]>
Commit | Line | Data |
---|---|---|
6001e347 | 1 | /////////////////////////////////////////////////////////////////////////////// |
80fdcdb9 | 2 | // Name: wx/strconv.h |
6001e347 | 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 | |
467e0479 VZ |
7 | // Copyright: (c) 1998 Ove Kaaven, Robert Roebling |
8 | // (c) 1998-2006 Vadim Zeitlin | |
65571936 | 9 | // Licence: wxWindows licence |
6001e347 RR |
10 | /////////////////////////////////////////////////////////////////////////////// |
11 | ||
d36c9347 VZ |
12 | #ifndef _WX_STRCONV_H_ |
13 | #define _WX_STRCONV_H_ | |
6001e347 | 14 | |
6001e347 | 15 | #include "wx/defs.h" |
e3f6cbd9 | 16 | #include "wx/chartype.h" |
6001e347 RR |
17 | #include "wx/buffer.h" |
18 | ||
7db39dd6 CE |
19 | #ifdef __DIGITALMARS__ |
20 | #include "typeinfo.h" | |
21 | #endif | |
22 | ||
9dea36ef DW |
23 | #if defined(__VISAGECPP__) && __IBMCPP__ >= 400 |
24 | # undef __BSEXCPT__ | |
25 | #endif | |
dccce9ea | 26 | |
6001e347 RR |
27 | #include <stdlib.h> |
28 | ||
b5dbe15d | 29 | class WXDLLIMPEXP_FWD_BASE wxString; |
86501081 | 30 | |
483b0434 VZ |
31 | // the error value returned by wxMBConv methods |
32 | #define wxCONV_FAILED ((size_t)-1) | |
33 | ||
e90c1d2a | 34 | // ---------------------------------------------------------------------------- |
bde4baac | 35 | // wxMBConv (abstract base class for conversions) |
e90c1d2a | 36 | // ---------------------------------------------------------------------------- |
6001e347 | 37 | |
509da451 VZ |
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 | // | |
d36c9347 VZ |
43 | // You also have to implement Clone() to allow copying the conversions |
44 | // polymorphically. | |
45 | // | |
509da451 | 46 | // And you might need to override GetMBNulLen() as well. |
bddd7a8d | 47 | class WXDLLIMPEXP_BASE wxMBConv |
6001e347 RR |
48 | { |
49 | public: | |
483b0434 VZ |
50 | // The functions doing actual conversion from/to narrow to/from wide |
51 | // character strings. | |
bde4baac | 52 | // |
483b0434 VZ |
53 | // On success, the return value is the length (i.e. the number of |
54 | // characters, not bytes) of the converted string including any trailing | |
55 | // L'\0' or (possibly multiple) '\0'(s). If the conversion fails or if | |
56 | // there is not enough space for everything, including the trailing NUL | |
467e0479 | 57 | // character(s), in the output buffer, wxCONV_FAILED is returned. |
483b0434 | 58 | // |
96132605 VZ |
59 | // In the special case when dst is NULL (the value of dstLen is ignored |
60 | // then) the return value is the length of the needed buffer but nothing | |
61 | // happens otherwise. If srcLen is wxNO_LEN, the entire string, up to and | |
467e0479 VZ |
62 | // including the trailing NUL(s), is converted, otherwise exactly srcLen |
63 | // bytes are. | |
483b0434 VZ |
64 | // |
65 | // Typical usage: | |
66 | // | |
67 | // size_t dstLen = conv.ToWChar(NULL, 0, src); | |
96132605 | 68 | // if ( dstLen == wxCONV_FAILED ) |
483b0434 VZ |
69 | // ... handle error ... |
70 | // wchar_t *wbuf = new wchar_t[dstLen]; | |
71 | // conv.ToWChar(wbuf, dstLen, src); | |
96132605 VZ |
72 | // ... work with wbuf ... |
73 | // delete [] wbuf; | |
483b0434 VZ |
74 | // |
75 | virtual size_t ToWChar(wchar_t *dst, size_t dstLen, | |
467e0479 | 76 | const char *src, size_t srcLen = wxNO_LEN) const; |
483b0434 VZ |
77 | |
78 | virtual size_t FromWChar(char *dst, size_t dstLen, | |
467e0479 | 79 | const wchar_t *src, size_t srcLen = wxNO_LEN) const; |
483b0434 | 80 | |
e90c1d2a | 81 | |
483b0434 VZ |
82 | // Convenience functions for translating NUL-terminated strings: returns |
