]>
Commit | Line | Data |
---|---|---|
1 | /////////////////////////////////////////////////////////////////////////////// | |
2 | // Name: strconv.h | |
3 | // Purpose: conversion routines for char sets any Unicode | |
4 | // Author: Ove Kaaven, Robert Roebling, Vadim Zeitlin | |
5 | // Modified by: | |
6 | // Created: 29/01/98 | |
7 | // RCS-ID: $Id$ | |
8 | // Copyright: (c) 1998 Ove Kaaven, Robert Roebling | |
9 | // (c) 1998-2006 Vadim Zeitlin | |
10 | // Licence: wxWindows licence | |
11 | /////////////////////////////////////////////////////////////////////////////// | |
12 | ||
13 | #ifndef _WX_STRCONV_H_ | |
14 | #define _WX_STRCONV_H_ | |
15 | ||
16 | #include "wx/defs.h" | |
17 | #include "wx/chartype.h" | |
18 | #include "wx/buffer.h" | |
19 | ||
20 | #ifdef __DIGITALMARS__ | |
21 | #include "typeinfo.h" | |
22 | #endif | |
23 | ||
24 | #if defined(__VISAGECPP__) && __IBMCPP__ >= 400 | |
25 | # undef __BSEXCPT__ | |
26 | #endif | |
27 | ||
28 | #ifndef __WXPALMOS5__ | |
29 | #include <stdlib.h> | |
30 | #endif // ! __WXPALMOS5__ | |
31 | ||
32 | #if wxUSE_WCHAR_T | |
33 | ||
34 | class WXDLLIMPEXP_FWD_BASE wxString; | |
35 | ||
36 | // the error value returned by wxMBConv methods | |
37 | #define wxCONV_FAILED ((size_t)-1) | |
38 | ||
39 | // the default value for some length parameters meaning that the string is | |
40 | // NUL-terminated | |
41 | #define wxNO_LEN ((size_t)-1) | |
42 | ||
43 | // ---------------------------------------------------------------------------- | |
44 | // wxMBConv (abstract base class for conversions) | |
45 | // ---------------------------------------------------------------------------- | |
46 | ||
47 | // When deriving a new class from wxMBConv you must reimplement ToWChar() and | |
48 | // FromWChar() methods which are not pure virtual only for historical reasons, | |
49 | // don't let the fact that the existing classes implement MB2WC/WC2MB() instead | |
50 | // confuse you. | |
51 | // | |
52 | // You also have to implement Clone() to allow copying the conversions | |
53 | // polymorphically. | |
54 | // | |
55 | // And you might need to override GetMBNulLen() as well. | |
56 | class WXDLLIMPEXP_BASE wxMBConv | |
57 | { | |
58 | public: | |
59 | // The functions doing actual conversion from/to narrow to/from wide | |
60 | // character strings. | |
61 | // | |
62 | // On success, the return value is the length (i.e. the number of | |
63 | // characters, not bytes) of the converted string including any trailing | |
64 | // L'\0' or (possibly multiple) '\0'(s). If the conversion fails or if | |
65 | // there is not enough space for everything, including the trailing NUL | |
66 | // character(s), in the output buffer, wxCONV_FAILED is returned. | |
67 | // | |
68 | // In the special case when dstLen is 0 (outputBuf may be NULL then) the | |
69 | // return value is the length of the needed buffer but nothing happens | |
70 | // otherwise. If srcLen is wxNO_LEN, the entire string, up to and | |
71 | // including the trailing NUL(s), is converted, otherwise exactly srcLen | |
72 | // bytes are. | |
73 | // | |
74 | // Typical usage: | |
75 | // | |
76 | // size_t dstLen = conv.ToWChar(NULL, 0, src); | |
77 | // if ( dstLen != wxCONV_FAILED ) | |
78 | // ... handle error ... | |
79 | // wchar_t *wbuf = new wchar_t[dstLen]; | |
80 | // conv.ToWChar(wbuf, dstLen, src); | |
81 | // | |
82 | virtual size_t ToWChar(wchar_t *dst, size_t dstLen, | |
83 | const char *src, size_t srcLen = wxNO_LEN) const; | |
84 | ||
85 | virtual size_t FromWChar(char *dst, size_t dstLen, | |
86 | const wchar_t *src, size_t srcLen = wxNO_LEN) const; | |
87 | ||
88 | ||
89 | // Convenience functions for translating NUL-terminated strings: returns | |
90 | // the buffer containing the converted string or NULL pointer if the | |
