?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
Merge branch '896-variant-type-docs' into 'master'
[glib.git] / glib / win_iconv.c
blob605e5f08d3da36dc0640bf8a40d1f6a7d7538727
1 /*
2 * iconv library implemented with Win32 API.
3 *
4 * This file is placed in the public domain.
5 *
6 * Maintainer: Yukihiro Nakadaira <yukihiro.nakadaira@gmail.com>
7 *
8 * If $WINICONV_LIBICONV_DLL environment variable was defined, win_iconv
9 * loads the specified DLL dynamically and uses it. If loading the DLL
10 * or iconv_open() failed, falls back to internal conversion.
11 * $WINICONV_LIBICONV_DLL is a comma separated list. The first loadable
12 * DLL is used. The specified DLL should have iconv_open(),
13 * iconv_close() and iconv() functions. Or these functions can be
14 * libiconv_open(), libiconv_close() and libiconv().
15 *
16 * Win32 API does not support strict encoding conversion for some
17 * codepage. And MLang function drop or replace invalid bytes and does
18 * not return useful error status as iconv. This implementation cannot
19 * be used for encoding validation purpose.
20 */
21
22 /* for WC_NO_BEST_FIT_CHARS */
23 #ifndef WINVER
24 # define WINVER 0x0500
25 #endif
26
27 #define STRICT
28 #include <windows.h>
29 #include <errno.h>
30 #include <string.h>
31 #include <stdlib.h>
32
33 #ifdef __GNUC__
34 #define UNUSED __attribute__((unused))
35 #else
36 #define UNUSED
37 #endif
38
39 /* WORKAROUND: */
40 #ifndef UNDER_CE
41 #define GetProcAddressA GetProcAddress
42 #endif
43
44 #if 0
45 # define MAKE_EXE
46 # define MAKE_DLL
47 # define USE_LIBICONV_DLL
48 #endif
49
50 #if !defined(DEFAULT_LIBICONV_DLL)
51 # define DEFAULT_LIBICONV_DLL ""
52 #endif
53
54 #define MB_CHAR_MAX 16
55
56 #define UNICODE_MODE_BOM_DONE 1
57 #define UNICODE_MODE_SWAPPED 2
58
59 #define FLAG_USE_BOM 1
60 #define FLAG_TRANSLIT 2 /* //TRANSLIT */
61 #define FLAG_IGNORE 4 /* //IGNORE */
62
63 typedef unsigned char uchar;
64 typedef unsigned short ushort;
65 typedef unsigned int uint;
66
67 typedef void* iconv_t;
68
69 iconv_t iconv_open(const char *tocode, const char *fromcode);
70 int iconv_close(iconv_t cd);
71 size_t iconv(iconv_t cd, /* const */ char **inbuf, size_t *inbytesleft, char **outbuf, size_t *outbytesleft);
72
73 /* libiconv interface for vim */
74 #if defined(MAKE_DLL)
75 int
76 iconvctl (iconv_t cd, int request, void* argument)
77 {
78 /* not supported */
79 return 0;
80 }
81 #endif
82
83 typedef struct compat_t compat_t;
84 typedef struct csconv_t csconv_t;
85 typedef struct rec_iconv_t rec_iconv_t;
86
87 typedef iconv_t (*f_iconv_open)(const char *tocode, const char *fromcode);
88 typedef int (*f_iconv_close)(iconv_t cd);
89 typedef size_t (*f_iconv)(iconv_t cd, /* const */ char **inbuf, size_t *inbytesleft, char **outbuf, size_t *outbytesleft);
90 typedef int* (*f_errno)(void);
91 typedef int (*f_mbtowc)(csconv_t *cv, const uchar *buf, int bufsize, ushort *wbuf, int *wbufsize);
92 typedef int (*f_wctomb)(csconv_t *cv, ushort *wbuf, int wbufsize, uchar *buf, int bufsize);
93 typedef int (*f_mblen)(csconv_t *cv, const uchar *buf, int bufsize);
94 typedef int (*f_flush)(csconv_t *cv, uchar *buf, int bufsize);
95
96 #define COMPAT_IN 1
97 #define COMPAT_OUT 2
98
99 /* unicode mapping for compatibility with other conversion table. */
100 struct compat_t {
101 uint in;
102 uint out;
103 uint flag;
104 };
105
106 struct csconv_t {
107 int codepage;
108 int flags;
109 f_mbtowc mbtowc;
110 f_wctomb wctomb;
111 f_mblen mblen;
112 f_flush flush;
113 DWORD mode;
114 compat_t *compat;
115 };
116
117 struct rec_iconv_t {
118 iconv_t cd;
119 f_iconv_close iconv_close;
120 f_iconv iconv;
121 f_errno _errno;
122 csconv_t from;
123 csconv_t to;
124 #if defined(USE_LIBICONV_DLL)
125 HMODULE hlibiconv;
126 #endif
127 };
128
129 static int win_iconv_open(rec_iconv_t *cd, const char *tocode, const char *fromcode);
130 static int win_iconv_close(iconv_t cd);
131 static size_t win_iconv(iconv_t cd, /* const */ char **inbuf, size_t *inbytesleft, char **outbuf, size_t *outbytesleft);
132
133 static int load_mlang(void);
134 static int make_csconv(const char *name, csconv_t *cv);
135 static int name_to_codepage(const char *name);
136 static uint utf16_to_ucs4(const ushort *wbuf);
137 static void ucs4_to_utf16(uint wc, ushort *wbuf, int *wbufsize);
138 static int mbtowc_flags(int codepage);
139 static int must_use_null_useddefaultchar(int codepage);
140 static char *strrstr(const char *str, const char *token);
141 static char *xstrndup(const char *s, size_t n);
142 static int seterror(int err);
143
144 #if defined(USE_LIBICONV_DLL)
145 static int libiconv_iconv_open(rec_iconv_t *cd, const char *tocode, const char *fromcode);
146 static PVOID MyImageDirectoryEntryToData(LPVOID Base, BOOLEAN MappedAsImage, USHORT DirectoryEntry, PULONG Size);
147 static FARPROC find_imported_function(HMODULE hModule, const char *funcname);
148
149 static HMODULE hwiniconv;
150 #endif
151
152 static int sbcs_mblen(csconv_t *cv, const uchar *buf, int bufsize);
153 static int dbcs_mblen(csconv_t *cv, const uchar *buf, int bufsize);
154 static int mbcs_mblen(csconv_t *cv, const uchar *buf, int bufsize);
155 static int utf8_mblen(csconv_t *cv, const uchar *buf, int bufsize);
156 static int eucjp_mblen(csconv_t *cv, const uchar *buf, int bufsize);
157
158 static int kernel_mbtowc(csconv_t *cv, const uchar *buf, int bufsize, ushort *wbuf, int *wbufsize);
159 static int kernel_wctomb(csconv_t *cv, ushort *wbuf, int wbufsize, uchar *buf, int bufsize);
160 static int mlang_mbtowc(csconv_t *cv, const uchar *buf, int bufsize, ushort *wbuf, int *wbufsize);
161 static int mlang_wctomb(csconv_t *cv, ushort *wbuf, int wbufsize, uchar *buf, int bufsize);
162 static int utf16_mbtowc(csconv_t *cv, const uchar *buf, int bufsize, ushort *wbuf, int *wbufsize);
163 static int utf16_wctomb(csconv_t *cv, ushort *wbuf, int wbufsize, uchar *buf, int bufsize);
164 static int utf32_mbtowc(csconv_t *cv, const uchar *buf, int bufsize, ushort *wbuf, int *wbufsize);
165 static int utf32_wctomb(csconv_t *cv, ushort *wbuf, int wbufsize, uchar *buf, int bufsize);
166 static int iso2022jp_mbtowc(csconv_t *cv, const uchar *buf, int bufsize, ushort *wbuf, int *wbufsize);
167 static int iso2022jp_wctomb(csconv_t *cv, ushort *wbuf, int wbufsize, uchar *buf, int bufsize);
168 static int iso2022jp_flush(csconv_t *cv, uchar *buf, int bufsize);
169
170 static struct {
171 int codepage;
172 const char *name;
173 } codepage_alias[] = {
174 {65001, "CP65001"},
175 {65001, "UTF8"},
176 {65001, "UTF-8"},
177
178 {1200, "CP1200"},
179 {1200, "UTF16LE"},
180 {1200, "UTF-16LE"},
181 {1200, "UCS2LE"},
182 {1200, "UCS-2LE"},
183
184 {1201, "CP1201"},
185 {1201, "UTF16BE"},
186 {1201, "UTF-16BE"},
187 {1201, "UCS2BE"},
188 {1201, "UCS-2BE"},
189 {1201, "unicodeFFFE"},
190
191 {12000, "CP12000"},
192 {12000, "UTF32LE"},
193 {12000, "UTF-32LE"},
194 {12000, "UCS4LE"},
195 {12000, "UCS-4LE"},
196
197 {12001, "CP12001"},
198 {12001, "UTF32BE"},
199 {12001, "UTF-32BE"},
200 {12001, "UCS4BE"},
201 {12001, "UCS-4BE"},
202
203 #ifndef GLIB_COMPILATION
204 /*
205 * Default is big endian.
