]>
Commit | Line | Data |
---|---|---|
6001e347 RR |
1 | ///////////////////////////////////////////////////////////////////////////// |
2 | // Name: strconv.cpp | |
3 | // Purpose: Unicode conversion classes | |
4 | // Author: Ove Kaaven, Robert Roebling, Vadim Zeitlin | |
5 | // Modified by: | |
6 | // Created: 29/01/98 | |
7 | // RCS-ID: $Id$ | |
8 | // Copyright: (c) 1999 Ove Kaaven, Robert Roebling, Vadim Zeitlin | |
9 | // Licence: wxWindows license | |
10 | ///////////////////////////////////////////////////////////////////////////// | |
11 | ||
f6bcfd97 BP |
12 | // ============================================================================ |
13 | // declarations | |
14 | // ============================================================================ | |
15 | ||
16 | // ---------------------------------------------------------------------------- | |
17 | // headers | |
18 | // ---------------------------------------------------------------------------- | |
19 | ||
6001e347 RR |
20 | #ifdef __GNUG__ |
21 | #pragma implementation "strconv.h" | |
22 | #endif | |
23 | ||
24 | // For compilers that support precompilation, includes "wx.h". | |
25 | #include "wx/wxprec.h" | |
26 | ||
27 | #ifdef __BORLANDC__ | |
28 | #pragma hdrstop | |
29 | #endif | |
30 | ||
0a1c1e62 GRG |
31 | #ifdef __WXMSW__ |
32 | #include "wx/msw/private.h" | |
33 | #endif | |
34 | ||
1cd52418 | 35 | #include <errno.h> |
6001e347 RR |
36 | #include <ctype.h> |
37 | #include <string.h> | |
38 | #include <stdlib.h> | |
39 | ||
40 | #ifdef __SALFORDC__ | |
41 | #include <clib.h> | |
42 | #endif | |
43 | ||
1cd52418 OK |
44 | #ifdef HAVE_ICONV_H |
45 | #include <iconv.h> | |
46 | #endif | |
47 | #ifdef HAVE_LANGINFO_H | |
48 | #include <langinfo.h> | |
49 | #endif | |
50 | ||
3e61dfb0 OK |
51 | #ifdef __WXMSW__ |
52 | #include <windows.h> | |
53 | #endif | |
54 | ||
6001e347 RR |
55 | #include "wx/debug.h" |
56 | #include "wx/strconv.h" | |
3caec1bb VS |
57 | #include "wx/intl.h" |
58 | #include "wx/log.h" | |
6001e347 | 59 | |
d43088ee | 60 | #if defined(WORDS_BIGENDIAN) || defined(__STDC_ISO_10646__) |
1cd52418 OK |
61 | #define BSWAP_UCS4(str, len) |
62 | #define BSWAP_UCS2(str, len) | |
63 | #else | |
64 | #define BSWAP_UCS4(str, len) { unsigned _c; for (_c=0; _c<len; _c++) str[_c]=wxUINT32_SWAP_ALWAYS(str[_c]); } | |
65 | #define BSWAP_UCS2(str, len) { unsigned _c; for (_c=0; _c<len; _c++) str[_c]=wxUINT16_SWAP_ALWAYS(str[_c]); } | |
66 | #define WC_NEED_BSWAP | |
67 | #endif | |
68 | #define BSWAP_UTF32(str, len) BSWAP_UCS4(str, len) | |
69 | #define BSWAP_UTF16(str, len) BSWAP_UCS2(str, len) | |
70 | ||
71 | #if SIZEOF_WCHAR_T == 4 | |
72 | #define WC_NAME "UCS4" | |
73 | #define WC_BSWAP BSWAP_UCS4 | |
74 | #elif SIZEOF_WCHAR_T == 2 | |
75 | #define WC_NAME "UTF16" | |
76 | #define WC_BSWAP BSWAP_UTF16 | |
77 | #define WC_UTF16 | |
78 | #endif | |
79 | ||
f6bcfd97 BP |
80 | // ---------------------------------------------------------------------------- |
81 | // globals | |
82 | // ---------------------------------------------------------------------------- | |
6001e347 RR |
83 | |
84 | WXDLLEXPORT_DATA(wxMBConv *) wxConvCurrent = &wxConvLibc; | |
85 | ||
f6bcfd97 BP |
86 | // ============================================================================ |
87 | // implementation | |
88 | // ============================================================================ | |
6001e347 | 89 | |
f6bcfd97 | 90 | #if wxUSE_WCHAR_T |
6001e347 | 91 | |
b0a6bb75 VZ |
92 | #ifdef WC_UTF16 |
93 | ||
1cd52418 OK |
94 | static size_t encode_utf16(wxUint32 input,wxUint16*output) |