91 | // conversion failed. | |
92 | const wxWCharBuffer cMB2WC(const char *in) const; | |
93 | const wxCharBuffer cWC2MB(const wchar_t *in) const; | |
94 | ||
95 | // Convenience functions for converting strings which may contain embedded | |
96 | // NULs and don't have to be NUL-terminated. | |
97 | // | |
98 | // inLen is the length of the buffer including trailing NUL if any: if the | |
99 | // last 4 bytes of the buffer are all NULs, these functions are more | |
100 | // efficient as they avoid copying the string, but otherwise a copy is made | |
101 | // internally which could be quite bad for (very) long strings. | |
102 | // | |
103 | // outLen receives, if not NULL, the length of the converted string or 0 if | |
104 | // the conversion failed (returning 0 and not -1 in this case makes it | |
105 | // difficult to distinguish between failed conversion and empty input but | |
106 | // this is done for backwards compatibility) | |
107 | const wxWCharBuffer | |
108 | cMB2WC(const char *in, size_t inLen, size_t *outLen) const; | |
109 | const wxCharBuffer | |
110 | cWC2MB(const wchar_t *in, size_t inLen, size_t *outLen) const; | |
111 | ||
112 | // convenience functions for converting MB or WC to/from wxWin default | |
113 | #if wxUSE_UNICODE | |
114 | const wxWCharBuffer cMB2WX(const char *psz) const { return cMB2WC(psz); } | |
115 | const wxCharBuffer cWX2MB(const wchar_t *psz) const { return cWC2MB(psz); } | |
116 | const wchar_t* cWC2WX(const wchar_t *psz) const { return psz; } | |
117 | const wchar_t* cWX2WC(const wchar_t *psz) const { return psz; } | |
118 | #else // ANSI | |
119 | const char* cMB2WX(const char *psz) const { return psz; } | |
120 | const char* cWX2MB(const char *psz) const { return psz; } | |
121 | const wxCharBuffer cWC2WX(const wchar_t *psz) const { return cWC2MB(psz); } | |
122 | const wxWCharBuffer cWX2WC(const char *psz) const { return cMB2WC(psz); } | |
123 | #endif // Unicode/ANSI | |
124 | ||
125 | // this function is used in the implementation of cMB2WC() to distinguish | |
126 | // between the following cases: | |
127 | // | |
128 | // a) var width encoding with strings terminated by a single NUL | |
129 | // (usual multibyte encodings): return 1 in this case | |
130 | // b) fixed width encoding with 2 bytes/char and so terminated by | |
131 | // 2 NULs (UTF-16/UCS-2 and variants): return 2 in this case | |
132 | // c) fixed width encoding with 4 bytes/char and so terminated by | |
133 | // 4 NULs (UTF-32/UCS-4 and variants): return 4 in this case | |
134 | // | |
135 | // anything else is not supported currently and -1 should be returned | |
136 | virtual size_t GetMBNulLen() const { return 1; } | |
137 | ||
138 | // return the maximal value currently returned by GetMBNulLen() for any | |
139 | // encoding | |
140 | static size_t GetMaxMBNulLen() { return 4 /* for UTF-32 */; } | |
141 | ||
142 | #if wxUSE_UNICODE_UTF8 | |
143 | // return true if the converter's charset is UTF-8, i.e. char* strings | |
144 | // decoded using this object can be directly copied to wxString's internal | |
145 | // storage without converting to WC and than back to UTF-8 MB string | |
146 | virtual bool IsUTF8() const { return false; } | |
147 | #endif | |
148 | ||
149 | // The old conversion functions. The existing classes currently mostly | |
150 | // implement these ones but we're in transition to using To/FromWChar() | |
151 | // instead and any new classes should implement just the new functions. | |
152 | // For now, however, we provide default implementation of To/FromWChar() in | |
153 | // this base class in terms of MB2WC/WC2MB() to avoid having to rewrite all | |
154 | // the conversions at once. | |
155 | // | |