83 | // the buffer containing the converted string or NULL pointer if the | |
84 | // conversion failed. | |
eec47cc6 VZ |
85 | const wxWCharBuffer cMB2WC(const char *in) const; |
86 | const wxCharBuffer cWC2MB(const wchar_t *in) const; | |
6001e347 | 87 | |
483b0434 VZ |
88 | // Convenience functions for converting strings which may contain embedded |
89 | // NULs and don't have to be NUL-terminated. | |
f5fb6871 | 90 | // |
f6a02087 VZ |
91 | // inLen is the length of the buffer including trailing NUL if any or |
92 | // wxNO_LEN if the input is NUL-terminated. | |
eec47cc6 VZ |
93 | // |
94 | // outLen receives, if not NULL, the length of the converted string or 0 if | |
95 | // the conversion failed (returning 0 and not -1 in this case makes it | |
96 | // difficult to distinguish between failed conversion and empty input but | |
f6a02087 VZ |
97 | // this is done for backwards compatibility). Notice that the rules for |
98 | // whether outLen accounts or not for the last NUL are the same as for | |
99 | // To/FromWChar() above: if inLen is specified, outLen is exactly the | |
100 | // number of characters converted, whether the last one of them was NUL or | |
101 | // not. But if inLen == wxNO_LEN then outLen doesn't account for the last | |
102 | // NUL even though it is present. | |
eec47cc6 VZ |
103 | const wxWCharBuffer |
104 | cMB2WC(const char *in, size_t inLen, size_t *outLen) const; | |
105 | const wxCharBuffer | |
106 | cWC2MB(const wchar_t *in, size_t inLen, size_t *outLen) const; | |
f5fb6871 | 107 | |
40ac5040 VZ |
108 | // And yet more convenience functions for converting the entire buffers: |
109 | // these are the simplest and least error-prone as you never need to bother | |
110 | // with lengths/sizes directly. | |
111 | const wxWCharBuffer cMB2WC(const wxScopedCharBuffer& in) const; | |
112 | const wxCharBuffer cWC2MB(const wxScopedWCharBuffer& in) const; | |
113 | ||
bde4baac | 114 | // convenience functions for converting MB or WC to/from wxWin default |
6001e347 | 115 | #if wxUSE_UNICODE |
e90c1d2a VZ |
116 | const wxWCharBuffer cMB2WX(const char *psz) const { return cMB2WC(psz); } |
117 | const wxCharBuffer cWX2MB(const wchar_t *psz) const { return cWC2MB(psz); } | |
118 | const wchar_t* cWC2WX(const wchar_t *psz) const { return psz; } | |
f6bcfd97 | 119 | const wchar_t* cWX2WC(const wchar_t *psz) const { return psz; } |
e90c1d2a VZ |
120 | #else // ANSI |
121 | const char* cMB2WX(const char *psz) const { return psz; } | |
122 | const char* cWX2MB(const char *psz) const { return psz; } | |
123 | const wxCharBuffer cWC2WX(const wchar_t *psz) const { return cWC2MB(psz); } | |
124 | const wxWCharBuffer cWX2WC(const char *psz) const { return cMB2WC(psz); } | |
125 | #endif // Unicode/ANSI | |
2b5f62a0 | 126 | |
c1464d9d VZ |
127 | // this function is used in the implementation of cMB2WC() to distinguish |
128 | // between the following cases: | |
eec47cc6 | 129 | // |
c1464d9d VZ |
130 | // a) var width encoding with strings terminated by a single NUL |
131 | // (usual multibyte encodings): return 1 in this case | |
132 | // b) fixed width encoding with 2 bytes/char and so terminated by | |
133 | // 2 NULs (UTF-16/UCS-2 and variants): return 2 in this case | |
134 | // c) fixed width encoding with 4 bytes/char and so terminated by | |
135 | // 4 NULs (UTF-32/UCS-4 and variants): return 4 in this case | |
136 | // | |
137 | // anything else is not supported currently and -1 should be returned | |
7ef3ab50 VZ |
138 | virtual size_t GetMBNulLen() const { return 1; } |
139 | ||