206 * See rfc2781 4.3 Interpreting text labelled as UTF-16.
207 */
208 {1201, "UTF16"},
209 {1201, "UTF-16"},
210 {1201, "UCS2"},
211 {1201, "UCS-2"},
212 {12001, "UTF32"},
213 {12001, "UTF-32"},
214 {12001, "UCS-4"},
215 {12001, "UCS4"},
216 #else
217 /* Default is little endian, because the platform is */
218 {1200, "UTF16"},
219 {1200, "UTF-16"},
220 {1200, "UCS2"},
221 {1200, "UCS-2"},
222 {12000, "UTF32"},
223 {12000, "UTF-32"},
224 {12000, "UCS4"},
225 {12000, "UCS-4"},
226 #endif
227
228 /* copy from libiconv `iconv -l` */
229 /* !IsValidCodePage(367) */
230 {20127, "ANSI_X3.4-1968"},
231 {20127, "ANSI_X3.4-1986"},
232 {20127, "ASCII"},
233 {20127, "CP367"},
234 {20127, "IBM367"},
235 {20127, "ISO-IR-6"},
236 {20127, "ISO646-US"},
237 {20127, "ISO_646.IRV:1991"},
238 {20127, "US"},
239 {20127, "US-ASCII"},
240 {20127, "CSASCII"},
241
242 /* !IsValidCodePage(819) */
243 {1252, "CP819"},
244 {1252, "IBM819"},
245 {28591, "ISO-8859-1"},
246 {28591, "ISO-IR-100"},
247 {28591, "ISO8859-1"},
248 {28591, "ISO_8859-1"},
249 {28591, "ISO_8859-1:1987"},
250 {28591, "L1"},
251 {28591, "LATIN1"},
252 {28591, "CSISOLATIN1"},
253
254 {1250, "CP1250"},
255 {1250, "MS-EE"},
256 {1250, "WINDOWS-1250"},
257
258 {1251, "CP1251"},
259 {1251, "MS-CYRL"},
260 {1251, "WINDOWS-1251"},
261
262 {1252, "CP1252"},
263 {1252, "MS-ANSI"},
264 {1252, "WINDOWS-1252"},
265
266 {1253, "CP1253"},
267 {1253, "MS-GREEK"},
268 {1253, "WINDOWS-1253"},
269
270 {1254, "CP1254"},
271 {1254, "MS-TURK"},
272 {1254, "WINDOWS-1254"},
273
274 {1255, "CP1255"},
275 {1255, "MS-HEBR"},
276 {1255, "WINDOWS-1255"},
277
278 {1256, "CP1256"},
279 {1256, "MS-ARAB"},
280 {1256, "WINDOWS-1256"},
281
282 {1257, "CP1257"},
283 {1257, "WINBALTRIM"},
284 {1257, "WINDOWS-1257"},
285
286 {1258, "CP1258"},
287 {1258, "WINDOWS-1258"},
288
289 {850, "850"},
290 {850, "CP850"},
291 {850, "IBM850"},
292 {850, "CSPC850MULTILINGUAL"},
293
294 /* !IsValidCodePage(862) */
295 {862, "862"},
296 {862, "CP862"},
297 {862, "IBM862"},
298 {862, "CSPC862LATINHEBREW"},
299
300 {866, "866"},
301 {866, "CP866"},
302 {866, "IBM866"},
303 {866, "CSIBM866"},
304
305 /* !IsValidCodePage(154) */
306 {154, "CP154"},
307 {154, "CYRILLIC-ASIAN"},
308 {154, "PT154"},
309 {154, "PTCP154"},
310 {154, "CSPTCP154"},
311
312 /* !IsValidCodePage(1133) */
313 {1133, "CP1133"},
314 {1133, "IBM-CP1133"},
315
316 {874, "CP874"},
317 {874, "WINDOWS-874"},
318
319 /* !IsValidCodePage(51932) */
320 {51932, "CP51932"},
321 {51932, "MS51932"},
322 {51932, "WINDOWS-51932"},
323 {51932, "EUC-JP"},
324
325 {932, "CP932"},
326 {932, "MS932"},
327 {932, "SHIFFT_JIS"},
328 {932, "SHIFFT_JIS-MS"},
329 {932, "SJIS"},
330 {932, "SJIS-MS"},
331 {932, "SJIS-OPEN"},
332 {932, "SJIS-WIN"},
333 {932, "WINDOWS-31J"},
334 {932, "WINDOWS-932"},
335 {932, "CSWINDOWS31J"},
336
337 {50221, "CP50221"},
338 {50221, "ISO-2022-JP"},
339 {50221, "ISO-2022-JP-MS"},
340 {50221, "ISO2022-JP"},
341 {50221, "ISO2022-JP-MS"},
342 {50221, "MS50221"},
343 {50221, "WINDOWS-50221"},
344
345 {936, "CP936"},
346 {936, "GBK"},
347 {936, "MS936"},
348 {936, "WINDOWS-936"},
349
350 {950, "CP950"},
351 {950, "BIG5"},
352 {950, "BIG5HKSCS"},
353 {950, "BIG5-HKSCS"},
354
355 {949, "CP949"},
356 {949, "UHC"},
357 {949, "EUC-KR"},
358
359 {1361, "CP1361"},
360 {1361, "JOHAB"},
361
362 {437, "437"},
363 {437, "CP437"},
364 {437, "IBM437"},
365 {437, "CSPC8CODEPAGE437"},
366
367 {737, "CP737"},
368
369 {775, "CP775"},
370 {775, "IBM775"},
371 {775, "CSPC775BALTIC"},
372
373 {852, "852"},
374 {852, "CP852"},
375 {852, "IBM852"},
376 {852, "CSPCP852"},
377
378 /* !IsValidCodePage(853) */
379 {853, "CP853"},
380
381 {855, "855"},
382 {855, "CP855"},
383 {855, "IBM855"},
384 {855, "CSIBM855"},
385
386 {857, "857"},
387 {857, "CP857"},
388 {857, "IBM857"},
389 {857, "CSIBM857"},
390
391 /* !IsValidCodePage(858) */
392 {858, "CP858"},
393
394 {860, "860"},
395 {860, "CP860"},
396 {860, "IBM860"},
397 {860, "CSIBM860"},
398
399 {861, "861"},
400 {861, "CP-IS"},
401 {861, "CP861"},
402 {861, "IBM861"},
403 {861, "CSIBM861"},
404
405 {863, "863"},
406 {863, "CP863"},
407 {863, "IBM863"},
408 {863, "CSIBM863"},
409
410 {864, "CP864"},
411 {864, "IBM864"},
412 {864, "CSIBM864"},
413
414 {865, "865"},
415 {865, "CP865"},
416 {865, "IBM865"},
417 {865, "CSIBM865"},
418
419 {869, "869"},
420 {869, "CP-GR"},
421 {869, "CP869"},
422 {869, "IBM869"},
423 {869, "CSIBM869"},
424
425 /* !IsValidCodePage(1152) */
426 {1125, "CP1125"},
427
428 /*
429 * Code Page Identifiers
430 * http://msdn2.microsoft.com/en-us/library/ms776446.aspx
431 */
432 {37, "IBM037"}, /* IBM EBCDIC US-Canada */
433 {437, "IBM437"}, /* OEM United States */
434 {500, "IBM500"}, /* IBM EBCDIC International */
435 {708, "ASMO-708"}, /* Arabic (ASMO 708) */
436 /* 709 Arabic (ASMO-449+, BCON V4) */
437 /* 710 Arabic - Transparent Arabic */
438 {720, "DOS-720"}, /* Arabic (Transparent ASMO); Arabic (DOS) */
439 {737, "ibm737"}, /* OEM Greek (formerly 437G); Greek (DOS) */
440 {775, "ibm775"}, /* OEM Baltic; Baltic (DOS) */
441 {850, "ibm850"}, /* OEM Multilingual Latin 1; Western European (DOS) */
442 {852, "ibm852"}, /* OEM Latin 2; Central European (DOS) */
443 {855, "IBM855"}, /* OEM Cyrillic (primarily Russian) */
444 {857, "ibm857"}, /* OEM Turkish; Turkish (DOS) */
445 {858, "IBM00858"}, /* OEM Multilingual Latin 1 + Euro symbol */
446 {860, "IBM860"}, /* OEM Portuguese; Portuguese (DOS) */