156 | // On success, the return value is the length (i.e. the number of | |
157 | // characters, not bytes) not counting the trailing NUL(s) of the converted | |
158 | // string. On failure, (size_t)-1 is returned. In the special case when | |
159 | // outputBuf is NULL the return value is the same one but nothing is | |
160 | // written to the buffer. | |
161 | // | |
162 | // Note that outLen is the length of the output buffer, not the length of | |
163 | // the input (which is always supposed to be terminated by one or more | |
164 | // NULs, as appropriate for the encoding)! | |
165 | virtual size_t MB2WC(wchar_t *out, const char *in, size_t outLen) const; | |
166 | virtual size_t WC2MB(char *out, const wchar_t *in, size_t outLen) const; | |
167 | ||
168 | ||
169 | // make a heap-allocated copy of this object | |
170 | virtual wxMBConv *Clone() const = 0; | |
171 | ||
172 | // virtual dtor for any base class | |
173 | virtual ~wxMBConv(); | |
174 | }; | |
175 | ||
176 | // ---------------------------------------------------------------------------- | |
177 | // wxMBConvLibc uses standard mbstowcs() and wcstombs() functions for | |
178 | // conversion (hence it depends on the current locale) | |
179 | // ---------------------------------------------------------------------------- | |
180 | ||
181 | class WXDLLIMPEXP_BASE wxMBConvLibc : public wxMBConv | |
182 | { | |
183 | public: | |
184 | virtual size_t MB2WC(wchar_t *outputBuf, const char *psz, size_t outputSize) const; | |
185 | virtual size_t WC2MB(char *outputBuf, const wchar_t *psz, size_t outputSize) const; | |
186 | ||
187 | virtual wxMBConv *Clone() const { return new wxMBConvLibc; } | |
188 | ||
189 | #if wxUSE_UNICODE_UTF8 | |
190 | virtual bool IsUTF8() const { return wxLocaleIsUtf8; } | |
191 | #endif | |
192 | }; | |
193 | ||
194 | #ifdef __UNIX__ | |
195 | ||
196 | // ---------------------------------------------------------------------------- | |
197 | // wxConvBrokenFileNames is made for Unix in Unicode mode when | |
198 | // files are accidentally written in an encoding which is not | |
199 | // the system encoding. Typically, the system encoding will be | |
200 | // UTF8 but there might be files stored in ISO8859-1 on disk. | |
201 | // ---------------------------------------------------------------------------- | |
202 | ||
203 | class WXDLLIMPEXP_BASE wxConvBrokenFileNames : public wxMBConv | |
204 | { | |
205 | public: | |
206 | wxConvBrokenFileNames(const wxString& charset); | |
207 | wxConvBrokenFileNames(const wxConvBrokenFileNames& conv) | |
208 | : wxMBConv(), | |
209 | m_conv(conv.m_conv ? conv.m_conv->Clone() : NULL) | |
210 | { | |
211 | } | |
212 | virtual ~wxConvBrokenFileNames() { delete m_conv; } | |
213 | ||
214 | virtual size_t MB2WC(wchar_t *out, const char *in, size_t outLen) const | |
215 | { | |
216 | return m_conv->MB2WC(out, in, outLen); | |
217 | } | |
218 | ||
219 | virtual size_t WC2MB(char *out, const wchar_t *in, size_t outLen) const | |
220 | { | |
221 | return m_conv->WC2MB(out, in, outLen); | |
222 | } | |
223 | ||
224 | virtual size_t GetMBNulLen() const | |
225 | { | |
226 | // cast needed to call a private function | |
227 | return m_conv->GetMBNulLen(); | |
228 | } | |
229 | ||
230 | #if wxUSE_UNICODE_UTF8 | |
231 | virtual bool IsUTF8() const { return m_conv->IsUTF8(); } | |
232 | #endif | |
233 | ||
234 | virtual wxMBConv *Clone() const { return new wxConvBrokenFileNames(*this); } | |
235 | ||
236 | private: | |
237 | // the conversion object we forward to | |
238 | wxMBConv *m_conv; | |
239 | ||
240 | DECLARE_NO_ASSIGN_CLASS(wxConvBrokenFileNames) | |
241 | }; | |
242 | ||
243 | #endif // __UNIX__ | |
244 | ||
245 | // ---------------------------------------------------------------------------- | |
246 | // wxMBConvUTF7 (for conversion using UTF7 encoding) | |