483b0434 VZ |
140 | // return the maximal value currently returned by GetMBNulLen() for any |
141 | // encoding | |
142 | static size_t GetMaxMBNulLen() { return 4 /* for UTF-32 */; } | |
143 | ||
111d9948 VS |
144 | #if wxUSE_UNICODE_UTF8 |
145 | // return true if the converter's charset is UTF-8, i.e. char* strings | |
146 | // decoded using this object can be directly copied to wxString's internal | |
147 | // storage without converting to WC and than back to UTF-8 MB string | |
148 | virtual bool IsUTF8() const { return false; } | |
149 | #endif | |
483b0434 VZ |
150 | |
151 | // The old conversion functions. The existing classes currently mostly | |
152 | // implement these ones but we're in transition to using To/FromWChar() | |
153 | // instead and any new classes should implement just the new functions. | |
154 | // For now, however, we provide default implementation of To/FromWChar() in | |
155 | // this base class in terms of MB2WC/WC2MB() to avoid having to rewrite all | |
156 | // the conversions at once. | |
157 | // | |
158 | // On success, the return value is the length (i.e. the number of | |
159 | // characters, not bytes) not counting the trailing NUL(s) of the converted | |
160 | // string. On failure, (size_t)-1 is returned. In the special case when | |
161 | // outputBuf is NULL the return value is the same one but nothing is | |
162 | // written to the buffer. | |
163 | // | |
164 | // Note that outLen is the length of the output buffer, not the length of | |
165 | // the input (which is always supposed to be terminated by one or more | |
166 | // NULs, as appropriate for the encoding)! | |
509da451 VZ |
167 | virtual size_t MB2WC(wchar_t *out, const char *in, size_t outLen) const; |
168 | virtual size_t WC2MB(char *out, const wchar_t *in, size_t outLen) const; | |
483b0434 VZ |
169 | |
170 | ||
d36c9347 VZ |
171 | // make a heap-allocated copy of this object |
172 | virtual wxMBConv *Clone() const = 0; | |
173 | ||
7ef3ab50 VZ |
174 | // virtual dtor for any base class |
175 | virtual ~wxMBConv(); | |
6001e347 RR |
176 | }; |
177 | ||
bde4baac VZ |
178 | // ---------------------------------------------------------------------------- |
179 | // wxMBConvLibc uses standard mbstowcs() and wcstombs() functions for | |
180 | // conversion (hence it depends on the current locale) | |
181 | // ---------------------------------------------------------------------------- | |
182 | ||
183 | class WXDLLIMPEXP_BASE wxMBConvLibc : public wxMBConv | |
184 | { | |
185 | public: | |
75736a9c DS |
186 | virtual size_t MB2WC(wchar_t *outputBuf, const char *psz, size_t outputSize) const; |
187 | virtual size_t WC2MB(char *outputBuf, const wchar_t *psz, size_t outputSize) const; | |
d36c9347 VZ |
188 | |
189 | virtual wxMBConv *Clone() const { return new wxMBConvLibc; } | |
111d9948 VS |
190 | |
191 | #if wxUSE_UNICODE_UTF8 | |
192 | virtual bool IsUTF8() const { return wxLocaleIsUtf8; } | |
193 | #endif | |
bde4baac VZ |
194 | }; |
195 | ||
5576edf8 RR |
196 | #ifdef __UNIX__ |
197 | ||
198 | // ---------------------------------------------------------------------------- | |
199 | // wxConvBrokenFileNames is made for Unix in Unicode mode when | |
200 | // files are accidentally written in an encoding which is not | |
201 | // the system encoding. Typically, the system encoding will be | |
202 | // UTF8 but there might be files stored in ISO8859-1 on disk. | |
203 | // ---------------------------------------------------------------------------- | |
204 | ||
205 | class WXDLLIMPEXP_BASE wxConvBrokenFileNames : public wxMBConv | |