95 | { | |
96 | if (input<=0xffff) { | |
97 | if (output) *output++ = input; | |
98 | return 1; | |
99 | } else | |
100 | if (input>=0x110000) { | |
101 | return (size_t)-1; | |
102 | } else { | |
103 | if (output) { | |
104 | *output++ = (input >> 10)+0xd7c0; | |
105 | *output++ = (input&0x3ff)+0xdc00; | |
106 | } | |
107 | return 2; | |
108 | } | |
109 | } | |
110 | ||
111 | static size_t decode_utf16(wxUint16*input,wxUint32&output) | |
112 | { | |
113 | if ((*input<0xd800) || (*input>0xdfff)) { | |
114 | output = *input; | |
115 | return 1; | |
116 | } else | |
117 | if ((input[1]<0xdc00) || (input[1]>=0xdfff)) { | |
118 | output = *input; | |
119 | return (size_t)-1; | |
120 | } else { | |
121 | output = ((input[0] - 0xd7c0) << 10) + (input[1] - 0xdc00); | |
122 | return 2; | |
123 | } | |
124 | } | |
125 | ||
b0a6bb75 VZ |
126 | #endif // WC_UTF16 |
127 | ||
f6bcfd97 | 128 | // ---------------------------------------------------------------------------- |
6001e347 | 129 | // wxMBConv |
f6bcfd97 | 130 | // ---------------------------------------------------------------------------- |
6001e347 RR |
131 | |
132 | WXDLLEXPORT_DATA(wxMBConv) wxConvLibc; | |
133 | ||
134 | size_t wxMBConv::MB2WC(wchar_t *buf, const char *psz, size_t n) const | |
135 | { | |
136 | return wxMB2WC(buf, psz, n); | |
137 | } | |
138 | ||
139 | size_t wxMBConv::WC2MB(char *buf, const wchar_t *psz, size_t n) const | |
140 | { | |
141 | return wxWC2MB(buf, psz, n); | |
142 | } | |
143 | ||
144 | const wxWCharBuffer wxMBConv::cMB2WC(const char *psz) const | |
145 | { | |
f6bcfd97 | 146 | if (psz) |
6001e347 RR |
147 | { |
148 | size_t nLen = MB2WC((wchar_t *) NULL, psz, 0); | |
f6bcfd97 BP |
149 | if (nLen == (size_t)-1) |
150 | return wxWCharBuffer((wchar_t *) NULL); | |
6001e347 RR |
151 | wxWCharBuffer buf(nLen); |
152 | MB2WC((wchar_t *)(const wchar_t *) buf, psz, nLen); | |
153 | return buf; | |
f6bcfd97 BP |
154 | } |
155 | else | |
6001e347 RR |
156 | return wxWCharBuffer((wchar_t *) NULL); |
157 | } | |
158 | ||
159 | const wxCharBuffer wxMBConv::cWC2MB(const wchar_t *psz) const | |
160 | { | |
f6bcfd97 | 161 | if (psz) |
6001e347 RR |
162 | { |
163 | size_t nLen = WC2MB((char *) NULL, psz, 0); | |
f6bcfd97 BP |
164 | if (nLen == (size_t)-1) |
165 | return wxCharBuffer((char *) NULL); | |
6001e347 RR |
166 | wxCharBuffer buf(nLen); |
167 | WC2MB((char *)(const char *) buf, psz, nLen); | |
168 | return buf; | |
f6bcfd97 BP |
169 | } |
170 | else | |
6001e347 RR |
171 | return wxCharBuffer((char *) NULL); |
172 | } | |
173 | ||
f6bcfd97 | 174 | // ---------------------------------------------------------------------------- |
6001e347 | 175 | // standard file conversion |
f6bcfd97 | 176 | // ---------------------------------------------------------------------------- |
6001e347 RR |
177 | |
178 | WXDLLEXPORT_DATA(wxMBConvFile) wxConvFile; | |
179 | ||
180 | // just use the libc conversion for now | |
181 | size_t wxMBConvFile::MB2WC(wchar_t *buf, const char *psz, size_t n) const | |
182 | { | |
183 | return wxMB2WC(buf, psz, n); | |
184 | } | |
185 | ||
186 | size_t wxMBConvFile::WC2MB(char *buf, const wchar_t *psz, size_t n) const | |
187 | { | |
188 | return wxWC2MB(buf, psz, n); | |
189 | } | |
190 | ||
f6bcfd97 | 191 | // ---------------------------------------------------------------------------- |
6001e347 | 192 | // standard gdk conversion |
f6bcfd97 BP |
193 | // ---------------------------------------------------------------------------- |
194 | ||
195 | #ifdef __WXGTK12__ | |
6001e347 RR |
196 | |
197 | WXDLLEXPORT_DATA(wxMBConvGdk) wxConvGdk; | |
