]>
Commit | Line | Data |
---|---|---|
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 | ||
12 | // ============================================================================ | |
13 | // declarations | |
14 | // ============================================================================ | |
15 | ||
16 | // ---------------------------------------------------------------------------- | |
17 | // headers | |
18 | // ---------------------------------------------------------------------------- | |
19 | ||
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 | ||
31 | #ifdef __WXMSW__ | |
32 | #include "wx/msw/private.h" | |
33 | #endif | |
34 | ||
35 | #include <errno.h> | |
36 | #include <ctype.h> | |
37 | #include <string.h> | |
38 | #include <stdlib.h> | |
39 | ||
40 | #ifdef __SALFORDC__ | |
41 | #include <clib.h> | |
42 | #endif | |
43 | ||
44 | #ifdef HAVE_ICONV_H | |
45 | #include <iconv.h> | |
46 | #endif | |
47 | ||
48 | #ifdef __WXMSW__ | |
49 | #include <windows.h> | |
50 | #endif | |
51 | ||
52 | #include "wx/debug.h" | |
53 | #include "wx/strconv.h" | |
54 | #include "wx/intl.h" | |
55 | #include "wx/log.h" | |
56 | ||
57 | #if defined(WORDS_BIGENDIAN) || defined(__STDC_ISO_10646__) | |
58 | #define BSWAP_UCS4(str, len) | |
59 | #define BSWAP_UCS2(str, len) | |
60 | #else | |
61 | #define BSWAP_UCS4(str, len) { unsigned _c; for (_c=0; _c<len; _c++) str[_c]=wxUINT32_SWAP_ALWAYS(str[_c]); } | |
62 | #define BSWAP_UCS2(str, len) { unsigned _c; for (_c=0; _c<len; _c++) str[_c]=wxUINT16_SWAP_ALWAYS(str[_c]); } | |
63 | #define WC_NEED_BSWAP | |
64 | #endif | |
65 | #define BSWAP_UTF32(str, len) BSWAP_UCS4(str, len) | |
66 | #define BSWAP_UTF16(str, len) BSWAP_UCS2(str, len) | |
67 | ||
68 | #if SIZEOF_WCHAR_T == 4 | |
69 | #define WC_NAME "UCS4" | |
70 | #define WC_BSWAP BSWAP_UCS4 | |
71 | #elif SIZEOF_WCHAR_T == 2 | |
72 | #define WC_NAME "UTF16" | |
73 | #define WC_BSWAP BSWAP_UTF16 | |
74 | #define WC_UTF16 | |
75 | #endif | |
76 | ||
77 | // ---------------------------------------------------------------------------- | |
78 | // globals | |
79 | // ---------------------------------------------------------------------------- | |
80 | ||
81 | WXDLLEXPORT_DATA(wxMBConv *) wxConvCurrent = &wxConvLibc; | |
82 | ||
83 | // ============================================================================ | |
84 | // implementation | |
85 | // ============================================================================ | |
86 | ||
87 | #if wxUSE_WCHAR_T | |
88 | ||
89 | #ifdef WC_UTF16 | |
90 | ||
91 | static size_t encode_utf16(wxUint32 input,wxUint16*output) | |
92 | { | |
93 | if (input<=0xffff) | |
94 | { | |
95 | if (output) *output++ = input; | |
96 | return 1; | |
97 | } | |
98 | else if (input>=0x110000) | |
99 | { | |
100 | return (size_t)-1; | |
101 | } | |
102 | else | |
103 | { | |
104 | if (output) | |
105 | { | |
106 | *output++ = (input >> 10)+0xd7c0; | |
107 | *output++ = (input&0x3ff)+0xdc00; | |
108 | } | |
109 | return 2; | |
110 | } | |
111 | } | |
112 | ||
113 | static size_t decode_utf16(wxUint16*input,wxUint32&output) | |
114 | { | |
115 | if ((*input<0xd800) || (*input>0xdfff)) | |
116 | { | |
117 | output = *input; | |
118 | return 1; | |
119 | } | |
