ ) && wxMBFILES
f5a1953b 639 #define wxFNCONV(name) wxConvFileName->cWC2MB( wxConvLocal.cWX2WC(name) )
6001e347 640#else
e90c1d2a 641 #define wxFNCONV(name) name
d5c8817c 642#endif
e90c1d2a 643 #define wxFNSTRINGCAST WXSTRINGCAST
6001e347
RR
644#endif
645
f5a1953b 646#else // !wxUSE_WCHAR_T
6001e347 647
e90c1d2a 648// ----------------------------------------------------------------------------
6001e347 649// stand-ins in absence of wchar_t
e90c1d2a 650// ----------------------------------------------------------------------------
6001e347 651
bddd7a8d 652class WXDLLIMPEXP_BASE wxMBConv
6001e347
RR
653{
654public:
e90c1d2a
VZ
655 const char* cMB2WX(const char *psz) const { return psz; }
656 const char* cWX2MB(const char *psz) const { return psz; }
6001e347 657};
e90c1d2a 658
bde4baac
VZ
659#define wxConvFile wxConvLocal
660
16cba29d 661extern WXDLLIMPEXP_DATA_BASE(wxMBConv) wxConvLibc,
8b04d4c4
VZ
662 wxConvLocal,
663 wxConvISO8859_1,
664 wxConvUTF8;
16cba29d 665extern WXDLLIMPEXP_DATA_BASE(wxMBConv *) wxConvCurrent;
6001e347
RR
666
667#define wxFNCONV(name) name
e90c1d2a 668#define wxFNSTRINGCAST WXSTRINGCAST
6001e347
RR
669
670#endif
671 // wxUSE_WCHAR_T
672
e90c1d2a
VZ
673// ----------------------------------------------------------------------------
674// macros for the most common conversions
675// ----------------------------------------------------------------------------
676
677#if wxUSE_UNICODE
678 #define wxConvertWX2MB(s) wxConvCurrent->cWX2MB(s)
679 #define wxConvertMB2WX(s) wxConvCurrent->cMB2WX(s)
69c928ef
VZ
680
681 // these functions should be used when the conversions really, really have
682 // to succeed (usually because we pass their results to a standard C
683 // function which would crash if we passed NULL to it), so these functions
684 // always return a valid pointer if their argument is non-NULL
685
686 // this function safety is achieved by trying wxConvLibc first, wxConvUTF8
687 // next if it fails and, finally, wxConvISO8859_1 which always succeeds
688 extern WXDLLIMPEXP_BASE wxWCharBuffer wxSafeConvertMB2WX(const char *s);
689
690 // this function uses wxConvLibc and wxConvUTF8(MAP_INVALID_UTF8_TO_OCTAL)
691 // if it fails
692 extern WXDLLIMPEXP_BASE wxCharBuffer wxSafeConvertWX2MB(const wchar_t *ws);
e90c1d2a
VZ
693#else // ANSI
694 // no conversions to do
695 #define wxConvertWX2MB(s) (s)
696 #define wxConvertMB2WX(s) (s)
69c928ef
VZ
697 #define wxSafeConvertMB2WX(s) (s)
698 #define wxSafeConvertWX2MB(s) (s)
e90c1d2a
VZ
699#endif // Unicode/ANSI
700
d36c9347 701#endif // _WX_STRCONV_H_
6001e347 702