247 | // ---------------------------------------------------------------------------- | |
248 | ||
249 | class WXDLLIMPEXP_BASE wxMBConvUTF7 : public wxMBConv | |
250 | { | |
251 | public: | |
252 | wxMBConvUTF7() { } | |
253 | ||
254 | // compiler-generated copy ctor, assignment operator and dtor are ok | |
255 | // (assuming it's ok to copy the shift state -- not really sure about it) | |
256 | ||
257 | virtual size_t ToWChar(wchar_t *dst, size_t dstLen, | |
258 | const char *src, size_t srcLen = wxNO_LEN) const; | |
259 | virtual size_t FromWChar(char *dst, size_t dstLen, | |
260 | const wchar_t *src, size_t srcLen = wxNO_LEN) const; | |
261 | ||
262 | virtual wxMBConv *Clone() const { return new wxMBConvUTF7; } | |
263 | ||
264 | private: | |
265 | // UTF-7 decoder/encoder may be in direct mode or in shifted mode after a | |
266 | // '+' (and until the '-' or any other non-base64 character) | |
267 | struct StateMode | |
268 | { | |
269 | enum Mode | |
270 | { | |
271 | Direct, // pass through state | |
272 | Shifted // after a '+' (and before '-') | |
273 | }; | |
274 | }; | |
275 | ||
276 | // the current decoder state: this is only used by ToWChar() if srcLen | |
277 | // parameter is not wxNO_LEN, when working on the entire NUL-terminated | |
278 | // strings we neither update nor use the state | |
279 | class DecoderState : StateMode | |
280 | { | |
281 | private: | |
282 | // current state: this one is private as we want to enforce the use of | |
283 | // ToDirect/ToShifted() methods below | |
284 | Mode mode; | |
285 | ||
286 | public: | |
287 | // the initial state is direct | |
288 | DecoderState() { mode = Direct; } | |
289 | ||
290 | // switch to/from shifted mode | |
291 | void ToDirect() { mode = Direct; } | |
292 | void ToShifted() { mode = Shifted; accum = bit = 0; isLSB = false; } | |
293 | ||
294 | bool IsDirect() const { return mode == Direct; } | |
295 | bool IsShifted() const { return mode == Shifted; } | |
296 | ||
297 | ||
298 | // these variables are only used in shifted mode | |
299 | ||
300 | unsigned int accum; // accumulator of the bit we've already got | |
301 | unsigned int bit; // the number of bits consumed mod 8 | |
302 | unsigned char msb; // the high byte of UTF-16 word | |
303 | bool isLSB; // whether we're decoding LSB or MSB of UTF-16 word | |
304 | }; | |
305 | ||
306 | DecoderState m_stateDecoder; | |
307 | ||
308 | ||
309 | // encoder state is simpler as we always receive entire Unicode characters | |
310 | // on input | |
311 | class EncoderState : StateMode | |
312 | { | |
313 | private: | |
314 | Mode mode; | |
315 | ||
316 | public: | |
317 | EncoderState() { mode = Direct; } | |
318 | ||
319 | void ToDirect() { mode = Direct; } | |
320 | void ToShifted() { mode = Shifted; accum = bit = 0; } | |
321 | ||
322 | bool IsDirect() const { return mode == Direct; } | |
323 | bool IsShifted() const { return mode == Shifted; } | |
324 | ||
325 | unsigned int accum; | |
326 | unsigned int bit; | |
327 | }; | |
328 | ||
329 | EncoderState m_stateEncoder; | |
330 | }; | |
331 | ||
332 | // ---------------------------------------------------------------------------- | |
333 | // wxMBConvUTF8 (for conversion using UTF8 encoding) | |
334 | // ---------------------------------------------------------------------------- | |
335 | ||
336 | // this is the real UTF-8 conversion class, it has to be called "strict UTF-8" | |
337 | // for compatibility reasons: the wxMBConvUTF8 class below also supports lossy | |
338 | // conversions if it is created with non default options | |
339 | class WXDLLIMPEXP_BASE wxMBConvStrictUTF8 : public wxMBConv | |
340 | { | |
341 | public: | |
342 | // compiler-generated default ctor and other methods are ok | |
343 | ||