198 | ||
199 | #include <gdk/gdk.h> | |
200 | ||
201 | size_t wxMBConvGdk::MB2WC(wchar_t *buf, const char *psz, size_t n) const | |
202 | { | |
203 | if (buf) { | |
204 | return gdk_mbstowcs((GdkWChar *)buf, psz, n); | |
205 | } else { | |
206 | GdkWChar *nbuf = new GdkWChar[n=strlen(psz)]; | |
207 | size_t len = gdk_mbstowcs(nbuf, psz, n); | |
208 | delete [] nbuf; | |
209 | return len; | |
210 | } | |
211 | } | |
212 | ||
213 | size_t wxMBConvGdk::WC2MB(char *buf, const wchar_t *psz, size_t n) const | |
214 | { | |
215 | char *mbstr = gdk_wcstombs((GdkWChar *)psz); | |
216 | size_t len = mbstr ? strlen(mbstr) : 0; | |
217 | if (buf) { | |
218 | if (len > n) len = n; | |
219 | memcpy(buf, psz, len); | |
220 | if (len < n) buf[len] = 0; | |
221 | } | |
222 | return len; | |
223 | } | |
f6bcfd97 | 224 | |
6001e347 RR |
225 | #endif // GTK > 1.0 |
226 | ||
227 | // ---------------------------------------------------------------------------- | |
228 | // UTF-7 | |
229 | // ---------------------------------------------------------------------------- | |
230 | ||
231 | WXDLLEXPORT_DATA(wxMBConvUTF7) wxConvUTF7; | |
232 | ||
233 | #if 0 | |
234 | static char utf7_setD[]="ABCDEFGHIJKLMNOPQRSTUVWXYZ" | |
235 | "abcdefghijklmnopqrstuvwxyz" | |
236 | "0123456789'(),-./:?"; | |
237 | static char utf7_setO[]="!\"#$%&*;<=>@[]^_`{|}"; | |
238 | static char utf7_setB[]="ABCDEFGHIJKLMNOPQRSTUVWXYZ" | |
239 | "abcdefghijklmnopqrstuvwxyz" | |
240 | "0123456789+/"; | |
241 | #endif | |
242 | ||
243 | // TODO: write actual implementations of UTF-7 here | |
244 | size_t wxMBConvUTF7::MB2WC(wchar_t * WXUNUSED(buf), | |
245 | const char * WXUNUSED(psz), | |
246 | size_t WXUNUSED(n)) const | |
247 | { | |
248 | return 0; | |
249 | } | |
250 | ||
251 | size_t wxMBConvUTF7::WC2MB(char * WXUNUSED(buf), | |
252 | const wchar_t * WXUNUSED(psz), | |
253 | size_t WXUNUSED(n)) const | |
254 | { | |
255 | return 0; | |
256 | } | |
257 | ||
f6bcfd97 | 258 | // ---------------------------------------------------------------------------- |
6001e347 | 259 | // UTF-8 |
f6bcfd97 | 260 | // ---------------------------------------------------------------------------- |
6001e347 RR |
261 | |
262 | WXDLLEXPORT_DATA(wxMBConvUTF8) wxConvUTF8; | |
263 | ||
1cd52418 | 264 | static wxUint32 utf8_max[]={0x7f,0x7ff,0xffff,0x1fffff,0x3ffffff,0x7fffffff,0xffffffff}; |
6001e347 RR |
265 | |
266 | size_t wxMBConvUTF8::MB2WC(wchar_t *buf, const char *psz, size_t n) const | |
267 | { | |
268 | size_t len = 0; | |
269 | ||
270 | while (*psz && ((!buf) || (len<n))) { | |
271 | unsigned char cc=*psz++, fc=cc; | |
272 | unsigned cnt; | |
273 | for (cnt=0; fc&0x80; cnt++) fc<<=1; | |
274 | if (!cnt) { | |
275 | // plain ASCII char | |
276 | if (buf) *buf++=cc; | |
277 | len++; | |
278 | } else { | |
279 | cnt--; | |
280 | if (!cnt) { | |
281 | // invalid UTF-8 sequence | |
282 | return (size_t)-1; | |
283 | } else { | |
284 | unsigned ocnt=cnt-1; | |
1cd52418 | 285 | wxUint32 res=cc&(0x3f>>cnt); |
6001e347 RR |
286 | while (cnt--) { |
287 | cc = *psz++; | |
288 | if ((cc&0xC0)!=0x80) { | |
289 | // invalid UTF-8 sequence | |
290 | return (size_t)-1; | |
291 | } | |
292 | res=(res<<6)|(cc&0x3f); | |
293 | } | |
294 | if (res<=utf8_max[ocnt]) { | |
295 | // illegal UTF-8 encoding | |
296 | return (size_t)-1; | |
297 | } | |
1cd52418 OK |
298 | #ifdef WC_UTF16 |
299 | size_t pa = encode_utf16(res, buf); | |
300 | if (pa == (size_t)-1) | |
301 | return (size_t)-1; | |
302 | if (buf) buf+=pa; | |
303 | len+=pa; | |
304 | #else | |
6001e347 RR |
305 | if (buf) *buf++=res; |
306 | len++; | |