447 {861, "ibm861"}, /* OEM Icelandic; Icelandic (DOS) */
448 {862, "DOS-862"}, /* OEM Hebrew; Hebrew (DOS) */
449 {863, "IBM863"}, /* OEM French Canadian; French Canadian (DOS) */
450 {864, "IBM864"}, /* OEM Arabic; Arabic (864) */
451 {865, "IBM865"}, /* OEM Nordic; Nordic (DOS) */
452 {866, "cp866"}, /* OEM Russian; Cyrillic (DOS) */
453 {869, "ibm869"}, /* OEM Modern Greek; Greek, Modern (DOS) */
454 {870, "IBM870"}, /* IBM EBCDIC Multilingual/ROECE (Latin 2); IBM EBCDIC Multilingual Latin 2 */
455 {874, "windows-874"}, /* ANSI/OEM Thai (same as 28605, ISO 8859-15); Thai (Windows) */
456 {875, "cp875"}, /* IBM EBCDIC Greek Modern */
457 {932, "shift_jis"}, /* ANSI/OEM Japanese; Japanese (Shift-JIS) */
458 {932, "shift-jis"}, /* alternative name for it */
459 {936, "gb2312"}, /* ANSI/OEM Simplified Chinese (PRC, Singapore); Chinese Simplified (GB2312) */
460 {949, "ks_c_5601-1987"}, /* ANSI/OEM Korean (Unified Hangul Code) */
461 {950, "big5"}, /* ANSI/OEM Traditional Chinese (Taiwan; Hong Kong SAR, PRC); Chinese Traditional (Big5) */
462 {950, "big5hkscs"}, /* ANSI/OEM Traditional Chinese (Hong Kong SAR); Chinese Traditional (Big5-HKSCS) */
463 {950, "big5-hkscs"}, /* alternative name for it */
464 {1026, "IBM1026"}, /* IBM EBCDIC Turkish (Latin 5) */
465 {1047, "IBM01047"}, /* IBM EBCDIC Latin 1/Open System */
466 {1140, "IBM01140"}, /* IBM EBCDIC US-Canada (037 + Euro symbol); IBM EBCDIC (US-Canada-Euro) */
467 {1141, "IBM01141"}, /* IBM EBCDIC Germany (20273 + Euro symbol); IBM EBCDIC (Germany-Euro) */
468 {1142, "IBM01142"}, /* IBM EBCDIC Denmark-Norway (20277 + Euro symbol); IBM EBCDIC (Denmark-Norway-Euro) */
469 {1143, "IBM01143"}, /* IBM EBCDIC Finland-Sweden (20278 + Euro symbol); IBM EBCDIC (Finland-Sweden-Euro) */
470 {1144, "IBM01144"}, /* IBM EBCDIC Italy (20280 + Euro symbol); IBM EBCDIC (Italy-Euro) */
471 {1145, "IBM01145"}, /* IBM EBCDIC Latin America-Spain (20284 + Euro symbol); IBM EBCDIC (Spain-Euro) */
472 {1146, "IBM01146"}, /* IBM EBCDIC United Kingdom (20285 + Euro symbol); IBM EBCDIC (UK-Euro) */
473 {1147, "IBM01147"}, /* IBM EBCDIC France (20297 + Euro symbol); IBM EBCDIC (France-Euro) */
474 {1148, "IBM01148"}, /* IBM EBCDIC International (500 + Euro symbol); IBM EBCDIC (International-Euro) */
475 {1149, "IBM01149"}, /* IBM EBCDIC Icelandic (20871 + Euro symbol); IBM EBCDIC (Icelandic-Euro) */
476 {1250, "windows-1250"}, /* ANSI Central European; Central European (Windows) */
477 {1251, "windows-1251"}, /* ANSI Cyrillic; Cyrillic (Windows) */
478 {1252, "windows-1252"}, /* ANSI Latin 1; Western European (Windows) */
479 {1253, "windows-1253"}, /* ANSI Greek; Greek (Windows) */
480 {1254, "windows-1254"}, /* ANSI Turkish; Turkish (Windows) */
481 {1255, "windows-1255"}, /* ANSI Hebrew; Hebrew (Windows) */
482 {1256, "windows-1256"}, /* ANSI Arabic; Arabic (Windows) */
483 {1257, "windows-1257"}, /* ANSI Baltic; Baltic (Windows) */
484 {1258, "windows-1258"}, /* ANSI/OEM Vietnamese; Vietnamese (Windows) */
485 {1361, "Johab"}, /* Korean (Johab) */
486 {10000, "macintosh"}, /* MAC Roman; Western European (Mac) */
487 {10001, "x-mac-japanese"}, /* Japanese (Mac) */
488 {10002, "x-mac-chinesetrad"}, /* MAC Traditional Chinese (Big5); Chinese Traditional (Mac) */
489 {10003, "x-mac-korean"}, /* Korean (Mac) */
490 {10004, "x-mac-arabic"}, /* Arabic (Mac) */
491 {10005, "x-mac-hebrew"}, /* Hebrew (Mac) */
492 {10006, "x-mac-greek"}, /* Greek (Mac) */
493 {10007, "x-mac-cyrillic"}, /* Cyrillic (Mac) */
494 {10008, "x-mac-chinesesimp"}, /* MAC Simplified Chinese (GB 2312); Chinese Simplified (Mac) */
495 {10010, "x-mac-romanian"}, /* Romanian (Mac) */
496 {10017, "x-mac-ukrainian"}, /* Ukrainian (Mac) */
497 {10021, "x-mac-thai"}, /* Thai (Mac) */
498 {10029, "x-mac-ce"}, /* MAC Latin 2; Central European (Mac) */
499 {10079, "x-mac-icelandic"}, /* Icelandic (Mac) */
500 {10081, "x-mac-turkish"}, /* Turkish (Mac) */
501 {10082, "x-mac-croatian"}, /* Croatian (Mac) */
502 {20000, "x-Chinese_CNS"}, /* CNS Taiwan; Chinese Traditional (CNS) */
503 {20001, "x-cp20001"}, /* TCA Taiwan */
504 {20002, "x_Chinese-Eten"}, /* Eten Taiwan; Chinese Traditional (Eten) */
505 {20003, "x-cp20003"}, /* IBM5550 Taiwan */
506 {20004, "x-cp20004"}, /* TeleText Taiwan */
507 {20005, "x-cp20005"}, /* Wang Taiwan */
508 {20105, "x-IA5"}, /* IA5 (IRV International Alphabet No. 5, 7-bit); Western European (IA5) */
509 {20106, "x-IA5-German"}, /* IA5 German (7-bit) */
510 {20107, "x-IA5-Swedish"}, /* IA5 Swedish (7-bit) */
511 {20108, "x-IA5-Norwegian"}, /* IA5 Norwegian (7-bit) */
512 {20127, "us-ascii"}, /* US-ASCII (7-bit) */
513 {20261, "x-cp20261"}, /* T.61 */
514 {20269, "x-cp20269"}, /* ISO 6937 Non-Spacing Accent */
515 {20273, "IBM273"}, /* IBM EBCDIC Germany */
516 {20277, "IBM277"}, /* IBM EBCDIC Denmark-Norway */
517 {20278, "IBM278"}, /* IBM EBCDIC Finland-Sweden */
518 {20280, "IBM280"}, /* IBM EBCDIC Italy */
519 {20284, "IBM284"}, /* IBM EBCDIC Latin America-Spain */
520 {20285, "IBM285"}, /* IBM EBCDIC United Kingdom */
521 {20290, "IBM290"}, /* IBM EBCDIC Japanese Katakana Extended */
522 {20297, "IBM297"}, /* IBM EBCDIC France */
523 {20420, "IBM420"}, /* IBM EBCDIC Arabic */
524 {20423, "IBM423"}, /* IBM EBCDIC Greek */
525 {20424, "IBM424"}, /* IBM EBCDIC Hebrew */
526 {20833, "x-EBCDIC-KoreanExtended"}, /* IBM EBCDIC Korean Extended */
527 {20838, "IBM-Thai"}, /* IBM EBCDIC Thai */
528 {20866, "koi8-r"}, /* Russian (KOI8-R); Cyrillic (KOI8-R) */