120 | else if ((input[1]<0xdc00) || (input[1]>=0xdfff)) | |
121 | { | |
122 | output = *input; | |
123 | return (size_t)-1; | |
124 | } | |
125 | else | |
126 | { | |
127 | output = ((input[0] - 0xd7c0) << 10) + (input[1] - 0xdc00); | |
128 | return 2; | |
129 | } | |
130 | } | |
131 | ||
132 | #endif // WC_UTF16 | |
133 | ||
134 | // ---------------------------------------------------------------------------- | |
135 | // wxMBConv | |
136 | // ---------------------------------------------------------------------------- | |
137 | ||
138 | WXDLLEXPORT_DATA(wxMBConv) wxConvLibc; | |
139 | ||
140 | size_t wxMBConv::MB2WC(wchar_t *buf, const char *psz, size_t n) const | |
141 | { | |
142 | return wxMB2WC(buf, psz, n); | |
143 | } | |
144 | ||
145 | size_t wxMBConv::WC2MB(char *buf, const wchar_t *psz, size_t n) const | |
146 | { | |
147 | return wxWC2MB(buf, psz, n); | |
148 | } | |
149 | ||
150 | const wxWCharBuffer wxMBConv::cMB2WC(const char *psz) const | |
151 | { | |
152 | if (psz) | |
153 | { | |
154 | size_t nLen = MB2WC((wchar_t *) NULL, psz, 0); | |
155 | if (nLen == (size_t)-1) | |
156 | return wxWCharBuffer((wchar_t *) NULL); | |
157 | wxWCharBuffer buf(nLen); | |
158 | MB2WC((wchar_t *)(const wchar_t *) buf, psz, nLen); | |
159 | return buf; | |
160 | } | |
161 | else | |
162 | return wxWCharBuffer((wchar_t *) NULL); | |
163 | } | |
164 | ||
165 | const wxCharBuffer wxMBConv::cWC2MB(const wchar_t *psz) const | |
166 | { | |
167 | if (psz) | |
168 | { | |
169 | size_t nLen = WC2MB((char *) NULL, psz, 0); | |
170 | if (nLen == (size_t)-1) | |
171 | return wxCharBuffer((char *) NULL); | |
172 | wxCharBuffer buf(nLen); | |
173 | WC2MB((char *)(const char *) buf, psz, nLen); | |
174 | return buf; | |
175 | } | |
176 | else | |
177 | return wxCharBuffer((char *) NULL); | |
178 | } | |
179 | ||
180 | // ---------------------------------------------------------------------------- | |
181 | // standard file conversion | |
182 | // ---------------------------------------------------------------------------- | |
183 | ||
184 | WXDLLEXPORT_DATA(wxMBConvFile) wxConvFile; | |
185 | ||
186 | // just use the libc conversion for now | |
187 | size_t wxMBConvFile::MB2WC(wchar_t *buf, const char *psz, size_t n) const | |
188 | { | |
189 | return wxMB2WC(buf, psz, n); | |
190 | } | |
191 | ||
192 | size_t wxMBConvFile::WC2MB(char *buf, const wchar_t *psz, size_t n) const | |
193 | { | |
194 | return wxWC2MB(buf, psz, n); | |
195 | } | |
196 | ||
197 | // ---------------------------------------------------------------------------- | |
198 | // standard gdk conversion | |
199 | // ---------------------------------------------------------------------------- | |
200 | ||
201 | #ifdef __WXGTK12__ | |
202 | ||
203 | WXDLLEXPORT_DATA(wxMBConvGdk) wxConvGdk; | |
204 | ||
205 | #include <gdk/gdk.h> | |
206 | ||
207 | size_t wxMBConvGdk::MB2WC(wchar_t *buf, const char *psz, size_t n) const | |
208 | { | |
209 | if (buf) | |
210 | { | |
211 | return gdk_mbstowcs((GdkWChar *)buf, psz, n); | |
212 | } | |
213 | else | |
214 | { | |
215 | GdkWChar *nbuf = new GdkWChar[n=strlen(psz)]; | |
216 | size_t len = gdk_mbstowcs(nbuf, psz, n); | |
217 | delete[] nbuf; | |
218 | return len; | |
219 | } | |
220 | } | |
221 | ||