1cd52418 | 307 | #endif |
6001e347 RR |
308 | } |
309 | } | |
310 | } | |
311 | if (buf && (len<n)) *buf = 0; | |
312 | return len; | |
313 | } | |
314 | ||
315 | size_t wxMBConvUTF8::WC2MB(char *buf, const wchar_t *psz, size_t n) const | |
316 | { | |
317 | size_t len = 0; | |
318 | ||
319 | while (*psz && ((!buf) || (len<n))) { | |
1cd52418 OK |
320 | wxUint32 cc; |
321 | #ifdef WC_UTF16 | |
322 | size_t pa = decode_utf16(psz,cc); | |
323 | psz += (pa == (size_t)-1) ? 1 : pa; | |
324 | #else | |
325 | cc=(*psz++)&0x7fffffff; | |
326 | #endif | |
6001e347 RR |
327 | unsigned cnt; |
328 | for (cnt=0; cc>utf8_max[cnt]; cnt++); | |
329 | if (!cnt) { | |
330 | // plain ASCII char | |
331 | if (buf) *buf++=cc; | |
332 | len++; | |
333 | } else { | |
334 | len+=cnt+1; | |
335 | if (buf) { | |
336 | *buf++=(-128>>cnt)|((cc>>(cnt*6))&(0x3f>>cnt)); | |
337 | while (cnt--) | |
338 | *buf++=0x80|((cc>>(cnt*6))&0x3f); | |
339 | } | |
340 | } | |
341 | } | |
342 | if (buf && (len<n)) *buf = 0; | |
343 | return len; | |
344 | } | |
345 | ||
346 | // ---------------------------------------------------------------------------- | |
347 | // specified character set | |
348 | // ---------------------------------------------------------------------------- | |
349 | ||
f6bcfd97 BP |
350 | WXDLLEXPORT_DATA(wxCSConv) wxConvLocal((const wxChar *)NULL); |
351 | ||
352 | #include "wx/encconv.h" | |
353 | #include "wx/fontmap.h" | |
6001e347 | 354 | |
1cd52418 OK |
355 | // TODO: add some tables here |
356 | // - perhaps common encodings to common codepages (for Win32) | |
357 | // - perhaps common encodings to objects ("UTF8" -> wxConvUTF8) | |
358 | // - move wxEncodingConverter meat in here | |
359 | ||
360 | #ifdef __WIN32__ | |
361 | #include "wx/msw/registry.h" | |
362 | // this should work if M$ Internet Exploiter is installed | |
363 | static long CharsetToCodepage(const wxChar *name) | |
364 | { | |
f1339c56 RR |
365 | if (!name) |
366 | return GetACP(); | |
367 | ||
368 | long CP=-1; | |
369 | ||
370 | wxString cn(name); | |
371 | do { | |
372 | wxString path( wxT("MIME\\Database\\Charset\\") ); | |
373 | path += cn; | |
374 | wxRegKey key( wxRegKey::HKCR, path ); | |
375 | ||
376 | /* two cases: either there's an AliasForCharset string, | |
377 | * or there are Codepage and InternetEncoding dwords. | |
378 | * The InternetEncoding gives us the actual encoding, | |
379 | * the Codepage just says which Windows character set to | |
380 | * use when displaying the data. | |
381 | */ | |
382 | if (key.QueryValue( wxT("InternetEncoding"), &CP )) break; | |
383 | ||
384 | // no encoding, see if it's an alias | |
385 | if (!key.QueryValue( wxT("AliasForCharset"), cn )) break; | |
386 | } while (1); | |
387 | ||
388 | return CP; | |
1cd52418 OK |
389 | } |
390 | #endif | |
391 | ||
6001e347 RR |
392 | class wxCharacterSet |
393 | { | |
1cd52418 | 394 | public: |
f1339c56 RR |
395 | wxCharacterSet(const wxChar*name) |
396 | : cname(name) {} | |
397 | virtual ~wxCharacterSet() | |
398 | {} | |
399 | virtual size_t MB2WC(wchar_t*buf, const char*psz, size_t n) | |
400 | { return (size_t)-1; } | |
401 | virtual size_t WC2MB(char*buf, const wchar_t*psz, size_t n) | |
402 | { return (size_t)-1; } | |
403 | virtual bool usable() | |
404 | { return FALSE; } | |
405 | public: | |
406 | const wxChar*cname; | |
1cd52418 OK |
407 | }; |
408 | ||
409 | class ID_CharSet : public wxCharacterSet | |
410 | { | |
411 | public: | |
f1339c56 RR |
412 | ID_CharSet(const wxChar*name,wxMBConv*cnv) |
413 | : wxCharacterSet(name), work(cnv) {} | |
414 | ||
415 | size_t MB2WC(wchar_t*buf, const char*psz, size_t n) | |