344 | virtual size_t ToWChar(wchar_t *dst, size_t dstLen, | |
345 | const char *src, size_t srcLen = wxNO_LEN) const; | |
346 | virtual size_t FromWChar(char *dst, size_t dstLen, | |
347 | const wchar_t *src, size_t srcLen = wxNO_LEN) const; | |
348 | ||
349 | virtual wxMBConv *Clone() const { return new wxMBConvStrictUTF8(); } | |
350 | ||
351 | #if wxUSE_UNICODE_UTF8 | |
352 | // NB: other mapping modes are not, strictly speaking, UTF-8, so we can't | |
353 | // take the shortcut in that case | |
354 | virtual bool IsUTF8() const { return true; } | |
355 | #endif | |
356 | }; | |
357 | ||
358 | class WXDLLIMPEXP_BASE wxMBConvUTF8 : public wxMBConvStrictUTF8 | |
359 | { | |
360 | public: | |
361 | enum | |
362 | { | |
363 | MAP_INVALID_UTF8_NOT = 0, | |
364 | MAP_INVALID_UTF8_TO_PUA = 1, | |
365 | MAP_INVALID_UTF8_TO_OCTAL = 2 | |
366 | }; | |
367 | ||
368 | wxMBConvUTF8(int options = MAP_INVALID_UTF8_NOT) : m_options(options) { } | |
369 | ||
370 | virtual size_t ToWChar(wchar_t *dst, size_t dstLen, | |
371 | const char *src, size_t srcLen = wxNO_LEN) const; | |
372 | virtual size_t FromWChar(char *dst, size_t dstLen, | |
373 | const wchar_t *src, size_t srcLen = wxNO_LEN) const; | |
374 | ||
375 | virtual wxMBConv *Clone() const { return new wxMBConvUTF8(m_options); } | |
376 | ||
377 | #if wxUSE_UNICODE_UTF8 | |
378 | // NB: other mapping modes are not, strictly speaking, UTF-8, so we can't | |
379 | // take the shortcut in that case | |
380 | virtual bool IsUTF8() const { return m_options == MAP_INVALID_UTF8_NOT; } | |
381 | #endif | |
382 | ||
383 | private: | |
384 | int m_options; | |
385 | }; | |
386 | ||
387 | // ---------------------------------------------------------------------------- | |
388 | // wxMBConvUTF16Base: for both LE and BE variants | |
389 | // ---------------------------------------------------------------------------- | |
390 | ||
391 | class WXDLLIMPEXP_BASE wxMBConvUTF16Base : public wxMBConv | |
392 | { | |
393 | public: | |
394 | enum { BYTES_PER_CHAR = 2 }; | |
395 | ||
396 | virtual size_t GetMBNulLen() const { return BYTES_PER_CHAR; } | |
397 | ||
398 | protected: | |
399 | // return the length of the buffer using srcLen if it's not wxNO_LEN and | |
400 | // computing the length ourselves if it is; also checks that the length is | |
401 | // even if specified as we need an entire number of UTF-16 characters and | |
402 | // returns wxNO_LEN which indicates error if it is odd | |
403 | static size_t GetLength(const char *src, size_t srcLen); | |
404 | }; | |
405 | ||
406 | // ---------------------------------------------------------------------------- | |
407 | // wxMBConvUTF16LE (for conversion using UTF16 Little Endian encoding) | |
408 | // ---------------------------------------------------------------------------- | |
409 | ||
410 | class WXDLLIMPEXP_BASE wxMBConvUTF16LE : public wxMBConvUTF16Base | |
411 | { | |
412 | public: | |
413 | virtual size_t ToWChar(wchar_t *dst, size_t dstLen, | |
414 | const char *src, size_t srcLen = wxNO_LEN) const; | |
415 | virtual size_t FromWChar(char *dst, size_t dstLen, | |
416 | const wchar_t *src, size_t srcLen = wxNO_LEN) const; | |
417 | virtual wxMBConv *Clone() const { return new wxMBConvUTF16LE; } | |
418 | }; | |
419 | ||
420 | // ---------------------------------------------------------------------------- | |
421 | // wxMBConvUTF16BE (for conversion using UTF16 Big Endian encoding) | |
422 | // ---------------------------------------------------------------------------- | |
423 | ||
424 | class WXDLLIMPEXP_BASE wxMBConvUTF16BE : public wxMBConvUTF16Base | |
425 | { | |
426 | public: | |
427 | virtual size_t ToWChar(wchar_t *dst, size_t dstLen, | |
428 | const char *src, size_t srcLen = wxNO_LEN) const; | |