222 | size_t wxMBConvGdk::WC2MB(char *buf, const wchar_t *psz, size_t n) const | |
223 | { | |
224 | char *mbstr = gdk_wcstombs((GdkWChar *)psz); | |
225 | size_t len = mbstr ? strlen(mbstr) : 0; | |
226 | if (buf) | |
227 | { | |
228 | if (len > n) | |
229 | len = n; | |
230 | memcpy(buf, psz, len); | |
231 | if (len < n) | |
232 | buf[len] = 0; | |
233 | } | |
234 | return len; | |
235 | } | |
236 | ||
237 | #endif // GTK > 1.0 | |
238 | ||
239 | // ---------------------------------------------------------------------------- | |
240 | // UTF-7 | |
241 | // ---------------------------------------------------------------------------- | |
242 | ||
243 | WXDLLEXPORT_DATA(wxMBConvUTF7) wxConvUTF7; | |
244 | ||
245 | #if 0 | |
246 | static char utf7_setD[]="ABCDEFGHIJKLMNOPQRSTUVWXYZ" | |
247 | "abcdefghijklmnopqrstuvwxyz" | |
248 | "0123456789'(),-./:?"; | |
249 | static char utf7_setO[]="!\"#$%&*;<=>@[]^_`{|}"; | |
250 | static char utf7_setB[]="ABCDEFGHIJKLMNOPQRSTUVWXYZ" | |
251 | "abcdefghijklmnopqrstuvwxyz" | |
252 | "0123456789+/"; | |
253 | #endif | |
254 | ||
255 | // TODO: write actual implementations of UTF-7 here | |
256 | size_t wxMBConvUTF7::MB2WC(wchar_t * WXUNUSED(buf), | |
257 | const char * WXUNUSED(psz), | |
258 | size_t WXUNUSED(n)) const | |
259 | { | |
260 | return 0; | |
261 | } | |
262 | ||
263 | size_t wxMBConvUTF7::WC2MB(char * WXUNUSED(buf), | |
264 | const wchar_t * WXUNUSED(psz), | |
265 | size_t WXUNUSED(n)) const | |
266 | { | |
267 | return 0; | |
268 | } | |
269 | ||
270 | // ---------------------------------------------------------------------------- | |
271 | // UTF-8 | |
272 | // ---------------------------------------------------------------------------- | |
273 | ||
274 | WXDLLEXPORT_DATA(wxMBConvUTF8) wxConvUTF8; | |
275 | ||
276 | static wxUint32 utf8_max[]= | |
277 | { 0x7f, 0x7ff, 0xffff, 0x1fffff, 0x3ffffff, 0x7fffffff, 0xffffffff }; | |
278 | ||
279 | size_t wxMBConvUTF8::MB2WC(wchar_t *buf, const char *psz, size_t n) const | |
280 | { | |
281 | size_t len = 0; | |
282 | ||
283 | while (*psz && ((!buf) || (len < n))) | |
284 | { | |
285 | unsigned char cc = *psz++, fc = cc; | |
286 | unsigned cnt; | |
287 | for (cnt = 0; fc & 0x80; cnt++) | |
288 | fc <<= 1; | |
289 | if (!cnt) | |
290 | { | |
291 | // plain ASCII char | |
292 | if (buf) | |
293 | *buf++ = cc; | |
294 | len++; | |
295 | } | |
296 | else | |
297 | { | |
298 | cnt--; | |
299 | if (!cnt) | |
300 | { | |
301 | // invalid UTF-8 sequence | |
302 | return (size_t)-1; | |
303 | } | |
304 | else | |
305 | { | |
306 | unsigned ocnt = cnt - 1; | |
307 | wxUint32 res = cc & (0x3f >> cnt); | |
308 | while (cnt--) | |
309 | { | |
310 | cc = *psz++; | |
311 | if ((cc & 0xC0) != 0x80) | |
312 | { | |
313 | // invalid UTF-8 sequence | |
314 | return (size_t)-1; | |
315 | } | |
316 | res = (res << 6) | (cc & 0x3f); | |
317 | } | |
318 | if (res <= utf8_max[ocnt]) | |
319 | { | |
320 | // illegal UTF-8 encoding | |
321 | return (size_t)-1; | |
322 | } | |
323 | #ifdef WC_UTF16 | |
324 | size_t pa = encode_utf16(res, buf); | |
325 | if (pa == (size_t)-1) | |
326 | return (size_t)-1; | |
327 | if (buf) | |
328 | buf += pa; | |
329 | len += pa; | |
330 | #else | |
331 | if (buf) | |
332 | *buf++ = res; | |
333 | len++; | |
334 | #endif | |
335 | } | |