206 | { | |
207 | public: | |
86501081 | 208 | wxConvBrokenFileNames(const wxString& charset); |
d36c9347 | 209 | wxConvBrokenFileNames(const wxConvBrokenFileNames& conv) |
2e2cf78d VZ |
210 | : wxMBConv(), |
211 | m_conv(conv.m_conv ? conv.m_conv->Clone() : NULL) | |
d36c9347 VZ |
212 | { |
213 | } | |
5576edf8 RR |
214 | virtual ~wxConvBrokenFileNames() { delete m_conv; } |
215 | ||
eec47cc6 VZ |
216 | virtual size_t MB2WC(wchar_t *out, const char *in, size_t outLen) const |
217 | { | |
218 | return m_conv->MB2WC(out, in, outLen); | |
219 | } | |
220 | ||
221 | virtual size_t WC2MB(char *out, const wchar_t *in, size_t outLen) const | |
222 | { | |
223 | return m_conv->WC2MB(out, in, outLen); | |
224 | } | |
5576edf8 | 225 | |
7ef3ab50 | 226 | virtual size_t GetMBNulLen() const |
eec47cc6 | 227 | { |
22886fb3 | 228 | // cast needed to call a private function |
7ef3ab50 | 229 | return m_conv->GetMBNulLen(); |
eec47cc6 VZ |
230 | } |
231 | ||
ba98e032 VS |
232 | #if wxUSE_UNICODE_UTF8 |
233 | virtual bool IsUTF8() const { return m_conv->IsUTF8(); } | |
234 | #endif | |
235 | ||
d36c9347 VZ |
236 | virtual wxMBConv *Clone() const { return new wxConvBrokenFileNames(*this); } |
237 | ||
7ef3ab50 | 238 | private: |
5576edf8 RR |
239 | // the conversion object we forward to |
240 | wxMBConv *m_conv; | |
d36c9347 | 241 | |
c0c133e1 | 242 | wxDECLARE_NO_ASSIGN_CLASS(wxConvBrokenFileNames); |
5576edf8 RR |
243 | }; |
244 | ||
eec47cc6 | 245 | #endif // __UNIX__ |
5576edf8 | 246 | |
e90c1d2a | 247 | // ---------------------------------------------------------------------------- |
6001e347 | 248 | // wxMBConvUTF7 (for conversion using UTF7 encoding) |
e90c1d2a | 249 | // ---------------------------------------------------------------------------- |
6001e347 | 250 | |
bddd7a8d | 251 | class WXDLLIMPEXP_BASE wxMBConvUTF7 : public wxMBConv |
6001e347 RR |
252 | { |
253 | public: | |
9d653e81 VZ |
254 | wxMBConvUTF7() { } |
255 | ||
256 | // compiler-generated copy ctor, assignment operator and dtor are ok | |
257 | // (assuming it's ok to copy the shift state -- not really sure about it) | |
258 | ||
259 | virtual size_t ToWChar(wchar_t *dst, size_t dstLen, | |
260 | const char *src, size_t srcLen = wxNO_LEN) const; | |
261 | virtual size_t FromWChar(char *dst, size_t dstLen, | |
262 | const wchar_t *src, size_t srcLen = wxNO_LEN) const; | |
d36c9347 VZ |
263 | |
264 | virtual wxMBConv *Clone() const { return new wxMBConvUTF7; } | |
9d653e81 VZ |
265 | |
266 | private: | |
267 | // UTF-7 decoder/encoder may be in direct mode or in shifted mode after a | |
268 | // '+' (and until the '-' or any other non-base64 character) | |
1bc82105 | 269 | struct StateMode |
9d653e81 | 270 | { |
1bc82105 VZ |
271 | enum Mode |
272 | { | |
273 | Direct, // pass through state | |
274 | Shifted // after a '+' (and before '-') | |
275 | }; | |
9d653e81 VZ |
276 | }; |
277 | ||
278 | // the current decoder state: this is only used by ToWChar() if srcLen | |
279 | // parameter is not wxNO_LEN, when working on the entire NUL-terminated | |
280 | // strings we neither update nor use the state | |
5c69ef61 | 281 | class DecoderState : private StateMode |
9d653e81 VZ |
282 | { |
283 | private: | |
284 | // current state: this one is private as we want to enforce the use of | |
285 | // ToDirect/ToShifted() methods below | |
286 | Mode mode; | |
287 | ||
288 | public: | |
289 | // the initial state is direct | |
290 | DecoderState() { mode = Direct; } | |
291 | ||
292 | // switch to/from shifted mode | |
293 | void ToDirect() { mode = Direct; } | |