429 | virtual size_t FromWChar(char *dst, size_t dstLen, | |
430 | const wchar_t *src, size_t srcLen = wxNO_LEN) const; | |
431 | virtual wxMBConv *Clone() const { return new wxMBConvUTF16BE; } | |
432 | }; | |
433 | ||
434 | // ---------------------------------------------------------------------------- | |
435 | // wxMBConvUTF32Base: base class for both LE and BE variants | |
436 | // ---------------------------------------------------------------------------- | |
437 | ||
438 | class WXDLLIMPEXP_BASE wxMBConvUTF32Base : public wxMBConv | |
439 | { | |
440 | public: | |
441 | enum { BYTES_PER_CHAR = 4 }; | |
442 | ||
443 | virtual size_t GetMBNulLen() const { return BYTES_PER_CHAR; } | |
444 | ||
445 | protected: | |
446 | // this is similar to wxMBConvUTF16Base method with the same name except | |
447 | // that, of course, it verifies that length is divisible by 4 if given and | |
448 | // not by 2 | |
449 | static size_t GetLength(const char *src, size_t srcLen); | |
450 | }; | |
451 | ||
452 | // ---------------------------------------------------------------------------- | |
453 | // wxMBConvUTF32LE (for conversion using UTF32 Little Endian encoding) | |
454 | // ---------------------------------------------------------------------------- | |
455 | ||
456 | class WXDLLIMPEXP_BASE wxMBConvUTF32LE : public wxMBConvUTF32Base | |
457 | { | |
458 | public: | |
459 | virtual size_t ToWChar(wchar_t *dst, size_t dstLen, | |
460 | const char *src, size_t srcLen = wxNO_LEN) const; | |
461 | virtual size_t FromWChar(char *dst, size_t dstLen, | |
462 | const wchar_t *src, size_t srcLen = wxNO_LEN) const; | |
463 | virtual wxMBConv *Clone() const { return new wxMBConvUTF32LE; } | |
464 | }; | |
465 | ||
466 | // ---------------------------------------------------------------------------- | |
467 | // wxMBConvUTF32BE (for conversion using UTF32 Big Endian encoding) | |
468 | // ---------------------------------------------------------------------------- | |
469 | ||
470 | class WXDLLIMPEXP_BASE wxMBConvUTF32BE : public wxMBConvUTF32Base | |
471 | { | |
472 | public: | |
473 | virtual size_t ToWChar(wchar_t *dst, size_t dstLen, | |
474 | const char *src, size_t srcLen = wxNO_LEN) const; | |
475 | virtual size_t FromWChar(char *dst, size_t dstLen, | |
476 | const wchar_t *src, size_t srcLen = wxNO_LEN) const; | |
477 | virtual wxMBConv *Clone() const { return new wxMBConvUTF32BE; } | |
478 | }; | |
479 | ||
480 | // ---------------------------------------------------------------------------- | |
481 | // wxCSConv (for conversion based on loadable char sets) | |
482 | // ---------------------------------------------------------------------------- | |
483 | ||
484 | #include "wx/fontenc.h" | |
485 | ||
486 | class WXDLLIMPEXP_BASE wxCSConv : public wxMBConv | |
487 | { | |
488 | public: | |
489 | // we can be created either from charset name or from an encoding constant | |
490 | // but we can't have both at once | |
491 | wxCSConv(const wxString& charset); | |
492 | wxCSConv(wxFontEncoding encoding); | |
493 | ||
494 | wxCSConv(const wxCSConv& conv); | |
495 | virtual ~wxCSConv(); | |
496 | ||
497 | wxCSConv& operator=(const wxCSConv& conv); | |
498 | ||
499 | virtual size_t ToWChar(wchar_t *dst, size_t dstLen, | |
500 | const char *src, size_t srcLen = wxNO_LEN) const; | |
501 | virtual size_t FromWChar(char *dst, size_t dstLen, | |
502 | const wchar_t *src, size_t srcLen = wxNO_LEN) const; | |
503 | virtual size_t MB2WC(wchar_t *outputBuf, const char *psz, size_t outputSize) const; | |
504 | virtual size_t WC2MB(char *outputBuf, const wchar_t *psz, size_t outputSize) const; | |
505 | virtual size_t GetMBNulLen() const; | |
506 | ||
507 | #if wxUSE_UNICODE_UTF8 | |
508 | virtual bool IsUTF8() const; | |
509 | #endif | |
510 | ||