529 {20871, "IBM871"}, /* IBM EBCDIC Icelandic */
530 {20880, "IBM880"}, /* IBM EBCDIC Cyrillic Russian */
531 {20905, "IBM905"}, /* IBM EBCDIC Turkish */
532 {20924, "IBM00924"}, /* IBM EBCDIC Latin 1/Open System (1047 + Euro symbol) */
533 {20932, "EUC-JP"}, /* Japanese (JIS 0208-1990 and 0121-1990) */
534 {20936, "x-cp20936"}, /* Simplified Chinese (GB2312); Chinese Simplified (GB2312-80) */
535 {20949, "x-cp20949"}, /* Korean Wansung */
536 {21025, "cp1025"}, /* IBM EBCDIC Cyrillic Serbian-Bulgarian */
537 /* 21027 (deprecated) */
538 {21866, "koi8-u"}, /* Ukrainian (KOI8-U); Cyrillic (KOI8-U) */
539 {28591, "iso-8859-1"}, /* ISO 8859-1 Latin 1; Western European (ISO) */
540 {28591, "iso8859-1"}, /* ISO 8859-1 Latin 1; Western European (ISO) */
541 {28591, "iso_8859-1"},
542 {28591, "iso_8859_1"},
543 {28592, "iso-8859-2"}, /* ISO 8859-2 Central European; Central European (ISO) */
544 {28592, "iso8859-2"}, /* ISO 8859-2 Central European; Central European (ISO) */
545 {28592, "iso_8859-2"},
546 {28592, "iso_8859_2"},
547 {28593, "iso-8859-3"}, /* ISO 8859-3 Latin 3 */
548 {28593, "iso8859-3"}, /* ISO 8859-3 Latin 3 */
549 {28593, "iso_8859-3"},
550 {28593, "iso_8859_3"},
551 {28594, "iso-8859-4"}, /* ISO 8859-4 Baltic */
552 {28594, "iso8859-4"}, /* ISO 8859-4 Baltic */
553 {28594, "iso_8859-4"},
554 {28594, "iso_8859_4"},
555 {28595, "iso-8859-5"}, /* ISO 8859-5 Cyrillic */
556 {28595, "iso8859-5"}, /* ISO 8859-5 Cyrillic */
557 {28595, "iso_8859-5"},
558 {28595, "iso_8859_5"},
559 {28596, "iso-8859-6"}, /* ISO 8859-6 Arabic */
560 {28596, "iso8859-6"}, /* ISO 8859-6 Arabic */
561 {28596, "iso_8859-6"},
562 {28596, "iso_8859_6"},
563 {28597, "iso-8859-7"}, /* ISO 8859-7 Greek */
564 {28597, "iso8859-7"}, /* ISO 8859-7 Greek */
565 {28597, "iso_8859-7"},
566 {28597, "iso_8859_7"},
567 {28598, "iso-8859-8"}, /* ISO 8859-8 Hebrew; Hebrew (ISO-Visual) */
568 {28598, "iso8859-8"}, /* ISO 8859-8 Hebrew; Hebrew (ISO-Visual) */
569 {28598, "iso_8859-8"},
570 {28598, "iso_8859_8"},
571 {28599, "iso-8859-9"}, /* ISO 8859-9 Turkish */
572 {28599, "iso8859-9"}, /* ISO 8859-9 Turkish */
573 {28599, "iso_8859-9"},
574 {28599, "iso_8859_9"},
575 {28603, "iso-8859-13"}, /* ISO 8859-13 Estonian */
576 {28603, "iso8859-13"}, /* ISO 8859-13 Estonian */
577 {28603, "iso_8859-13"},
578 {28603, "iso_8859_13"},
579 {28605, "iso-8859-15"}, /* ISO 8859-15 Latin 9 */
580 {28605, "iso8859-15"}, /* ISO 8859-15 Latin 9 */
581 {28605, "iso_8859-15"},
582 {28605, "iso_8859_15"},
583 {29001, "x-Europa"}, /* Europa 3 */
584 {38598, "iso-8859-8-i"}, /* ISO 8859-8 Hebrew; Hebrew (ISO-Logical) */
585 {38598, "iso8859-8-i"}, /* ISO 8859-8 Hebrew; Hebrew (ISO-Logical) */
586 {38598, "iso_8859-8-i"},
587 {38598, "iso_8859_8-i"},
588 {50220, "iso-2022-jp"}, /* ISO 2022 Japanese with no halfwidth Katakana; Japanese (JIS) */
589 {50221, "csISO2022JP"}, /* ISO 2022 Japanese with halfwidth Katakana; Japanese (JIS-Allow 1 byte Kana) */
590 {50222, "iso-2022-jp"}, /* ISO 2022 Japanese JIS X 0201-1989; Japanese (JIS-Allow 1 byte Kana - SO/SI) */
591 {50225, "iso-2022-kr"}, /* ISO 2022 Korean */
592 {50225, "iso2022-kr"}, /* ISO 2022 Korean */
593 {50227, "x-cp50227"}, /* ISO 2022 Simplified Chinese; Chinese Simplified (ISO 2022) */
594 /* 50229 ISO 2022 Traditional Chinese */
595 /* 50930 EBCDIC Japanese (Katakana) Extended */
596 /* 50931 EBCDIC US-Canada and Japanese */
597 /* 50933 EBCDIC Korean Extended and Korean */
598 /* 50935 EBCDIC Simplified Chinese Extended and Simplified Chinese */
599 /* 50936 EBCDIC Simplified Chinese */
600 /* 50937 EBCDIC US-Canada and Traditional Chinese */
601 /* 50939 EBCDIC Japanese (Latin) Extended and Japanese */
602 {51932, "euc-jp"}, /* EUC Japanese */
603 {51936, "EUC-CN"}, /* EUC Simplified Chinese; Chinese Simplified (EUC) */
604 {51949, "euc-kr"}, /* EUC Korean */
605 /* 51950 EUC Traditional Chinese */
606 {52936, "hz-gb-2312"}, /* HZ-GB2312 Simplified Chinese; Chinese Simplified (HZ) */
607 {54936, "GB18030"}, /* Windows XP and later: GB18030 Simplified Chinese (4 byte); Chinese Simplified (GB18030) */
608 {57002, "x-iscii-de"}, /* ISCII Devanagari */
609 {57003, "x-iscii-be"}, /* ISCII Bengali */
610 {57004, "x-iscii-ta"}, /* ISCII Tamil */
611 {57005, "x-iscii-te"}, /* ISCII Telugu */
612 {57006, "x-iscii-as"}, /* ISCII Assamese */
613 {57007, "x-iscii-or"}, /* ISCII Oriya */
614 {57008, "x-iscii-ka"}, /* ISCII Kannada */
615 {57009, "x-iscii-ma"}, /* ISCII Malayalam */
616 {57010, "x-iscii-gu"}, /* ISCII Gujarati */
617 {57011, "x-iscii-pa"}, /* ISCII Punjabi */
618
619 {0, NULL}
620 };
621
622 /*
623 * SJIS SHIFTJIS table CP932 table
624 * ---- --------------------------- --------------------------------
625 * 5C U+00A5 YEN SIGN U+005C REVERSE SOLIDUS
626 * 7E U+203E OVERLINE U+007E TILDE
627 * 815C U+2014 EM DASH U+2015 HORIZONTAL BAR
628 * 815F U+005C REVERSE SOLIDUS U+FF3C FULLWIDTH REVERSE SOLIDUS
629 * 8160 U+301C WAVE DASH U+FF5E FULLWIDTH TILDE
630 * 8161 U+2016 DOUBLE VERTICAL LINE U+2225 PARALLEL TO
631 * 817C U+2212 MINUS SIGN U+FF0D FULLWIDTH HYPHEN-MINUS
632 * 8191 U+00A2 CENT SIGN U+FFE0 FULLWIDTH CENT SIGN
633 * 8192 U+00A3 POUND SIGN U+FFE1 FULLWIDTH POUND SIGN
634 * 81CA U+00AC NOT SIGN U+FFE2 FULLWIDTH NOT SIGN
635 *
636 * EUC-JP and ISO-2022-JP should be compatible with CP932.
637 *
638 * Kernel and MLang have different Unicode mapping table. Make sure
639 * which API is used.