294 | void ToShifted() { mode = Shifted; accum = bit = 0; isLSB = false; } | |
295 | ||
296 | bool IsDirect() const { return mode == Direct; } | |
297 | bool IsShifted() const { return mode == Shifted; } | |
298 | ||
299 | ||
300 | // these variables are only used in shifted mode | |
301 | ||
302 | unsigned int accum; // accumulator of the bit we've already got | |
303 | unsigned int bit; // the number of bits consumed mod 8 | |
304 | unsigned char msb; // the high byte of UTF-16 word | |
305 | bool isLSB; // whether we're decoding LSB or MSB of UTF-16 word | |
306 | }; | |
307 | ||
308 | DecoderState m_stateDecoder; | |
309 | ||
310 | ||
311 | // encoder state is simpler as we always receive entire Unicode characters | |
312 | // on input | |
5c69ef61 | 313 | class EncoderState : private StateMode |
9d653e81 VZ |
314 | { |
315 | private: | |
316 | Mode mode; | |
317 | ||
318 | public: | |
319 | EncoderState() { mode = Direct; } | |
320 | ||
321 | void ToDirect() { mode = Direct; } | |
322 | void ToShifted() { mode = Shifted; accum = bit = 0; } | |
323 | ||
324 | bool IsDirect() const { return mode == Direct; } | |
325 | bool IsShifted() const { return mode == Shifted; } | |
326 | ||
327 | unsigned int accum; | |
328 | unsigned int bit; | |
329 | }; | |
330 | ||
331 | EncoderState m_stateEncoder; | |
6001e347 RR |
332 | }; |
333 | ||
e90c1d2a | 334 | // ---------------------------------------------------------------------------- |
6001e347 | 335 | // wxMBConvUTF8 (for conversion using UTF8 encoding) |
e90c1d2a | 336 | // ---------------------------------------------------------------------------- |
6001e347 | 337 | |
0286d08d VZ |
338 | // this is the real UTF-8 conversion class, it has to be called "strict UTF-8" |
339 | // for compatibility reasons: the wxMBConvUTF8 class below also supports lossy | |
340 | // conversions if it is created with non default options | |
341 | class WXDLLIMPEXP_BASE wxMBConvStrictUTF8 : public wxMBConv | |
342 | { | |
343 | public: | |
344 | // compiler-generated default ctor and other methods are ok | |
345 | ||
346 | virtual size_t ToWChar(wchar_t *dst, size_t dstLen, | |
347 | const char *src, size_t srcLen = wxNO_LEN) const; | |
348 | virtual size_t FromWChar(char *dst, size_t dstLen, | |
349 | const wchar_t *src, size_t srcLen = wxNO_LEN) const; | |
350 | ||
351 | virtual wxMBConv *Clone() const { return new wxMBConvStrictUTF8(); } | |
352 | ||
353 | #if wxUSE_UNICODE_UTF8 | |
354 | // NB: other mapping modes are not, strictly speaking, UTF-8, so we can't | |
355 | // take the shortcut in that case | |
356 | virtual bool IsUTF8() const { return true; } | |
357 | #endif | |
358 | }; | |
359 | ||
360 | class WXDLLIMPEXP_BASE wxMBConvUTF8 : public wxMBConvStrictUTF8 | |
6001e347 RR |
361 | { |
362 | public: | |
d36c9347 VZ |
363 | enum |
364 | { | |
ea8ce907 RR |
365 | MAP_INVALID_UTF8_NOT = 0, |
366 | MAP_INVALID_UTF8_TO_PUA = 1, | |
367 | MAP_INVALID_UTF8_TO_OCTAL = 2 | |
368 | }; | |
369 | ||
370 | wxMBConvUTF8(int options = MAP_INVALID_UTF8_NOT) : m_options(options) { } | |
d16d0917 VZ |
371 | |
372 | virtual size_t ToWChar(wchar_t *dst, size_t dstLen, | |
373 | const char *src, size_t srcLen = wxNO_LEN) const; | |
374 | virtual size_t FromWChar(char *dst, size_t dstLen, | |
375 | const wchar_t *src, size_t srcLen = wxNO_LEN) const; | |
eec47cc6 | 376 | |
d36c9347 VZ |
377 | virtual wxMBConv *Clone() const { return new wxMBConvUTF8(m_options); } |
378 | ||
111d9948 VS |
379 | #if wxUSE_UNICODE_UTF8 |
380 | // NB: other mapping modes are not, strictly speaking, UTF-8, so we can't | |
381 | // take the shortcut in that case | |