336 | } | |
337 | } | |
338 | if (buf && (len < n)) | |
339 | *buf = 0; | |
340 | return len; | |
341 | } | |
342 | ||
343 | size_t wxMBConvUTF8::WC2MB(char *buf, const wchar_t *psz, size_t n) const | |
344 | { | |
345 | size_t len = 0; | |
346 | ||
347 | while (*psz && ((!buf) || (len < n))) | |
348 | { | |
349 | wxUint32 cc; | |
350 | #ifdef WC_UTF16 | |
351 | size_t pa = decode_utf16(psz,cc); | |
352 | psz += (pa == (size_t)-1) ? 1 : pa; | |
353 | #else | |
354 | cc=(*psz++) & 0x7fffffff; | |
355 | #endif | |
356 | unsigned cnt; | |
357 | for (cnt = 0; cc > utf8_max[cnt]; cnt++) {} | |
358 | if (!cnt) | |
359 | { | |
360 | // plain ASCII char | |
361 | if (buf) | |
362 | *buf++ = cc; | |
363 | len++; | |
364 | } | |
365 | ||
366 | else | |
367 | { | |
368 | len += cnt + 1; | |
369 | if (buf) | |
370 | { | |
371 | *buf++ = (-128 >> cnt) | ((cc >> (cnt * 6)) & (0x3f >> cnt)); | |
372 | while (cnt--) | |
373 | *buf++ = 0x80 | ((cc >> (cnt * 6)) & 0x3f); | |
374 | } | |
375 | } | |
376 | } | |
377 | ||
378 | if (buf && (len<n)) *buf = 0; | |
379 | return len; | |
380 | } | |
381 | ||
382 | // ---------------------------------------------------------------------------- | |
383 | // specified character set | |
384 | // ---------------------------------------------------------------------------- | |
385 | ||
386 | WXDLLEXPORT_DATA(wxCSConv) wxConvLocal((const wxChar *)NULL); | |
387 | ||
388 | #include "wx/encconv.h" | |
389 | #include "wx/fontmap.h" | |
390 | ||
391 | // TODO: add some tables here | |
392 | // - perhaps common encodings to common codepages (for Win32) | |
393 | // - perhaps common encodings to objects ("UTF8" -> wxConvUTF8) | |
394 | // - move wxEncodingConverter meat in here | |
395 | ||
396 | #ifdef __WIN32__ | |
397 | #include "wx/msw/registry.h" | |
398 | // this should work if M$ Internet Exploiter is installed | |
399 | static long CharsetToCodepage(const wxChar *name) | |
400 | { | |
401 | if (!name) | |
402 | return GetACP(); | |
403 | ||
404 | long CP=-1; | |
405 | ||
406 | wxString cn(name); | |
407 | do { | |
408 | wxString path(wxT("MIME\\Database\\Charset\\")); | |
409 | path += cn; | |
410 | wxRegKey key(wxRegKey::HKCR, path); | |
411 | ||
412 | if (!key.Exists()) continue; | |
413 | ||
414 | // two cases: either there's an AliasForCharset string, | |
415 | // or there are Codepage and InternetEncoding dwords. | |
416 | // The InternetEncoding gives us the actual encoding, | |
417 | // the Codepage just says which Windows character set to | |
418 | // use when displaying the data. | |
419 | if (key.HasValue(wxT("InternetEncoding")) && | |
420 | key.QueryValue(wxT("InternetEncoding"), &CP)) break; | |
421 | ||
422 | // no encoding, see if it's an alias | |
423 | if (!key.HasValue(wxT("AliasForCharset")) || | |
424 | !key.QueryValue(wxT("AliasForCharset"), cn)) break; | |
425 | } while (1); | |
426 | ||
427 | return CP; | |
428 | } | |
429 | #endif | |
430 | ||
431 | class wxCharacterSet | |
432 | { | |
433 | public: | |
434 | wxCharacterSet(const wxChar*name) | |
435 | : cname(name) {} | |
436 | virtual ~wxCharacterSet() | |
437 | {} | |
438 | virtual size_t MB2WC(wchar_t *buf, const char *psz, size_t n) | |
439 | { return (size_t)-1; } | |