640 */
641 static compat_t cp932_compat[] = {
642 {0x00A5, 0x005C, COMPAT_OUT},
643 {0x203E, 0x007E, COMPAT_OUT},
644 {0x2014, 0x2015, COMPAT_OUT},
645 {0x301C, 0xFF5E, COMPAT_OUT},
646 {0x2016, 0x2225, COMPAT_OUT},
647 {0x2212, 0xFF0D, COMPAT_OUT},
648 {0x00A2, 0xFFE0, COMPAT_OUT},
649 {0x00A3, 0xFFE1, COMPAT_OUT},
650 {0x00AC, 0xFFE2, COMPAT_OUT},
651 {0, 0, 0}
652 };
653
654 static compat_t cp20932_compat[] = {
655 {0x00A5, 0x005C, COMPAT_OUT},
656 {0x203E, 0x007E, COMPAT_OUT},
657 {0x2014, 0x2015, COMPAT_OUT},
658 {0xFF5E, 0x301C, COMPAT_OUT|COMPAT_IN},
659 {0x2225, 0x2016, COMPAT_OUT|COMPAT_IN},
660 {0xFF0D, 0x2212, COMPAT_OUT|COMPAT_IN},
661 {0xFFE0, 0x00A2, COMPAT_OUT|COMPAT_IN},
662 {0xFFE1, 0x00A3, COMPAT_OUT|COMPAT_IN},
663 {0xFFE2, 0x00AC, COMPAT_OUT|COMPAT_IN},
664 {0, 0, 0}
665 };
666
667 static compat_t *cp51932_compat = cp932_compat;
668
669 /* cp20932_compat for kernel. cp932_compat for mlang. */
670 static compat_t *cp5022x_compat = cp932_compat;
671
672 typedef HRESULT (WINAPI *CONVERTINETSTRING)(
673 LPDWORD lpdwMode,
674 DWORD dwSrcEncoding,
675 DWORD dwDstEncoding,
676 LPCSTR lpSrcStr,
677 LPINT lpnSrcSize,
678 LPBYTE lpDstStr,
679 LPINT lpnDstSize
680 );
681 typedef HRESULT (WINAPI *CONVERTINETMULTIBYTETOUNICODE)(
682 LPDWORD lpdwMode,
683 DWORD dwSrcEncoding,
684 LPCSTR lpSrcStr,
685 LPINT lpnMultiCharCount,
686 LPWSTR lpDstStr,
687 LPINT lpnWideCharCount
688 );
689 typedef HRESULT (WINAPI *CONVERTINETUNICODETOMULTIBYTE)(
690 LPDWORD lpdwMode,
691 DWORD dwEncoding,
692 LPCWSTR lpSrcStr,
693 LPINT lpnWideCharCount,
694 LPSTR lpDstStr,
695 LPINT lpnMultiCharCount
696 );
697 typedef HRESULT (WINAPI *ISCONVERTINETSTRINGAVAILABLE)(
698 DWORD dwSrcEncoding,
699 DWORD dwDstEncoding
700 );
701 typedef HRESULT (WINAPI *LCIDTORFC1766A)(
702 LCID Locale,
703 LPSTR pszRfc1766,
704 int nChar
705 );
706 typedef HRESULT (WINAPI *LCIDTORFC1766W)(
707 LCID Locale,
708 LPWSTR pszRfc1766,
709 int nChar
710 );
711 typedef HRESULT (WINAPI *RFC1766TOLCIDA)(
712 LCID *pLocale,
713 LPSTR pszRfc1766
714 );
715 typedef HRESULT (WINAPI *RFC1766TOLCIDW)(
716 LCID *pLocale,
717 LPWSTR pszRfc1766
718 );
719 static CONVERTINETSTRING ConvertINetString;
720 static CONVERTINETMULTIBYTETOUNICODE ConvertINetMultiByteToUnicode;
721 static CONVERTINETUNICODETOMULTIBYTE ConvertINetUnicodeToMultiByte;
722 static ISCONVERTINETSTRINGAVAILABLE IsConvertINetStringAvailable;
723 static LCIDTORFC1766A LcidToRfc1766A;
724 static RFC1766TOLCIDA Rfc1766ToLcidA;
725
726 static int
727 load_mlang(void)
728 {
729 HMODULE h;
730 if (ConvertINetString != NULL)
731 return TRUE;
732 h = LoadLibrary(TEXT("mlang.dll"));
733 if (!h)
734 return FALSE;
735 ConvertINetString = (CONVERTINETSTRING)GetProcAddressA(h, "ConvertINetString");
736 ConvertINetMultiByteToUnicode = (CONVERTINETMULTIBYTETOUNICODE)GetProcAddressA(h, "ConvertINetMultiByteToUnicode");
737 ConvertINetUnicodeToMultiByte = (CONVERTINETUNICODETOMULTIBYTE)GetProcAddressA(h, "ConvertINetUnicodeToMultiByte");
738 IsConvertINetStringAvailable = (ISCONVERTINETSTRINGAVAILABLE)GetProcAddressA(h, "IsConvertINetStringAvailable");
739 LcidToRfc1766A = (LCIDTORFC1766A)GetProcAddressA(h, "LcidToRfc1766A");
740 Rfc1766ToLcidA = (RFC1766TOLCIDA)GetProcAddressA(h, "Rfc1766ToLcidA");
741 return TRUE;
742 }
743
744 iconv_t
745 iconv_open(const char *tocode, const char *fromcode)
746 {
747 rec_iconv_t *cd;
748
749 cd = (rec_iconv_t *)calloc(1, sizeof(rec_iconv_t));
750 if (cd == NULL)
751 return (iconv_t)(-1);
752
753 #if defined(USE_LIBICONV_DLL)
754 errno = 0;
755 if (libiconv_iconv_open(cd, tocode, fromcode))
756 return (iconv_t)cd;
757 #endif
758
759 /* reset the errno to prevent reporting wrong error code.
760 * 0 for unsorted error. */
761 errno = 0;
762 if (win_iconv_open(cd, tocode, fromcode))
763 return (iconv_t)cd;
764
765 free(cd);
766
767 return (iconv_t)(-1);
768 }
769
770 int
771 iconv_close(iconv_t _cd)
772 {
773 rec_iconv_t *cd = (rec_iconv_t *)_cd;
774 int r = cd->iconv_close(cd->cd);
775 int e = *(cd->_errno());
776 #if defined(USE_LIBICONV_DLL)
777 if (cd->hlibiconv != NULL)
778 FreeLibrary(cd->hlibiconv);
779 #endif
780 free(cd);
781 errno = e;
782 return r;
783 }
784
785 size_t
786 iconv(iconv_t _cd, /* const */ char **inbuf, size_t *inbytesleft, char **outbuf, size_t *outbytesleft)
787 {
788 rec_iconv_t *cd = (rec_iconv_t *)_cd;
789 size_t r = cd->iconv(cd->cd, inbuf, inbytesleft, outbuf, outbytesleft);
790 errno = *(cd->_errno());
791 return r;
792 }
793
794 static int
795 win_iconv_open(rec_iconv_t *cd, const char *tocode, const char *fromcode)
796 {
797 if (!make_csconv(fromcode, &cd->from) || !make_csconv(tocode, &cd->to))
798 return FALSE;
799 cd->iconv_close = win_iconv_close;
800 cd->iconv = win_iconv;
801 cd->_errno = _errno;
802 cd->cd = (iconv_t)cd;
803 return TRUE;
804 }
805
806 static int
807 win_iconv_close(iconv_t cd UNUSED)
808 {
809 return 0;
810 }
811
812 static size_t
813 win_iconv(iconv_t _cd, /* const */ char **inbuf, size_t *inbytesleft, char **outbuf, size_t *outbytesleft)
814 {
815 rec_iconv_t *cd = (rec_iconv_t *)_cd;
816 ushort wbuf[MB_CHAR_MAX]; /* enough room for one character */
817 int insize;
818 int outsize;
819 int wsize;
820 DWORD frommode;
821 DWORD tomode;
822 uint wc;
823 compat_t *cp;
824 int i;
825
826 if (inbuf == NULL || *inbuf == NULL)
827 {
828 if (outbuf != NULL && *outbuf != NULL && cd->to.flush != NULL)
829 {
830 tomode = cd->to.mode;
831 outsize = cd->to.flush(&cd->to, (uchar *)*outbuf, *outbytesleft);
832 if (outsize == -1)
833 {
834 if ((cd->to.flags & FLAG_IGNORE) && errno != E2BIG)
835 {
836 outsize = 0;
837 }
838 else
839 {
840 cd->to.mode = tomode;
841 return (size_t)(-1);
842 }
843 }
844 *outbuf += outsize;
845 *outbytesleft -= outsize;
846 }
847 cd->from.mode = 0;
848 cd->to.mode = 0;
849 return 0;
850 }
851
852 while (*inbytesleft != 0)
853 {
854 frommode = cd->from.mode;
855 tomode = cd->to.mode;
856 wsize = MB_CHAR_MAX;
857
858 insize = cd->from.mbtowc(&cd->from, (const uchar *)*inbuf, *inbytesleft, wbuf, &wsize);
859 if (insize == -1)
860 {
861 if (cd->to.flags & FLAG_IGNORE)
862 {
863 cd->from.mode = frommode;
864 insize = 1;
865 wsize = 0;
866 }
867 else
868 {
869 cd->from.mode = frommode;
870 return (size_t)(-1);
871 }
872 }
873
874 if (wsize == 0)
875 {
876 *inbuf += insize;
877 *inbytesleft -= insize;