382 | virtual bool IsUTF8() const { return m_options == MAP_INVALID_UTF8_NOT; } | |
383 | #endif | |
384 | ||
ea8ce907 RR |
385 | private: |
386 | int m_options; | |
6001e347 RR |
387 | }; |
388 | ||
eec47cc6 VZ |
389 | // ---------------------------------------------------------------------------- |
390 | // wxMBConvUTF16Base: for both LE and BE variants | |
391 | // ---------------------------------------------------------------------------- | |
392 | ||
393 | class WXDLLIMPEXP_BASE wxMBConvUTF16Base : public wxMBConv | |
394 | { | |
7ef3ab50 | 395 | public: |
467e0479 VZ |
396 | enum { BYTES_PER_CHAR = 2 }; |
397 | ||
398 | virtual size_t GetMBNulLen() const { return BYTES_PER_CHAR; } | |
399 | ||
400 | protected: | |
401 | // return the length of the buffer using srcLen if it's not wxNO_LEN and | |
402 | // computing the length ourselves if it is; also checks that the length is | |
403 | // even if specified as we need an entire number of UTF-16 characters and | |
404 | // returns wxNO_LEN which indicates error if it is odd | |
405 | static size_t GetLength(const char *src, size_t srcLen); | |
eec47cc6 VZ |
406 | }; |
407 | ||
e90c1d2a | 408 | // ---------------------------------------------------------------------------- |
c91830cb VZ |
409 | // wxMBConvUTF16LE (for conversion using UTF16 Little Endian encoding) |
410 | // ---------------------------------------------------------------------------- | |
411 | ||
eec47cc6 | 412 | class WXDLLIMPEXP_BASE wxMBConvUTF16LE : public wxMBConvUTF16Base |
c91830cb VZ |
413 | { |
414 | public: | |
467e0479 VZ |
415 | virtual size_t ToWChar(wchar_t *dst, size_t dstLen, |
416 | const char *src, size_t srcLen = wxNO_LEN) const; | |
417 | virtual size_t FromWChar(char *dst, size_t dstLen, | |
418 | const wchar_t *src, size_t srcLen = wxNO_LEN) const; | |
d36c9347 | 419 | virtual wxMBConv *Clone() const { return new wxMBConvUTF16LE; } |
c91830cb VZ |
420 | }; |
421 | ||
422 | // ---------------------------------------------------------------------------- | |
423 | // wxMBConvUTF16BE (for conversion using UTF16 Big Endian encoding) | |
424 | // ---------------------------------------------------------------------------- | |
425 | ||
eec47cc6 | 426 | class WXDLLIMPEXP_BASE wxMBConvUTF16BE : public wxMBConvUTF16Base |
c91830cb VZ |
427 | { |
428 | public: | |
467e0479 VZ |
429 | virtual size_t ToWChar(wchar_t *dst, size_t dstLen, |
430 | const char *src, size_t srcLen = wxNO_LEN) const; | |
431 | virtual size_t FromWChar(char *dst, size_t dstLen, | |
432 | const wchar_t *src, size_t srcLen = wxNO_LEN) const; | |
d36c9347 | 433 | virtual wxMBConv *Clone() const { return new wxMBConvUTF16BE; } |
c91830cb VZ |
434 | }; |
435 | ||
eec47cc6 VZ |
436 | // ---------------------------------------------------------------------------- |
437 | // wxMBConvUTF32Base: base class for both LE and BE variants | |
438 | // ---------------------------------------------------------------------------- | |
439 | ||
440 | class WXDLLIMPEXP_BASE wxMBConvUTF32Base : public wxMBConv | |
441 | { | |
7ef3ab50 | 442 | public: |
467e0479 VZ |
443 | enum { BYTES_PER_CHAR = 4 }; |
444 | ||
445 | virtual size_t GetMBNulLen() const { return BYTES_PER_CHAR; } | |
446 | ||
447 | protected: | |
448 | // this is similar to wxMBConvUTF16Base method with the same name except | |
449 | // that, of course, it verifies that length is divisible by 4 if given and | |
450 | // not by 2 | |
451 | static size_t GetLength(const char *src, size_t srcLen); | |
eec47cc6 VZ |
452 | }; |
453 | ||
c91830cb | 454 | // ---------------------------------------------------------------------------- |