416 | { return work ? work->MB2WC(buf,psz,n) : (size_t)-1; } | |
417 | ||
418 | size_t WC2MB(char*buf, const wchar_t*psz, size_t n) | |
419 | { return work ? work->WC2MB(buf,psz,n) : (size_t)-1; } | |
420 | ||
421 | bool usable() | |
422 | { return work!=NULL; } | |
423 | public: | |
424 | wxMBConv*work; | |
1cd52418 OK |
425 | }; |
426 | ||
3caec1bb | 427 | |
1cd52418 | 428 | #ifdef HAVE_ICONV_H |
3caec1bb VS |
429 | |
430 | // VS: glibc 2.1.3 is broken in that iconv() conversion to/from UCS4 fails with E2BIG | |
431 | // if output buffer is _exactly_ as big as needed. Such case is (unless there's | |
432 | // yet another bug in glibc) the only case when iconv() returns with (size_t)-1 | |
433 | // (which means error) and says there are 0 bytes left in the input buffer -- | |
434 | // when _real_ error occurs, bytes-left-in-input buffer is non-zero. Hence, | |
435 | // this alternative test for iconv() failure. | |
436 | // [This bug does not appear in glibc 2.2.] | |
437 | #if defined(__GLIBC__) && __GLIBC__ == 2 && __GLIBC_MINOR__ <= 1 | |
438 | #define ICONV_FAILED(cres, bufLeft) ((cres == (size_t)-1) && \ | |
439 | (errno != E2BIG || bufLeft != 0)) | |
440 | #else | |
441 | #define ICONV_FAILED(cres, bufLeft) (cres == (size_t)-1) | |
442 | #endif | |
443 | ||
1cd52418 OK |
444 | class IC_CharSet : public wxCharacterSet |
445 | { | |
446 | public: | |
f1339c56 | 447 | IC_CharSet(const wxChar*name) |
3caec1bb | 448 | : wxCharacterSet(name) |
f1339c56 | 449 | { |
3caec1bb VS |
450 | m2w = iconv_open(WC_NAME, wxConvLibc.cWX2MB(cname)); |
451 | w2m = iconv_open(wxConvLibc.cWX2MB(cname), WC_NAME); | |
f1339c56 RR |
452 | } |
453 | ||
3caec1bb VS |
454 | ~IC_CharSet() |
455 | { | |
456 | if ( m2w != (iconv_t)-1 ) | |
457 | iconv_close(m2w); | |
458 | if ( w2m != (iconv_t)-1 ) | |
459 | iconv_close(w2m); | |
f1339c56 RR |
460 | } |
461 | ||
3caec1bb | 462 | size_t MB2WC(wchar_t *buf, const char *psz, size_t n) |
f1339c56 | 463 | { |
f1339c56 | 464 | size_t inbuf = strlen(psz); |
3caec1bb | 465 | size_t outbuf = n * SIZEOF_WCHAR_T; |
f1339c56 | 466 | size_t res, cres; |
3caec1bb VS |
467 | // VS: Use these instead of psz, buf because iconv() modifies its arguments: |
468 | wchar_t *bufPtr = buf; | |
469 | const char *pszPtr = psz; | |
470 | ||
f1339c56 RR |
471 | if (buf) |
472 | { | |
473 | // have destination buffer, convert there | |
95c8801c | 474 | #ifdef WX_ICONV_TAKES_CHAR |
3caec1bb | 475 | cres = iconv(m2w, (char**)&pszPtr, &inbuf, (char**)&bufPtr, &outbuf); |
95c8801c | 476 | #else |
3caec1bb | 477 | cres = iconv(m2w, &pszPtr, &inbuf, (char**)&bufPtr, &outbuf); |
95c8801c | 478 | #endif |
3caec1bb | 479 | res = n - (outbuf / SIZEOF_WCHAR_T); |
f1339c56 | 480 | // convert to native endianness |
3caec1bb VS |
481 | #ifdef WC_NEED_BSWAP |
482 | WC_BSWAP(buf /* _not_ bufPtr */, res) | |
483 | #endif | |
f1339c56 RR |
484 | } |
485 | else | |
486 | { | |
487 | // no destination buffer... convert using temp buffer | |
488 | // to calculate destination buffer requirement | |
489 | wchar_t tbuf[8]; | |
490 | res = 0; | |
491 | do { | |
3caec1bb | 492 | bufPtr = tbuf; outbuf = 8*SIZEOF_WCHAR_T; |
95c8801c | 493 | #ifdef WX_ICONV_TAKES_CHAR |
3caec1bb | 494 | cres = iconv( m2w, (char**)&pszPtr, &inbuf, (char**)&bufPtr, &outbuf ); |
95c8801c | 495 | #else |
3caec1bb | 496 | cres = iconv( m2w, &pszPtr, &inbuf, (char**)&bufPtr, &outbuf ); |
95c8801c | 497 | #endif |
f1339c56 RR |
498 | res += 8-(outbuf/SIZEOF_WCHAR_T); |
499 | } while ((cres==(size_t)-1) && (errno==E2BIG)); | |
500 | } | |
501 | ||