440 | virtual size_t WC2MB(char *buf, const wchar_t *psz, size_t n) | |
441 | { return (size_t)-1; } | |
442 | virtual bool usable() | |
443 | { return FALSE; } | |
444 | public: | |
445 | const wxChar*cname; | |
446 | }; | |
447 | ||
448 | class ID_CharSet : public wxCharacterSet | |
449 | { | |
450 | public: | |
451 | ID_CharSet(const wxChar *name,wxMBConv *cnv) | |
452 | : wxCharacterSet(name), work(cnv) {} | |
453 | ||
454 | size_t MB2WC(wchar_t *buf, const char *psz, size_t n) | |
455 | { return work ? work->MB2WC(buf,psz,n) : (size_t)-1; } | |
456 | ||
457 | size_t WC2MB(char *buf, const wchar_t *psz, size_t n) | |
458 | { return work ? work->WC2MB(buf,psz,n) : (size_t)-1; } | |
459 | ||
460 | bool usable() | |
461 | { return work!=NULL; } | |
462 | public: | |
463 | wxMBConv*work; | |
464 | }; | |
465 | ||
466 | ||
467 | #ifdef HAVE_ICONV_H | |
468 | ||
469 | // VS: glibc 2.1.3 is broken in that iconv() conversion to/from UCS4 fails with E2BIG | |
470 | // if output buffer is _exactly_ as big as needed. Such case is (unless there's | |
471 | // yet another bug in glibc) the only case when iconv() returns with (size_t)-1 | |
472 | // (which means error) and says there are 0 bytes left in the input buffer -- | |
473 | // when _real_ error occurs, bytes-left-in-input buffer is non-zero. Hence, | |
474 | // this alternative test for iconv() failure. | |
475 | // [This bug does not appear in glibc 2.2.] | |
476 | #if defined(__GLIBC__) && __GLIBC__ == 2 && __GLIBC_MINOR__ <= 1 | |
477 | #define ICONV_FAILED(cres, bufLeft) ((cres == (size_t)-1) && \ | |
478 | (errno != E2BIG || bufLeft != 0)) | |
479 | #else | |
480 | #define ICONV_FAILED(cres, bufLeft) (cres == (size_t)-1) | |
481 | #endif | |
482 | ||
483 | class IC_CharSet : public wxCharacterSet | |
484 | { | |
485 | public: | |
486 | IC_CharSet(const wxChar *name) | |
487 | : wxCharacterSet(name) | |
488 | { | |
489 | m2w = iconv_open(WC_NAME, wxConvLibc.cWX2MB(cname)); | |
490 | w2m = iconv_open(wxConvLibc.cWX2MB(cname), WC_NAME); | |
491 | } | |
492 | ||
493 | ~IC_CharSet() | |
494 | { | |
495 | if ( m2w != (iconv_t)-1 ) | |
496 | iconv_close(m2w); | |
497 | if ( w2m != (iconv_t)-1 ) | |
498 | iconv_close(w2m); | |
499 | } | |
500 | ||
501 | size_t MB2WC(wchar_t *buf, const char *psz, size_t n) | |
502 | { | |
503 | size_t inbuf = strlen(psz); | |
504 | size_t outbuf = n * SIZEOF_WCHAR_T; | |
505 | size_t res, cres; | |
506 | // VS: Use these instead of psz, buf because iconv() modifies its arguments: | |
507 | wchar_t *bufPtr = buf; | |
508 | const char *pszPtr = psz; | |
509 | ||
510 | if (buf) | |
511 | { | |
512 | // have destination buffer, convert there | |
513 | #ifdef WX_ICONV_TAKES_CHAR | |
514 | cres = iconv(m2w, (char**)&pszPtr, &inbuf, (char**)&bufPtr, &outbuf); | |
515 | #else | |
516 | cres = iconv(m2w, &pszPtr, &inbuf, (char**)&bufPtr, &outbuf); | |
517 | #endif | |
518 | res = n - (outbuf / SIZEOF_WCHAR_T); | |
519 | // convert to native endianness | |
520 | #ifdef WC_NEED_BSWAP | |
521 | WC_BSWAP(buf /* _not_ bufPtr */, res) | |
522 | #endif | |
523 | } | |
524 | else | |
525 | { | |
526 | // no destination buffer... convert using temp buffer | |
527 | // to calculate destination buffer requirement | |
528 | wchar_t tbuf[8]; | |
529 | res = 0; | |