8b9e1f43 | 455 | // wxMBConvUTF32LE (for conversion using UTF32 Little Endian encoding) |
c91830cb VZ |
456 | // ---------------------------------------------------------------------------- |
457 | ||
eec47cc6 | 458 | class WXDLLIMPEXP_BASE wxMBConvUTF32LE : public wxMBConvUTF32Base |
c91830cb VZ |
459 | { |
460 | public: | |
467e0479 VZ |
461 | virtual size_t ToWChar(wchar_t *dst, size_t dstLen, |
462 | const char *src, size_t srcLen = wxNO_LEN) const; | |
463 | virtual size_t FromWChar(char *dst, size_t dstLen, | |
464 | const wchar_t *src, size_t srcLen = wxNO_LEN) const; | |
d36c9347 | 465 | virtual wxMBConv *Clone() const { return new wxMBConvUTF32LE; } |
c91830cb VZ |
466 | }; |
467 | ||
468 | // ---------------------------------------------------------------------------- | |
8b9e1f43 | 469 | // wxMBConvUTF32BE (for conversion using UTF32 Big Endian encoding) |
c91830cb VZ |
470 | // ---------------------------------------------------------------------------- |
471 | ||
eec47cc6 | 472 | class WXDLLIMPEXP_BASE wxMBConvUTF32BE : public wxMBConvUTF32Base |
c91830cb VZ |
473 | { |
474 | public: | |
467e0479 VZ |
475 | virtual size_t ToWChar(wchar_t *dst, size_t dstLen, |
476 | const char *src, size_t srcLen = wxNO_LEN) const; | |
477 | virtual size_t FromWChar(char *dst, size_t dstLen, | |
478 | const wchar_t *src, size_t srcLen = wxNO_LEN) const; | |
d36c9347 | 479 | virtual wxMBConv *Clone() const { return new wxMBConvUTF32BE; } |
c91830cb VZ |
480 | }; |
481 | ||
482 | // ---------------------------------------------------------------------------- | |
e90c1d2a VZ |
483 | // wxCSConv (for conversion based on loadable char sets) |
484 | // ---------------------------------------------------------------------------- | |
6001e347 | 485 | |
8b04d4c4 VZ |
486 | #include "wx/fontenc.h" |
487 | ||
bddd7a8d | 488 | class WXDLLIMPEXP_BASE wxCSConv : public wxMBConv |
6001e347 | 489 | { |
6001e347 | 490 | public: |
e95354ec VZ |
491 | // we can be created either from charset name or from an encoding constant |
492 | // but we can't have both at once | |
86501081 | 493 | wxCSConv(const wxString& charset); |
8b04d4c4 | 494 | wxCSConv(wxFontEncoding encoding); |
e95354ec | 495 | |
54380f29 | 496 | wxCSConv(const wxCSConv& conv); |
e90c1d2a VZ |
497 | virtual ~wxCSConv(); |
498 | ||
54380f29 | 499 | wxCSConv& operator=(const wxCSConv& conv); |
2b5f62a0 | 500 | |
1c714a5d VZ |
501 | virtual size_t ToWChar(wchar_t *dst, size_t dstLen, |
502 | const char *src, size_t srcLen = wxNO_LEN) const; | |
503 | virtual size_t FromWChar(char *dst, size_t dstLen, | |
504 | const wchar_t *src, size_t srcLen = wxNO_LEN) const; | |
7ef3ab50 | 505 | virtual size_t GetMBNulLen() const; |
1c714a5d | 506 | |
ba98e032 VS |
507 | #if wxUSE_UNICODE_UTF8 |
508 | virtual bool IsUTF8() const; | |
509 | #endif | |
510 | ||
d36c9347 | 511 | virtual wxMBConv *Clone() const { return new wxCSConv(*this); } |
e90c1d2a | 512 | |
d36c9347 | 513 | void Clear(); |
65e50848 | 514 | |
a08a37d0 | 515 | // return true if the conversion could be initialized successfully |
0f0298b1 | 516 | bool IsOk() const; |
0f0298b1 | 517 | |
e90c1d2a | 518 | private: |
8b04d4c4 VZ |
519 | // common part of all ctors |
520 | void Init(); | |
521 | ||
6c4d607e VZ |
522 | // Creates the conversion to use, called from all ctors to initialize |
523 | // m_convReal. | |
e95354ec VZ |
524 | wxMBConv *DoCreate() const; |
525 | ||
6c4d607e VZ |
526 | // Set the name (may be only called when m_name == NULL), makes copy of |
527 | // the charset string. | |