3caec1bb | 502 | if (ICONV_FAILED(cres, inbuf)) |
f1339c56 | 503 | return (size_t)-1; |
3caec1bb | 504 | |
f1339c56 RR |
505 | return res; |
506 | } | |
507 | ||
508 | size_t WC2MB(char*buf, const wchar_t*psz, size_t n) | |
509 | { | |
1cd52418 | 510 | #if defined(__BORLANDC__) && (__BORLANDC__ > 0x530) |
3caec1bb | 511 | size_t inbuf = std::wcslen(psz) * SIZEOF_WCHAR_T; |
1cd52418 | 512 | #else |
3caec1bb | 513 | size_t inbuf = ::wcslen(psz) * SIZEOF_WCHAR_T; |
1cd52418 | 514 | #endif |
f1339c56 RR |
515 | size_t outbuf = n; |
516 | size_t res, cres; | |
3caec1bb | 517 | |
1cd52418 | 518 | #ifdef WC_NEED_BSWAP |
f1339c56 RR |
519 | // need to copy to temp buffer to switch endianness |
520 | // this absolutely doesn't rock! | |
521 | // (no, doing WC_BSWAP twice on the original buffer won't help, as it | |
522 | // could be in read-only memory, or be accessed in some other thread) | |
523 | wchar_t*tmpbuf=(wchar_t*)malloc((inbuf+1)*SIZEOF_WCHAR_T); | |
524 | memcpy(tmpbuf,psz,(inbuf+1)*SIZEOF_WCHAR_T); | |
525 | WC_BSWAP(tmpbuf, inbuf) | |
526 | psz=tmpbuf; | |
1cd52418 | 527 | #endif |
f1339c56 RR |
528 | if (buf) |
529 | { | |
530 | // have destination buffer, convert there | |
95c8801c | 531 | #ifdef WX_ICONV_TAKES_CHAR |
f1339c56 | 532 | cres = iconv( w2m, (char**)&psz, &inbuf, &buf, &outbuf ); |
95c8801c VS |
533 | #else |
534 | cres = iconv( w2m, (const char**)&psz, &inbuf, &buf, &outbuf ); | |
535 | #endif | |
f1339c56 RR |
536 | res = n-outbuf; |
537 | } | |
538 | else | |
539 | { | |
540 | // no destination buffer... convert using temp buffer | |
541 | // to calculate destination buffer requirement | |
542 | char tbuf[16]; | |
543 | res = 0; | |
544 | do { | |
545 | buf = tbuf; outbuf = 16; | |
95c8801c | 546 | #ifdef WX_ICONV_TAKES_CHAR |
f1339c56 | 547 | cres = iconv( w2m, (char**)&psz, &inbuf, &buf, &outbuf ); |
95c8801c VS |
548 | #else |
549 | cres = iconv( w2m, (const char**)&psz, &inbuf, &buf, &outbuf ); | |
550 | #endif | |
f1339c56 RR |
551 | res += 16 - outbuf; |
552 | } while ((cres==(size_t)-1) && (errno==E2BIG)); | |
553 | } | |
1cd52418 | 554 | #ifdef WC_NEED_BSWAP |
f1339c56 | 555 | free(tmpbuf); |
1cd52418 | 556 | #endif |
3caec1bb | 557 | if (ICONV_FAILED(cres, inbuf)) |
f1339c56 RR |
558 | return (size_t)-1; |
559 | ||
560 | return res; | |
561 | } | |
562 | ||
563 | bool usable() | |
3caec1bb | 564 | { return (m2w != (iconv_t)-1) && (w2m != (iconv_t)-1); } |
f1339c56 RR |
565 | |
566 | public: | |
567 | iconv_t m2w, w2m; | |
1cd52418 OK |
568 | }; |
569 | #endif | |
570 | ||
571 | #ifdef __WIN32__ | |
572 | class CP_CharSet : public wxCharacterSet | |
573 | { | |
574 | public: | |
f1339c56 RR |
575 | CP_CharSet(const wxChar*name) |
576 | : wxCharacterSet(name), CodePage(CharsetToCodepage(name)) {} | |
577 | ||
578 | size_t MB2WC(wchar_t*buf, const char*psz, size_t n) | |
579 | { | |
580 | size_t len = MultiByteToWideChar(CodePage,0,psz,-1,buf,buf?n:0); | |
581 | return len ? len : (size_t)-1; | |
582 | } | |
583 | ||
584 | size_t WC2MB(char*buf, const wchar_t*psz, size_t n) | |
585 | { | |
586 | size_t len = WideCharToMultiByte(CodePage,0,psz,-1,buf,buf?n:0,NULL,NULL); | |
587 | return len ? len : (size_t)-1; | |
588 | } | |
589 | ||
590 | bool usable() | |
591 | { return CodePage!=-1; } | |
592 | ||
593 | public: | |
594 | long CodePage; | |
1cd52418 OK |
595 | }; |
596 | #endif | |
597 | ||
598 | class EC_CharSet : public wxCharacterSet | |
599 | { | |
6001e347 | 600 | public: |
f1339c56 RR |
601 | // temporarily just use wxEncodingConverter stuff, |
602 | // so that it works while a better implementation is built | |