530 | do { | |
531 | bufPtr = tbuf; outbuf = 8*SIZEOF_WCHAR_T; | |
532 | #ifdef WX_ICONV_TAKES_CHAR | |
533 | cres = iconv( m2w, (char**)&pszPtr, &inbuf, (char**)&bufPtr, &outbuf ); | |
534 | #else | |
535 | cres = iconv( m2w, &pszPtr, &inbuf, (char**)&bufPtr, &outbuf ); | |
536 | #endif | |
537 | res += 8-(outbuf/SIZEOF_WCHAR_T); | |
538 | } while ((cres==(size_t)-1) && (errno==E2BIG)); | |
539 | } | |
540 | ||
541 | if (ICONV_FAILED(cres, inbuf)) | |
542 | return (size_t)-1; | |
543 | ||
544 | return res; | |
545 | } | |
546 | ||
547 | size_t WC2MB(char *buf, const wchar_t *psz, size_t n) | |
548 | { | |
549 | #if defined(__BORLANDC__) && (__BORLANDC__ > 0x530) | |
550 | size_t inbuf = std::wcslen(psz) * SIZEOF_WCHAR_T; | |
551 | #else | |
552 | size_t inbuf = ::wcslen(psz) * SIZEOF_WCHAR_T; | |
553 | #endif | |
554 | size_t outbuf = n; | |
555 | size_t res, cres; | |
556 | ||
557 | #ifdef WC_NEED_BSWAP | |
558 | // need to copy to temp buffer to switch endianness | |
559 | // this absolutely doesn't rock! | |
560 | // (no, doing WC_BSWAP twice on the original buffer won't help, as it | |
561 | // could be in read-only memory, or be accessed in some other thread) | |
562 | wchar_t *tmpbuf=(wchar_t*)malloc((inbuf+1)*SIZEOF_WCHAR_T); | |
563 | memcpy(tmpbuf,psz,(inbuf+1)*SIZEOF_WCHAR_T); | |
564 | WC_BSWAP(tmpbuf, inbuf) | |
565 | psz=tmpbuf; | |
566 | #endif | |
567 | if (buf) | |
568 | { | |
569 | // have destination buffer, convert there | |
570 | #ifdef WX_ICONV_TAKES_CHAR | |
571 | cres = iconv( w2m, (char**)&psz, &inbuf, &buf, &outbuf ); | |
572 | #else | |
573 | cres = iconv( w2m, (const char**)&psz, &inbuf, &buf, &outbuf ); | |
574 | #endif | |
575 | res = n-outbuf; | |
576 | } | |
577 | else | |
578 | { | |
579 | // no destination buffer... convert using temp buffer | |
580 | // to calculate destination buffer requirement | |
581 | char tbuf[16]; | |
582 | res = 0; | |
583 | do { | |
584 | buf = tbuf; outbuf = 16; | |
585 | #ifdef WX_ICONV_TAKES_CHAR | |
586 | cres = iconv( w2m, (char**)&psz, &inbuf, &buf, &outbuf ); | |
587 | #else | |
588 | cres = iconv( w2m, (const char**)&psz, &inbuf, &buf, &outbuf ); | |
589 | #endif | |
590 | res += 16 - outbuf; | |
591 | } while ((cres==(size_t)-1) && (errno==E2BIG)); | |
592 | } | |
593 | #ifdef WC_NEED_BSWAP | |
594 | free(tmpbuf); | |
595 | #endif | |
596 | if (ICONV_FAILED(cres, inbuf)) | |
597 | return (size_t)-1; | |
598 | ||
599 | return res; | |
600 | } | |
601 | ||
602 | bool usable() | |
603 | { return (m2w != (iconv_t)-1) && (w2m != (iconv_t)-1); } | |
604 | ||
605 | public: | |
606 | iconv_t m2w, w2m; | |
607 | }; | |
608 | #endif | |
609 | ||
610 | #ifdef __WIN32__ | |
611 | class CP_CharSet : public wxCharacterSet | |
612 | { | |
613 | public: | |
614 | CP_CharSet(const wxChar*name) | |
615 | : wxCharacterSet(name), CodePage(CharsetToCodepage(name)) {} | |
616 | ||
617 | size_t MB2WC(wchar_t *buf, const char *psz, size_t n) | |
618 | { | |
619 | size_t len = | |
620 | MultiByteToWideChar(CodePage, 0, psz, -1, buf, buf ? n : 0); | |
621 | return len ? len : (size_t)-1; | |
622 | } | |
623 | ||
624 | size_t WC2MB(char *buf, const wchar_t *psz, size_t n) | |
625 | { | |
626 | size_t len = WideCharToMultiByte(CodePage, 0, psz, -1, buf, | |