511 | virtual wxMBConv *Clone() const { return new wxCSConv(*this); } | |
512 | ||
513 | void Clear(); | |
514 | ||
515 | // return true if the conversion could be initilized successfully | |
516 | bool IsOk() const; | |
517 | ||
518 | private: | |
519 | // common part of all ctors | |
520 | void Init(); | |
521 | ||
522 | // creates m_convReal if necessary | |
523 | void CreateConvIfNeeded() const; | |
524 | ||
525 | // do create m_convReal (unconditionally) | |
526 | wxMBConv *DoCreate() const; | |
527 | ||
528 | // set the name (may be only called when m_name == NULL), makes copy of | |
529 | // charset string | |
530 | void SetName(const char *charset); | |
531 | ||
532 | ||
533 | // note that we can't use wxString here because of compilation | |
534 | // dependencies: we're included from wx/string.h | |
535 | char *m_name; | |
536 | wxFontEncoding m_encoding; | |
537 | ||
538 | // use CreateConvIfNeeded() before accessing m_convReal! | |
539 | wxMBConv *m_convReal; | |
540 | bool m_deferred; | |
541 | }; | |
542 | ||
543 | ||
544 | // ---------------------------------------------------------------------------- | |
545 | // declare predefined conversion objects | |
546 | // ---------------------------------------------------------------------------- | |
547 | ||
548 | // Note: this macro is an implementation detail (see the comment in | |
549 | // strconv.cpp). The wxGet_XXX() and wxGet_XXXPtr() functions shouldn't be | |
550 | // used by user code and neither should XXXPtr, use the wxConvXXX macro | |
551 | // instead. | |
552 | #define WX_DECLARE_GLOBAL_CONV(klass, name) \ | |
553 | extern WXDLLIMPEXP_DATA_BASE(klass*) name##Ptr; \ | |
554 | extern WXDLLIMPEXP_BASE klass* wxGet_##name##Ptr(); \ | |
555 | inline klass& wxGet_##name() \ | |
556 | { \ | |
557 | if ( !name##Ptr ) \ | |
558 | name##Ptr = wxGet_##name##Ptr(); \ | |
559 | return *name##Ptr; \ | |
560 | } | |
561 | ||
562 | ||
563 | // conversion to be used with all standard functions affected by locale, e.g. | |
564 | // strtol(), strftime(), ... | |
565 | WX_DECLARE_GLOBAL_CONV(wxMBConv, wxConvLibc) | |
566 | #define wxConvLibc wxGet_wxConvLibc() | |
567 | ||
568 | // conversion ISO-8859-1/UTF-7/UTF-8 <-> wchar_t | |
569 | WX_DECLARE_GLOBAL_CONV(wxCSConv, wxConvISO8859_1) | |
570 | #define wxConvISO8859_1 wxGet_wxConvISO8859_1() | |
571 | ||
572 | WX_DECLARE_GLOBAL_CONV(wxMBConvStrictUTF8, wxConvUTF8) | |
573 | #define wxConvUTF8 wxGet_wxConvUTF8() | |
574 | ||
575 | WX_DECLARE_GLOBAL_CONV(wxMBConvUTF7, wxConvUTF7) | |
576 | #define wxConvUTF7 wxGet_wxConvUTF7() | |
577 | ||
578 | // conversion used for the file names on the systems where they're not Unicode | |
579 | // (basically anything except Windows) | |
580 | // | |
581 | // this is used by all file functions, can be changed by the application | |
582 | // | |
583 | // by default UTF-8 under Mac OS X and wxConvLibc elsewhere (but it's not used | |
584 | // under Windows normally) | |
585 | extern WXDLLIMPEXP_DATA_BASE(wxMBConv *) wxConvFileName; | |
586 | ||
587 | // backwards compatible define | |
588 | #define wxConvFile (*wxConvFileName) | |
589 | ||
590 | // the current conversion object, may be set to any conversion, is used by | |
591 | // default in a couple of places inside wx (initially same as wxConvLibc) | |
592 | extern WXDLLIMPEXP_DATA_BASE(wxMBConv *) wxConvCurrent; | |
593 | ||
594 | // the conversion corresponding to the current locale | |
595 | WX_DECLARE_GLOBAL_CONV(wxCSConv, wxConvLocal) | |
596 | #define wxConvLocal wxGet_wxConvLocal() | |
597 | ||
598 | // the conversion corresponding to the encoding of the standard UI elements | |
599 | // | |
600 | // by default this is the same as wxConvLocal but may be changed if the program | |
601 | // needs to use a fixed encoding | |