878 continue;
879 }
880
881 if (cd->from.compat != NULL)
882 {
883 wc = utf16_to_ucs4(wbuf);
884 cp = cd->from.compat;
885 for (i = 0; cp[i].in != 0; ++i)
886 {
887 if ((cp[i].flag & COMPAT_IN) && cp[i].out == wc)
888 {
889 ucs4_to_utf16(cp[i].in, wbuf, &wsize);
890 break;
891 }
892 }
893 }
894
895 if (cd->to.compat != NULL)
896 {
897 wc = utf16_to_ucs4(wbuf);
898 cp = cd->to.compat;
899 for (i = 0; cp[i].in != 0; ++i)
900 {
901 if ((cp[i].flag & COMPAT_OUT) && cp[i].in == wc)
902 {
903 ucs4_to_utf16(cp[i].out, wbuf, &wsize);
904 break;
905 }
906 }
907 }
908
909 outsize = cd->to.wctomb(&cd->to, wbuf, wsize, (uchar *)*outbuf, *outbytesleft);
910 if (outsize == -1)
911 {
912 if ((cd->to.flags & FLAG_IGNORE) && errno != E2BIG)
913 {
914 cd->to.mode = tomode;
915 outsize = 0;
916 }
917 else
918 {
919 cd->from.mode = frommode;
920 cd->to.mode = tomode;
921 return (size_t)(-1);
922 }
923 }
924
925 *inbuf += insize;
926 *outbuf += outsize;
927 *inbytesleft -= insize;
928 *outbytesleft -= outsize;
929 }
930
931 return 0;
932 }
933
934 static int
935 make_csconv(const char *_name, csconv_t *cv)
936 {
937 CPINFO cpinfo;
938 int use_compat = TRUE;
939 int flag = 0;
940 char *name;
941 char *p;
942
943 name = xstrndup(_name, strlen(_name));
944 if (name == NULL)
945 return FALSE;
946
947 /* check for option "enc_name//opt1//opt2" */
948 while ((p = strrstr(name, "//")) != NULL)
949 {
950 if (_stricmp(p + 2, "nocompat") == 0)
951 use_compat = FALSE;
952 else if (_stricmp(p + 2, "translit") == 0)
953 flag |= FLAG_TRANSLIT;
954 else if (_stricmp(p + 2, "ignore") == 0)
955 flag |= FLAG_IGNORE;
956 *p = 0;
957 }
958
959 cv->mode = 0;
960 cv->flags = flag;
961 cv->mblen = NULL;
962 cv->flush = NULL;
963 cv->compat = NULL;
964 cv->codepage = name_to_codepage(name);
965 if (cv->codepage == 1200 || cv->codepage == 1201)
966 {
967 cv->mbtowc = utf16_mbtowc;
968 cv->wctomb = utf16_wctomb;
969 if (_stricmp(name, "UTF-16") == 0 || _stricmp(name, "UTF16") == 0 ||
970 _stricmp(name, "UCS-2") == 0 || _stricmp(name, "UCS2") == 0)
971 cv->flags |= FLAG_USE_BOM;
972 }
973 else if (cv->codepage == 12000 || cv->codepage == 12001)
974 {
975 cv->mbtowc = utf32_mbtowc;
976 cv->wctomb = utf32_wctomb;
977 if (_stricmp(name, "UTF-32") == 0 || _stricmp(name, "UTF32") == 0 ||
978 _stricmp(name, "UCS-4") == 0 || _stricmp(name, "UCS4") == 0)
979 cv->flags |= FLAG_USE_BOM;
980 }
981 else if (cv->codepage == 65001)
982 {
983 cv->mbtowc = kernel_mbtowc;
984 cv->wctomb = kernel_wctomb;
985 cv->mblen = utf8_mblen;
986 }
987 else if ((cv->codepage == 50220 || cv->codepage == 50221 || cv->codepage == 50222) && load_mlang())
988 {
989 cv->mbtowc = iso2022jp_mbtowc;
990 cv->wctomb = iso2022jp_wctomb;
991 cv->flush = iso2022jp_flush;
992 }
993 else if (cv->codepage == 51932 && load_mlang())
994 {
995 cv->mbtowc = mlang_mbtowc;
996 cv->wctomb = mlang_wctomb;
997 cv->mblen = eucjp_mblen;
998 }
999 else if (IsValidCodePage(cv->codepage)
1000 && GetCPInfo(cv->codepage, &cpinfo) != 0)
1001 {
1002 cv->mbtowc = kernel_mbtowc;
1003 cv->wctomb = kernel_wctomb;
1004 if (cpinfo.MaxCharSize == 1)
1005 cv->mblen = sbcs_mblen;
1006 else if (cpinfo.MaxCharSize == 2)
1007 cv->mblen = dbcs_mblen;
1008 else
1009 cv->mblen = mbcs_mblen;
1010 }
1011 else
1012 {
1013 /* not supported */
1014 free(name);
1015 errno = EINVAL;
1016 return FALSE;
1017 }
1018
1019 if (use_compat)
1020 {
1021 switch (cv->codepage)
1022 {
1023 case 932: cv->compat = cp932_compat; break;
1024 case 20932: cv->compat = cp20932_compat; break;
1025 case 51932: cv->compat = cp51932_compat; break;
1026 case 50220: case 50221: case 50222: cv->compat = cp5022x_compat; break;
1027 }
1028 }
1029
1030 free(name);
1031
1032 return TRUE;
1033 }
1034
1035 static int
1036 name_to_codepage(const char *name)
1037 {
1038 int i;
1039
1040 if (*name == '\0' ||
1041 strcmp(name, "char") == 0)
1042 return GetACP();
1043 else if (strcmp(name, "wchar_t") == 0)
1044 return 1200;
1045 else if (_strnicmp(name, "cp", 2) == 0)
1046 return atoi(name + 2); /* CP123 */
1047 else if ('0' <= name[0] && name[0] <= '9')
1048 return atoi(name); /* 123 */
1049 else if (_strnicmp(name, "xx", 2) == 0)
1050 return atoi(name + 2); /* XX123 for debug */
1051
1052 for (i = 0; codepage_alias[i].name != NULL; ++i)
1053 if (_stricmp(name, codepage_alias[i].name) == 0)
1054 return codepage_alias[i].codepage;
1055 return -1;
1056 }
1057
1058 /*
1059 * http://www.faqs.org/rfcs/rfc2781.html
1060 */
1061 static uint
1062 utf16_to_ucs4(const ushort *wbuf)
1063 {
1064 uint wc = wbuf[0];
1065 if (0xD800 <= wbuf[0] && wbuf[0] <= 0xDBFF)
1066 wc = ((wbuf[0] & 0x3FF) << 10) + (wbuf[1] & 0x3FF) + 0x10000;
1067 return wc;
1068 }
1069
1070 static void
1071 ucs4_to_utf16(uint wc, ushort *wbuf, int *wbufsize)
1072 {
1073 if (wc < 0x10000)
1074 {
1075 wbuf[0] = wc;
1076 *wbufsize = 1;
1077 }
1078 else
1079 {
1080 wc -= 0x10000;
1081 wbuf[0] = 0xD800 | ((wc >> 10) & 0x3FF);
1082 wbuf[1] = 0xDC00 | (wc & 0x3FF);
1083 *wbufsize = 2;
1084 }
1085 }
1086
1087 /*
1088 * Check if codepage is one of those for which the dwFlags parameter
1089 * to MultiByteToWideChar() must be zero. Return zero or
1090 * MB_ERR_INVALID_CHARS. The docs in Platform SDK for for Windows
1091 * Server 2003 R2 claims that also codepage 65001 is one of these, but
1092 * that doesn't seem to be the case. The MSDN docs for MSVS2008 leave
1093 * out 65001 (UTF-8), and that indeed seems to be the case on XP, it
1094 * works fine to pass MB_ERR_INVALID_CHARS in dwFlags when converting
1095 * from UTF-8.
1096 */
1097 static int
1098 mbtowc_flags(int codepage)
1099 {
1100 return (codepage == 50220 || codepage == 50221 ||
1101 codepage == 50222 || codepage == 50225 ||
1102 codepage == 50227 || codepage == 50229 ||
1103 codepage == 52936 || codepage == 54936 ||
1104 (codepage >= 57002 && codepage <= 57011) ||
1105 codepage == 65000 || codepage == 42) ? 0 : MB_ERR_INVALID_CHARS;
1106 }
1107
1108 /*
1109 * Check if codepage is one those for which the lpUsedDefaultChar
1110 * parameter to WideCharToMultiByte() must be NULL. The docs in
1111 * Platform SDK for Windows Server 2003 R2 claims that this is the
1112 * list below, while the MSDN docs for MSVS2008 claim that it is only
1113 * for 65000 (UTF-7) and 65001 (UTF-8). This time the earlier Platform
1114 * SDK seems to be correct, at least for XP.