603 | EC_CharSet(const wxChar*name) : wxCharacterSet(name), enc(wxFONTENCODING_SYSTEM) | |
604 | { | |
605 | if (name) | |
606 | enc = wxTheFontMapper->CharsetToEncoding(name, FALSE); | |
607 | m2w.Init(enc, wxFONTENCODING_UNICODE); | |
608 | w2m.Init(wxFONTENCODING_UNICODE, enc); | |
609 | } | |
610 | ||
611 | size_t MB2WC(wchar_t*buf, const char*psz, size_t n) | |
612 | { | |
613 | size_t inbuf = strlen(psz); | |
f1339c56 RR |
614 | if (buf) m2w.Convert(psz,buf); |
615 | return inbuf; | |
616 | } | |
617 | ||
618 | size_t WC2MB(char*buf, const wchar_t*psz, size_t n) | |
619 | { | |
1cd52418 | 620 | #if defined(__BORLANDC__) && (__BORLANDC__ > 0x530) |
f1339c56 | 621 | size_t inbuf = std::wcslen(psz); |
1cd52418 | 622 | #else |
f1339c56 | 623 | size_t inbuf = ::wcslen(psz); |
1cd52418 | 624 | #endif |
f1339c56 RR |
625 | if (buf) |
626 | w2m.Convert(psz,buf); | |
627 | ||
628 | return inbuf; | |
629 | } | |
630 | ||
631 | bool usable() | |
632 | { return (enc!=wxFONTENCODING_SYSTEM) && (enc!=wxFONTENCODING_DEFAULT); } | |
633 | ||
634 | public: | |
635 | wxFontEncoding enc; | |
636 | wxEncodingConverter m2w, w2m; | |
f6bcfd97 | 637 | }; |
6001e347 | 638 | |
f6bcfd97 | 639 | static wxCharacterSet *wxGetCharacterSet(const wxChar *name) |
6001e347 | 640 | { |
f1339c56 RR |
641 | wxCharacterSet *cset = NULL; |
642 | if (name) | |
643 | { | |
644 | if (!wxStricmp(name, wxT("UTF8")) || !wxStricmp(name, wxT("UTF-8"))) | |
645 | { | |
646 | cset = new ID_CharSet(name, &wxConvUTF8); | |
647 | } | |
648 | else | |
649 | { | |
1cd52418 | 650 | #ifdef HAVE_ICONV_H |
f1339c56 | 651 | cset = new IC_CharSet(name); // may not take NULL |
1cd52418 | 652 | #endif |
f1339c56 | 653 | } |
1cd52418 | 654 | } |
f1339c56 RR |
655 | |
656 | if (cset && cset->usable()) return cset; | |
657 | if (cset) delete cset; | |
3caec1bb | 658 | cset = NULL; |
1cd52418 | 659 | #ifdef __WIN32__ |
f1339c56 RR |
660 | cset = new CP_CharSet(name); // may take NULL |
661 | if (cset->usable()) return cset; | |
1cd52418 | 662 | #endif |
f1339c56 RR |
663 | if (cset) delete cset; |
664 | cset = new EC_CharSet(name); | |
665 | if (cset->usable()) return cset; | |
666 | delete cset; | |
3caec1bb | 667 | wxLogError(_("Unknown encoding '%s'!"), name); |
f1339c56 | 668 | return NULL; |
6001e347 RR |
669 | } |
670 | ||
6001e347 RR |
671 | wxCSConv::wxCSConv(const wxChar *charset) |
672 | { | |
f1339c56 RR |
673 | m_name = (wxChar *) NULL; |
674 | m_cset = (wxCharacterSet *) NULL; | |
675 | m_deferred = TRUE; | |
676 | SetName(charset); | |
6001e347 RR |
677 | } |
678 | ||
679 | wxCSConv::~wxCSConv() | |
680 | { | |
f1339c56 RR |
681 | if (m_name) free(m_name); |
682 | if (m_cset) delete m_cset; | |
6001e347 RR |
683 | } |
684 | ||
685 | void wxCSConv::SetName(const wxChar *charset) | |
686 | { | |
f1339c56 RR |
687 | if (charset) |
688 | { | |
689 | m_name = wxStrdup(charset); | |
690 | m_deferred = TRUE; | |
691 | } | |
6001e347 RR |
692 | } |
693 | ||
694 | void wxCSConv::LoadNow() | |
695 | { | |
f1339c56 RR |
696 | if (m_deferred) |
697 | { | |
698 | if (!m_name) | |
699 | { | |
6001e347 | 700 | #ifdef __UNIX__ |
1cd52418 | 701 | #if defined(HAVE_LANGINFO_H) && defined(CODESET) |
f1339c56 RR |
702 | // GNU libc provides current character set this way |
703 | char*alang = nl_langinfo(CODESET); | |
704 | if (alang) | |
705 | { | |
706 | SetName(wxConvLibc.cMB2WX(alang)); | |
707 | } | |
708 | else | |
1cd52418 | 709 | #endif |
f1339c56 RR |
710 | { |
711 | // if we can't get at the character set directly, | |
712 | // try to see if it's in the environment variables | |