627 | buf ? n : 0, NULL, NULL); | |
628 | return len ? len : (size_t)-1; | |
629 | } | |
630 | ||
631 | bool usable() | |
632 | { return CodePage != -1; } | |
633 | ||
634 | public: | |
635 | long CodePage; | |
636 | }; | |
637 | #endif | |
638 | ||
639 | class EC_CharSet : public wxCharacterSet | |
640 | { | |
641 | public: | |
642 | // temporarily just use wxEncodingConverter stuff, | |
643 | // so that it works while a better implementation is built | |
644 | EC_CharSet(const wxChar*name) : wxCharacterSet(name), | |
645 | enc(wxFONTENCODING_SYSTEM) | |
646 | { | |
647 | if (name) | |
648 | enc = wxTheFontMapper->CharsetToEncoding(name, FALSE); | |
649 | m2w.Init(enc, wxFONTENCODING_UNICODE); | |
650 | w2m.Init(wxFONTENCODING_UNICODE, enc); | |
651 | } | |
652 | ||
653 | size_t MB2WC(wchar_t *buf, const char *psz, size_t n) | |
654 | { | |
655 | size_t inbuf = strlen(psz); | |
656 | if (buf) | |
657 | m2w.Convert(psz,buf); | |
658 | return inbuf; | |
659 | } | |
660 | ||
661 | size_t WC2MB(char *buf, const wchar_t *psz, size_t n) | |
662 | { | |
663 | #if defined(__BORLANDC__) && (__BORLANDC__ > 0x530) | |
664 | size_t inbuf = std::wcslen(psz); | |
665 | #else | |
666 | size_t inbuf = ::wcslen(psz); | |
667 | #endif | |
668 | if (buf) | |
669 | w2m.Convert(psz,buf); | |
670 | ||
671 | return inbuf; | |
672 | } | |
673 | ||
674 | bool usable() | |
675 | { return (enc!=wxFONTENCODING_SYSTEM) && (enc!=wxFONTENCODING_DEFAULT); } | |
676 | ||
677 | public: | |
678 | wxFontEncoding enc; | |
679 | wxEncodingConverter m2w, w2m; | |
680 | }; | |
681 | ||
682 | static wxCharacterSet *wxGetCharacterSet(const wxChar *name) | |
683 | { | |
684 | wxCharacterSet *cset = NULL; | |
685 | if (name) | |
686 | { | |
687 | if (wxStricmp(name, wxT("UTF8")) == 0 || wxStricmp(name, wxT("UTF-8")) == 0) | |
688 | { | |
689 | cset = new ID_CharSet(name, &wxConvUTF8); | |
690 | } | |
691 | else | |
692 | { | |
693 | #ifdef HAVE_ICONV_H | |
694 | cset = new IC_CharSet(name); // may not take NULL | |
695 | #endif | |
696 | } | |
697 | } | |
698 | ||
699 | if (cset && cset->usable()) return cset; | |
700 | if (cset) | |
701 | { | |
702 | delete cset; | |
703 | cset = NULL; | |
704 | } | |
705 | ||
706 | #ifdef __WIN32__ | |
707 | cset = new CP_CharSet(name); // may take NULL | |
708 | if (cset->usable()) | |
709 | return cset; | |
710 | ||
711 | delete cset; | |
712 | #endif // __WIN32__ | |
713 | ||
714 | cset = new EC_CharSet(name); | |
715 | if (cset->usable()) | |
716 | return cset; | |
717 | ||
718 | delete cset; | |
719 | wxLogError(_("Unknown encoding '%s'!"), name); | |
720 | return NULL; | |
721 | } | |
722 | ||
723 | wxCSConv::wxCSConv(const wxChar *charset) | |
724 | { | |
725 | m_name = (wxChar *)NULL; | |
726 | m_cset = (wxCharacterSet *) NULL; | |
727 | m_deferred = TRUE; | |
728 | ||
729 | SetName(charset); | |
730 | } | |
731 | ||
732 | wxCSConv::~wxCSConv() | |
733 | { | |
734 | free(m_name); | |
735 | delete m_cset; | |
736 | } | |
737 | ||
738 | void wxCSConv::SetName(const wxChar *charset) | |
739 | { | |
740 | if (charset) | |
741 | { | |
742 | m_name = wxStrdup(charset); | |
743 | m_deferred = TRUE; | |
744 | } | |
745 | } | |
746 | ||
747 | void wxCSConv::LoadNow() | |
748 | { | |
749 | if (m_deferred) | |
750 | { | |