1115 */
1116 static int
1117 must_use_null_useddefaultchar(int codepage)
1118 {
1119 return (codepage == 65000 || codepage == 65001 ||
1120 codepage == 50220 || codepage == 50221 ||
1121 codepage == 50222 || codepage == 50225 ||
1122 codepage == 50227 || codepage == 50229 ||
1123 codepage == 52936 || codepage == 54936 ||
1124 (codepage >= 57002 && codepage <= 57011) ||
1125 codepage == 42);
1126 }
1127
1128 static char *
1129 strrstr(const char *str, const char *token)
1130 {
1131 int len = strlen(token);
1132 const char *p = str + strlen(str);
1133
1134 while (str <= --p)
1135 if (p[0] == token[0] && strncmp(p, token, len) == 0)
1136 return (char *)p;
1137 return NULL;
1138 }
1139
1140 static char *
1141 xstrndup(const char *s, size_t n)
1142 {
1143 char *p;
1144
1145 p = (char *)malloc(n + 1);
1146 if (p == NULL)
1147 return NULL;
1148 memcpy(p, s, n);
1149 p[n] = '\0';
1150 return p;
1151 }
1152
1153 static int
1154 seterror(int err)
1155 {
1156 errno = err;
1157 return -1;
1158 }
1159
1160 #if defined(USE_LIBICONV_DLL)
1161 static int
1162 libiconv_iconv_open(rec_iconv_t *cd, const char *tocode, const char *fromcode)
1163 {
1164 HMODULE hlibiconv = NULL;
1165 char *dllname;
1166 const char *p;
1167 const char *e;
1168 f_iconv_open _iconv_open;
1169
1170 /*
1171 * always try to load dll, so that we can switch dll in runtime.
1172 */
1173
1174 /* XXX: getenv() can't get variable set by SetEnvironmentVariable() */
1175 p = getenv("WINICONV_LIBICONV_DLL");
1176 if (p == NULL)
1177 p = DEFAULT_LIBICONV_DLL;
1178 /* parse comma separated value */
1179 for ( ; *p != 0; p = (*e == ',') ? e + 1 : e)
1180 {
1181 e = strchr(p, ',');
1182 if (p == e)
1183 continue;
1184 else if (e == NULL)
1185 e = p + strlen(p);
1186 dllname = xstrndup(p, e - p);
1187 if (dllname == NULL)
1188 return FALSE;
1189 hlibiconv = LoadLibraryA(dllname);
1190 free(dllname);
1191 if (hlibiconv != NULL)
1192 {
1193 if (hlibiconv == hwiniconv)
1194 {
1195 FreeLibrary(hlibiconv);
1196 hlibiconv = NULL;
1197 continue;
1198 }
1199 break;
1200 }
1201 }
1202
1203 if (hlibiconv == NULL)
1204 goto failed;
1205
1206 _iconv_open = (f_iconv_open)GetProcAddressA(hlibiconv, "libiconv_open");
1207 if (_iconv_open == NULL)
1208 _iconv_open = (f_iconv_open)GetProcAddressA(hlibiconv, "iconv_open");
1209 cd->iconv_close = (f_iconv_close)GetProcAddressA(hlibiconv, "libiconv_close");
1210 if (cd->iconv_close == NULL)
1211 cd->iconv_close = (f_iconv_close)GetProcAddressA(hlibiconv, "iconv_close");
1212 cd->iconv = (f_iconv)GetProcAddressA(hlibiconv, "libiconv");
1213 if (cd->iconv == NULL)
1214 cd->iconv = (f_iconv)GetProcAddressA(hlibiconv, "iconv");
1215 cd->_errno = (f_errno)find_imported_function(hlibiconv, "_errno");
1216 if (_iconv_open == NULL || cd->iconv_close == NULL
1217 || cd->iconv == NULL || cd->_errno == NULL)
1218 goto failed;
1219
1220 cd->cd = _iconv_open(tocode, fromcode);
1221 if (cd->cd == (iconv_t)(-1))
1222 goto failed;
1223
1224 cd->hlibiconv = hlibiconv;
1225 return TRUE;
1226
1227 failed:
1228 if (hlibiconv != NULL)
1229 FreeLibrary(hlibiconv);
1230 return FALSE;
1231 }
1232
1233 /*
1234 * Reference:
1235 * http://forums.belution.com/ja/vc/000/234/78s.shtml
1236 * http://nienie.com/~masapico/api_ImageDirectoryEntryToData.html
1237 *
1238 * The formal way is
1239 * imagehlp.h or dbghelp.h
1240 * imagehlp.lib or dbghelp.lib
1241 * ImageDirectoryEntryToData()
1242 */
1243 #define TO_DOS_HEADER(base) ((PIMAGE_DOS_HEADER)(base))
1244 #define TO_NT_HEADERS(base) ((PIMAGE_NT_HEADERS)((LPBYTE)(base) + TO_DOS_HEADER(base)->e_lfanew))
1245 static PVOID
1246 MyImageDirectoryEntryToData(LPVOID Base, BOOLEAN MappedAsImage, USHORT DirectoryEntry, PULONG Size)
1247 {
1248 /* TODO: MappedAsImage? */
1249 PIMAGE_DATA_DIRECTORY p;
1250 p = TO_NT_HEADERS(Base)->OptionalHeader.DataDirectory + DirectoryEntry;
1251 if (p->VirtualAddress == 0) {
1252 *Size = 0;
1253 return NULL;
1254 }
1255 *Size = p->Size;
1256 return (PVOID)((LPBYTE)Base + p->VirtualAddress);
1257 }
1258
1259 static FARPROC
1260 find_imported_function(HMODULE hModule, const char *funcname)
1261 {
1262 DWORD_PTR Base;
1263 ULONG Size;
1264 PIMAGE_IMPORT_DESCRIPTOR Imp;
1265 PIMAGE_THUNK_DATA Address; /* Import Address Table */
1266 PIMAGE_THUNK_DATA Name; /* Import Name Table */
1267 PIMAGE_IMPORT_BY_NAME ImpName;
1268
1269 Base = (DWORD_PTR)hModule;
1270 Imp = (PIMAGE_IMPORT_DESCRIPTOR)MyImageDirectoryEntryToData(
1271 (LPVOID)Base,
1272 TRUE,
1273 IMAGE_DIRECTORY_ENTRY_IMPORT,
1274 &Size);
1275 if (Imp == NULL)
1276 return NULL;
1277 for ( ; Imp->OriginalFirstThunk != 0; ++Imp)
1278 {
1279 Address = (PIMAGE_THUNK_DATA)(Base + Imp->FirstThunk);
1280 Name = (PIMAGE_THUNK_DATA)(Base + Imp->OriginalFirstThunk);
1281 for ( ; Name->u1.Ordinal != 0; ++Name, ++Address)
1282 {
1283 if (!IMAGE_SNAP_BY_ORDINAL(Name->u1.Ordinal))
1284 {
1285 ImpName = (PIMAGE_IMPORT_BY_NAME)
1286 (Base + (DWORD_PTR)Name->u1.AddressOfData);
1287 if (strcmp((char *)ImpName->Name, funcname) == 0)
1288 return (FARPROC)Address->u1.Function;
1289 }
1290 }
1291 }
1292 return NULL;
1293 }
1294 #endif
1295
1296 static int
1297 sbcs_mblen(csconv_t *cv UNUSED, const uchar *buf UNUSED, int bufsize UNUSED)
1298 {
1299 return 1;
1300 }
1301
1302 static int
1303 dbcs_mblen(csconv_t *cv, const uchar *buf, int bufsize)
1304 {
1305 int len = IsDBCSLeadByteEx(cv->codepage, buf[0]) ? 2 : 1;
1306 if (bufsize < len)
1307 return seterror(EINVAL);
1308 return len;
1309 }
1310
1311 static int
1312 mbcs_mblen(csconv_t *cv, const uchar *buf, int bufsize)
1313 {
1314 int len = 0;
1315
1316 if (cv->codepage == 54936) {
1317 if (buf[0] <= 0x7F) len = 1;
1318 else if (buf[0] >= 0x81 && buf[0] <= 0xFE &&
1319 bufsize >= 2 &&
1320 ((buf[1] >= 0x40 && buf[1] <= 0x7E) ||
1321 (buf[1] >= 0x80 && buf[1] <= 0xFE))) len = 2;
1322 else if (buf[0] >= 0x81 && buf[0] <= 0xFE &&
1323 bufsize >= 4 &&