713 | // (in most cases this won't work, but I was out of ideas) | |
714 | wxChar *lang = wxGetenv(wxT("LC_ALL")); | |
715 | if (!lang) lang = wxGetenv(wxT("LC_CTYPE")); | |
716 | if (!lang) lang = wxGetenv(wxT("LANG")); | |
717 | wxChar *dot = lang ? wxStrchr(lang, wxT('.')) : (wxChar *)NULL; | |
718 | if (dot) SetName(dot+1); | |
719 | } | |
6001e347 | 720 | #endif |
f1339c56 RR |
721 | } |
722 | m_cset = wxGetCharacterSet(m_name); | |
723 | m_deferred = FALSE; | |
6001e347 | 724 | } |
6001e347 RR |
725 | } |
726 | ||
727 | size_t wxCSConv::MB2WC(wchar_t *buf, const char *psz, size_t n) const | |
728 | { | |
f1339c56 RR |
729 | ((wxCSConv *)this)->LoadNow(); // discard constness |
730 | ||
731 | if (m_cset) | |
732 | return m_cset->MB2WC(buf, psz, n); | |
733 | ||
734 | // latin-1 (direct) | |
735 | size_t len=strlen(psz); | |
736 | ||
737 | if (buf) | |
738 | { | |
739 | for (size_t c=0; c<=len; c++) | |
740 | buf[c] = (unsigned char)(psz[c]); | |
741 | } | |
742 | ||
743 | return len; | |
6001e347 RR |
744 | } |
745 | ||
746 | size_t wxCSConv::WC2MB(char *buf, const wchar_t *psz, size_t n) const | |
747 | { | |
f1339c56 RR |
748 | ((wxCSConv *)this)->LoadNow(); // discard constness |
749 | ||
750 | if (m_cset) | |
751 | return m_cset->WC2MB(buf, psz, n); | |
1cd52418 | 752 | |
f1339c56 | 753 | // latin-1 (direct) |
d834f22c | 754 | #if defined(__BORLANDC__) && (__BORLANDC__ > 0x530) |
f1339c56 | 755 | size_t len=std::wcslen(psz); |
d834f22c | 756 | #else |
f1339c56 | 757 | size_t len=::wcslen(psz); |
d834f22c | 758 | #endif |
f1339c56 RR |
759 | if (buf) |
760 | { | |
761 | for (size_t c=0; c<=len; c++) | |
762 | buf[c] = (psz[c]>0xff) ? '?' : psz[c]; | |
763 | } | |
764 | ||
765 | return len; | |
6001e347 RR |
766 | } |
767 | ||
1cd52418 OK |
768 | #ifdef HAVE_ICONV_H |
769 | class IC_CharSetConverter | |
770 | { | |
771 | public: | |
f1339c56 RR |
772 | IC_CharSetConverter(IC_CharSet*from,IC_CharSet*to) |
773 | { cnv = iconv_open(wxConvLibc.cWX2MB(to->cname),wxConvLibc.cWX2MB(from->cname)); } | |
774 | ||
775 | ~IC_CharSetConverter() | |
776 | { if (cnv!=(iconv_t)-1) iconv_close(cnv); } | |
777 | ||
778 | size_t Convert(char*buf, const char*psz, size_t n) | |
779 | { | |
780 | size_t inbuf = strlen(psz); | |
781 | size_t outbuf = n; | |
95c8801c | 782 | #ifdef WX_ICONV_TAKES_CHAR |
f1339c56 | 783 | size_t res = iconv( cnv, (char**)&psz, &inbuf, &buf, &outbuf ); |
95c8801c VS |
784 | #else |
785 | size_t res = iconv( cnv, &psz, &inbuf, &buf, &outbuf ); | |
786 | #endif | |
f1339c56 RR |
787 | if (res==(size_t)-1) return (size_t)-1; |
788 | return n-outbuf; | |
789 | } | |
790 | ||
791 | public: | |
792 | iconv_t cnv; | |
1cd52418 OK |
793 | }; |
794 | #endif | |
795 | ||
796 | class EC_CharSetConverter | |
797 | { | |
798 | public: | |
f1339c56 RR |
799 | EC_CharSetConverter(EC_CharSet*from,EC_CharSet*to) |
800 | { cnv.Init(from->enc,to->enc); } | |
801 | ||
802 | size_t Convert(char*buf, const char*psz, size_t n) | |
803 | { | |
804 | size_t inbuf = strlen(psz); | |
805 | if (buf) cnv.Convert(psz,buf); | |
806 | return inbuf; | |
807 | } | |
808 | ||
809 | public: | |
810 | wxEncodingConverter cnv; | |
1cd52418 OK |
811 | }; |
812 | ||
f6bcfd97 BP |
813 | #else // !wxUSE_WCHAR_T |
814 | ||
815 | // ---------------------------------------------------------------------------- | |
816 | // stand-ins in absence of wchar_t | |
817 | // ---------------------------------------------------------------------------- | |
818 | ||
819 | WXDLLEXPORT_DATA(wxMBConv) wxConvLibc, wxConvFile; | |
820 | ||
821 | #endif // wxUSE_WCHAR_T | |
6001e347 RR |
822 | |
823 |