751 | if ( !m_name ) | |
752 | { | |
753 | wxString name = wxLocale::GetSystemEncodingName(); | |
754 | if ( !name.empty() ) | |
755 | SetName(name); | |
756 | } | |
757 | ||
758 | m_cset = wxGetCharacterSet(m_name); | |
759 | m_deferred = FALSE; | |
760 | } | |
761 | } | |
762 | ||
763 | size_t wxCSConv::MB2WC(wchar_t *buf, const char *psz, size_t n) const | |
764 | { | |
765 | ((wxCSConv *)this)->LoadNow(); // discard constness | |
766 | ||
767 | if (m_cset) | |
768 | return m_cset->MB2WC(buf, psz, n); | |
769 | ||
770 | // latin-1 (direct) | |
771 | size_t len = strlen(psz); | |
772 | ||
773 | if (buf) | |
774 | { | |
775 | for (size_t c = 0; c <= len; c++) | |
776 | buf[c] = (unsigned char)(psz[c]); | |
777 | } | |
778 | ||
779 | return len; | |
780 | } | |
781 | ||
782 | size_t wxCSConv::WC2MB(char *buf, const wchar_t *psz, size_t n) const | |
783 | { | |
784 | ((wxCSConv *)this)->LoadNow(); // discard constness | |
785 | ||
786 | if (m_cset) | |
787 | return m_cset->WC2MB(buf, psz, n); | |
788 | ||
789 | // latin-1 (direct) | |
790 | #if defined(__BORLANDC__) && (__BORLANDC__ > 0x530) | |
791 | size_t len=std::wcslen(psz); | |
792 | #else | |
793 | size_t len=::wcslen(psz); | |
794 | #endif | |
795 | if (buf) | |
796 | { | |
797 | for (size_t c = 0; c <= len; c++) | |
798 | buf[c] = (psz[c] > 0xff) ? '?' : psz[c]; | |
799 | } | |
800 | ||
801 | return len; | |
802 | } | |
803 | ||
804 | #ifdef HAVE_ICONV_H | |
805 | ||
806 | class IC_CharSetConverter | |
807 | { | |
808 | public: | |
809 | IC_CharSetConverter(IC_CharSet *from, IC_CharSet *to) | |
810 | { | |
811 | cnv = iconv_open(wxConvLibc.cWX2MB(to->cname), | |
812 | wxConvLibc.cWX2MB(from->cname)); | |
813 | } | |
814 | ||
815 | ~IC_CharSetConverter() | |
816 | { | |
817 | if (cnv != (iconv_t)-1) | |
818 | iconv_close(cnv); | |
819 | } | |
820 | ||
821 | size_t Convert(char *buf, const char *psz, size_t n) | |
822 | { | |
823 | size_t inbuf = strlen(psz); | |
824 | size_t outbuf = n; | |
825 | #ifdef WX_ICONV_TAKES_CHAR | |
826 | size_t res = iconv( cnv, (char**)&psz, &inbuf, &buf, &outbuf ); | |
827 | #else | |
828 | size_t res = iconv( cnv, &psz, &inbuf, &buf, &outbuf ); | |
829 | #endif | |
830 | if (res == (size_t)-1) | |
831 | return (size_t)-1; | |
832 | return (n - outbuf); | |
833 | } | |
834 | ||
835 | public: | |
836 | iconv_t cnv; | |
837 | }; | |
838 | ||
839 | #endif // HAVE_ICONV_H | |
840 | ||
841 | class EC_CharSetConverter | |
842 | { | |
843 | public: | |
844 | EC_CharSetConverter(EC_CharSet*from,EC_CharSet*to) | |
845 | { cnv.Init(from->enc,to->enc); } | |
846 | ||
847 | size_t Convert(char*buf, const char*psz, size_t n) | |
848 | { | |
849 | size_t inbuf = strlen(psz); | |
850 | if (buf) cnv.Convert(psz,buf); | |
851 | return inbuf; | |
852 | } | |
853 | ||
854 | public: | |
855 | wxEncodingConverter cnv; | |
856 | }; | |
857 | ||
858 | #else // !wxUSE_WCHAR_T | |
859 | ||
860 | // ---------------------------------------------------------------------------- | |
861 | // stand-ins in absence of wchar_t | |
862 | // ---------------------------------------------------------------------------- | |
863 | ||
864 | WXDLLEXPORT_DATA(wxMBConv) wxConvLibc, wxConvFile; | |
865 | ||
866 | #endif // wxUSE_WCHAR_T | |
867 | ||
868 |