86501081 | 528 | void SetName(const char *charset); |
e90c1d2a | 529 | |
6c4d607e VZ |
530 | // Set m_encoding field respecting the rules below, i.e. making sure it has |
531 | // a valid value if m_name == NULL (thus this should be always called after | |
532 | // SetName()). | |
a08a37d0 | 533 | // |
6c4d607e VZ |
534 | // Input encoding may be valid or not. |
535 | void SetEncoding(wxFontEncoding encoding); | |
a08a37d0 | 536 | |
6c4d607e VZ |
537 | |
538 | // The encoding we use is specified by the two fields below: | |
539 | // | |
540 | // 1. If m_name != NULL, m_encoding corresponds to it if it's one of | |
541 | // encodings we know about (i.e. member of wxFontEncoding) or is | |
542 | // wxFONTENCODING_SYSTEM otherwise. | |
a08a37d0 | 543 | // |
6c4d607e VZ |
544 | // 2. If m_name == NULL, m_encoding is always valid, i.e. not one of |
545 | // wxFONTENCODING_{SYSTEM,DEFAULT,MAX}. | |
546 | char *m_name; | |
8b04d4c4 | 547 | wxFontEncoding m_encoding; |
e95354ec | 548 | |
6c4d607e VZ |
549 | // The conversion object for our encoding or NULL if we failed to create it |
550 | // in which case we fall back to hard-coded ISO8859-1 conversion. | |
e95354ec | 551 | wxMBConv *m_convReal; |
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 |
576 | WX_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 |
580 | WX_DECLARE_GLOBAL_CONV(wxCSConv, wxConvISO8859_1) |
581 | #define wxConvISO8859_1 wxGet_wxConvISO8859_1() | |
582 | ||
0286d08d | 583 | WX_DECLARE_GLOBAL_CONV(wxMBConvStrictUTF8, wxConvUTF8) |
1e50d914 VS |
584 | #define wxConvUTF8 wxGet_wxConvUTF8() |
585 | ||
586 | WX_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) | |
596 | extern 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 | 603 | extern WXDLLIMPEXP_DATA_BASE(wxMBConv *) wxConvCurrent; |
6001e347 | 604 | |
0e052272 | 605 | // the conversion corresponding to the current locale |
1e50d914 VS |
606 | WX_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 | |
613 | extern 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 | ||
e90c1d2a VZ |
646 | // ---------------------------------------------------------------------------- |
647 | // macros for the most common conversions | |
648 | // ---------------------------------------------------------------------------- | |
649 | ||
650 | #if wxUSE_UNICODE | |
651 | #define wxConvertWX2MB(s) wxConvCurrent->cWX2MB(s) | |
652 | #define wxConvertMB2WX(s) wxConvCurrent->cMB2WX(s) | |
69c928ef VZ |
653 | |
654 | // these functions should be used when the conversions really, really have | |
655 | // to succeed (usually because we pass their results to a standard C | |
656 | // function which would crash if we passed NULL to it), so these functions | |
657 | // always return a valid pointer if their argument is non-NULL | |
658 | ||
659 | // this function safety is achieved by trying wxConvLibc first, wxConvUTF8 | |
660 | // next if it fails and, finally, wxConvISO8859_1 which always succeeds | |
661 | extern WXDLLIMPEXP_BASE wxWCharBuffer wxSafeConvertMB2WX(const char *s); | |
662 | ||
663 | // this function uses wxConvLibc and wxConvUTF8(MAP_INVALID_UTF8_TO_OCTAL) | |
664 | // if it fails | |
665 | extern WXDLLIMPEXP_BASE wxCharBuffer wxSafeConvertWX2MB(const wchar_t *ws); | |
e90c1d2a VZ |
666 | #else // ANSI |
667 | // no conversions to do | |
668 | #define wxConvertWX2MB(s) (s) | |
669 | #define wxConvertMB2WX(s) (s) | |
69c928ef VZ |
670 | #define wxSafeConvertMB2WX(s) (s) |
671 | #define wxSafeConvertWX2MB(s) (s) | |
e90c1d2a VZ |
672 | #endif // Unicode/ANSI |
673 | ||
d36c9347 | 674 | #endif // _WX_STRCONV_H_ |
6001e347 | 675 |