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