602 | extern WXDLLIMPEXP_DATA_BASE(wxMBConv *) wxConvUI; | |
603 | ||
604 | #undef WX_DECLARE_GLOBAL_CONV | |
605 | ||
606 | // ---------------------------------------------------------------------------- | |
607 | // endianness-dependent conversions | |
608 | // ---------------------------------------------------------------------------- | |
609 | ||
610 | #ifdef WORDS_BIGENDIAN | |
611 | typedef wxMBConvUTF16BE wxMBConvUTF16; | |
612 | typedef wxMBConvUTF32BE wxMBConvUTF32; | |
613 | #else | |
614 | typedef wxMBConvUTF16LE wxMBConvUTF16; | |
615 | typedef wxMBConvUTF32LE wxMBConvUTF32; | |
616 | #endif | |
617 | ||
618 | // ---------------------------------------------------------------------------- | |
619 | // filename conversion macros | |
620 | // ---------------------------------------------------------------------------- | |
621 | ||
622 | // filenames are multibyte on Unix and widechar on Windows | |
623 | #if wxMBFILES && wxUSE_UNICODE | |
624 | #define wxFNCONV(name) wxConvFileName->cWX2MB(name) | |
625 | #define wxFNSTRINGCAST wxMBSTRINGCAST | |
626 | #else | |
627 | #if defined( __WXOSX__ ) && wxMBFILES | |
628 | #define wxFNCONV(name) wxConvFileName->cWC2MB( wxConvLocal.cWX2WC(name) ) | |
629 | #else | |
630 | #define wxFNCONV(name) name | |
631 | #endif | |
632 | #define wxFNSTRINGCAST WXSTRINGCAST | |
633 | #endif | |
634 | ||
635 | #else // !wxUSE_WCHAR_T | |
636 | ||
637 | // ---------------------------------------------------------------------------- | |
638 | // stand-ins in absence of wchar_t | |
639 | // ---------------------------------------------------------------------------- | |
640 | ||
641 | class WXDLLIMPEXP_BASE wxMBConv | |
642 | { | |
643 | public: | |
644 | const char* cMB2WX(const char *psz) const { return psz; } | |
645 | const char* cWX2MB(const char *psz) const { return psz; } | |
646 | }; | |
647 | ||
648 | #define wxConvFile wxConvLocal | |
649 | ||
650 | extern WXDLLIMPEXP_DATA_BASE(wxMBConv) wxConvLibc, | |
651 | wxConvLocal, | |
652 | wxConvISO8859_1, | |
653 | wxConvUTF8; | |
654 | extern WXDLLIMPEXP_DATA_BASE(wxMBConv *) wxConvCurrent; | |
655 | ||
656 | #define wxFNCONV(name) name | |
657 | #define wxFNSTRINGCAST WXSTRINGCAST | |
658 | ||
659 | #endif | |
660 | // wxUSE_WCHAR_T | |
661 | ||
662 | // ---------------------------------------------------------------------------- | |
663 | // macros for the most common conversions | |
664 | // ---------------------------------------------------------------------------- | |
665 | ||
666 | #if wxUSE_UNICODE | |
667 | #define wxConvertWX2MB(s) wxConvCurrent->cWX2MB(s) | |
668 | #define wxConvertMB2WX(s) wxConvCurrent->cMB2WX(s) | |
669 | ||
670 | // these functions should be used when the conversions really, really have | |
671 | // to succeed (usually because we pass their results to a standard C | |
672 | // function which would crash if we passed NULL to it), so these functions | |
673 | // always return a valid pointer if their argument is non-NULL | |
674 | ||
675 | // this function safety is achieved by trying wxConvLibc first, wxConvUTF8 | |
676 | // next if it fails and, finally, wxConvISO8859_1 which always succeeds | |
677 | extern WXDLLIMPEXP_BASE wxWCharBuffer wxSafeConvertMB2WX(const char *s); | |
678 | ||
679 | // this function uses wxConvLibc and wxConvUTF8(MAP_INVALID_UTF8_TO_OCTAL) | |
680 | // if it fails | |
681 | extern WXDLLIMPEXP_BASE wxCharBuffer wxSafeConvertWX2MB(const wchar_t *ws); | |
682 | #else // ANSI | |
683 | // no conversions to do | |
684 | #define wxConvertWX2MB(s) (s) | |
685 | #define wxConvertMB2WX(s) (s) | |
686 | #define wxSafeConvertMB2WX(s) (s) | |
687 | #define wxSafeConvertWX2MB(s) (s) | |
688 | #endif // Unicode/ANSI | |
689 | ||
690 | #endif // _WX_STRCONV_H_ | |
691 |