1324 buf[1] >= 0x30 && buf[1] <= 0x39) len = 4;
1325 else
1326 return seterror(EINVAL);
1327 return len;
1328 }
1329 else
1330 return seterror(EINVAL);
1331 }
1332
1333 static int
1334 utf8_mblen(csconv_t *cv UNUSED, const uchar *buf, int bufsize)
1335 {
1336 int len = 0;
1337
1338 if (buf[0] < 0x80) len = 1;
1339 else if ((buf[0] & 0xE0) == 0xC0) len = 2;
1340 else if ((buf[0] & 0xF0) == 0xE0) len = 3;
1341 else if ((buf[0] & 0xF8) == 0xF0) len = 4;
1342 else if ((buf[0] & 0xFC) == 0xF8) len = 5;
1343 else if ((buf[0] & 0xFE) == 0xFC) len = 6;
1344
1345 if (len == 0)
1346 return seterror(EILSEQ);
1347 else if (bufsize < len)
1348 return seterror(EINVAL);
1349 return len;
1350 }
1351
1352 static int
1353 eucjp_mblen(csconv_t *cv UNUSED, const uchar *buf, int bufsize)
1354 {
1355 if (buf[0] < 0x80) /* ASCII */
1356 return 1;
1357 else if (buf[0] == 0x8E) /* JIS X 0201 */
1358 {
1359 if (bufsize < 2)
1360 return seterror(EINVAL);
1361 else if (!(0xA1 <= buf[1] && buf[1] <= 0xDF))
1362 return seterror(EILSEQ);
1363 return 2;
1364 }
1365 else if (buf[0] == 0x8F) /* JIS X 0212 */
1366 {
1367 if (bufsize < 3)
1368 return seterror(EINVAL);
1369 else if (!(0xA1 <= buf[1] && buf[1] <= 0xFE)
1370 || !(0xA1 <= buf[2] && buf[2] <= 0xFE))
1371 return seterror(EILSEQ);
1372 return 3;
1373 }
1374 else /* JIS X 0208 */
1375 {
1376 if (bufsize < 2)
1377 return seterror(EINVAL);
1378 else if (!(0xA1 <= buf[0] && buf[0] <= 0xFE)
1379 || !(0xA1 <= buf[1] && buf[1] <= 0xFE))
1380 return seterror(EILSEQ);
1381 return 2;
1382 }
1383 }
1384
1385 static int
1386 kernel_mbtowc(csconv_t *cv, const uchar *buf, int bufsize, ushort *wbuf, int *wbufsize)
1387 {
1388 int len;
1389
1390 len = cv->mblen(cv, buf, bufsize);
1391 if (len == -1)
1392 return -1;
1393 /* If converting from ASCII, reject 8bit
1394 * chars. MultiByteToWideChar() doesn't. Note that for ASCII we
1395 * know that the mblen function is sbcs_mblen() so len is 1.
1396 */
1397 if (cv->codepage == 20127 && buf[0] >= 0x80)
1398 return seterror(EILSEQ);
1399 *wbufsize = MultiByteToWideChar(cv->codepage, mbtowc_flags (cv->codepage),
1400 (const char *)buf, len, (wchar_t *)wbuf, *wbufsize);
1401 if (*wbufsize == 0)
1402 return seterror(EILSEQ);
1403 return len;
1404 }
1405
1406 static int
1407 kernel_wctomb(csconv_t *cv, ushort *wbuf, int wbufsize, uchar *buf, int bufsize)
1408 {
1409 BOOL usedDefaultChar = 0;
1410 BOOL *p = NULL;
1411 int flags = 0;
1412 int len;
1413
1414 if (bufsize == 0)
1415 return seterror(E2BIG);
1416 if (!must_use_null_useddefaultchar(cv->codepage))
1417 {
1418 p = &usedDefaultChar;
1419 #ifdef WC_NO_BEST_FIT_CHARS
1420 if (!(cv->flags & FLAG_TRANSLIT))
1421 flags |= WC_NO_BEST_FIT_CHARS;
1422 #endif
1423 }
1424 len = WideCharToMultiByte(cv->codepage, flags,
1425 (const wchar_t *)wbuf, wbufsize, (char *)buf, bufsize, NULL, p);
1426 if (len == 0)
1427 {
1428 if (GetLastError() == ERROR_INSUFFICIENT_BUFFER)
1429 return seterror(E2BIG);
1430 return seterror(EILSEQ);
1431 }
1432 else if (usedDefaultChar && !(cv->flags & FLAG_TRANSLIT))
1433 return seterror(EILSEQ);
1434 else if (cv->mblen(cv, buf, len) != len) /* validate result */
1435 return seterror(EILSEQ);
1436 return len;
1437 }
1438
1439 /*
1440 * It seems that the mode (cv->mode) is fixnum.
1441 * For example, when converting iso-2022-jp(cp50221) to unicode:
1442 * in ascii sequence: mode=0xC42C0000
1443 * in jisx0208 sequence: mode=0xC42C0001
1444 * "C42C" is same for each convert session.
1445 * It should be: ((codepage-1)<<16)|state
1446 */
1447 static int
1448 mlang_mbtowc(csconv_t *cv, const uchar *buf, int bufsize, ushort *wbuf, int *wbufsize)
1449 {
1450 int len;
1451 int insize;
1452 HRESULT hr;
1453
1454 len = cv->mblen(cv, buf, bufsize);
1455 if (len == -1)
1456 return -1;
1457 insize = len;
1458 hr = ConvertINetMultiByteToUnicode(&cv->mode, cv->codepage,
1459 (const char *)buf, &insize, (wchar_t *)wbuf, wbufsize);
1460 if (hr != S_OK || insize != len)
1461 return seterror(EILSEQ);
1462 return len;
1463 }
1464
1465 static int
1466 mlang_wctomb(csconv_t *cv, ushort *wbuf, int wbufsize, uchar *buf, int bufsize)
1467 {
1468 char tmpbuf[MB_CHAR_MAX]; /* enough room for one character */
1469 int tmpsize = MB_CHAR_MAX;
1470 int insize = wbufsize;
1471 HRESULT hr;
1472
1473 hr = ConvertINetUnicodeToMultiByte(&cv->mode, cv->codepage,
1474 (const wchar_t *)wbuf, &wbufsize, tmpbuf, &tmpsize);
1475 if (hr != S_OK || insize != wbufsize)
1476 return seterror(EILSEQ);
1477 else if (bufsize < tmpsize)
1478 return seterror(E2BIG);
1479 else if (cv->mblen(cv, (uchar *)tmpbuf, tmpsize) != tmpsize)
1480 return seterror(EILSEQ);
1481 memcpy(buf, tmpbuf, tmpsize);
1482 return tmpsize;
1483 }
1484
1485 static int
1486 utf16_mbtowc(csconv_t *cv, const uchar *buf, int bufsize, ushort *wbuf, int *wbufsize)
1487 {
1488 int codepage = cv->codepage;
1489
1490 /* swap endian: 1200 <-> 1201 */
1491 if (cv->mode & UNICODE_MODE_SWAPPED)
1492 codepage ^= 1;
1493
1494 if (bufsize < 2)
1495 return seterror(EINVAL);
1496 if (codepage == 1200) /* little endian */
1497 wbuf[0] = (buf[1] << 8) | buf[0];
1498 else if (codepage == 1201) /* big endian */
1499 wbuf[0] = (buf[0] << 8) | buf[1];
1500
1501 if ((cv->flags & FLAG_USE_BOM) && !(cv->mode & UNICODE_MODE_BOM_DONE))
1502 {
1503 cv->mode |= UNICODE_MODE_BOM_DONE;
1504 if (wbuf[0] == 0xFFFE)
1505 {
1506 cv->mode |= UNICODE_MODE_SWAPPED;
1507 *wbufsize = 0;
1508 return 2;
1509 }
1510 else if (wbuf[0] == 0xFEFF)
1511 {
1512 *wbufsize = 0;
1513 return 2;
1514 }
1515 }
1516
1517 if (0xDC00 <= wbuf[0] && wbuf[0] <= 0xDFFF)
1518 return seterror(EILSEQ);
1519 if (0xD800 <= wbuf[0] && wbuf[0] <= 0xDBFF)
1520 {
1521 if (bufsize < 4)
1522 return seterror(EINVAL);
1523 if (codepage == 1200) /* little endian */
1524 wbuf[1] = (buf[3] << 8) | buf[2];
1525 else if (codepage == 1201) /* big endian */
1526 wbuf[1] = (buf[2] << 8) | buf[3];
1527 if (!(0xDC00 <= wbuf[1] && wbuf[1] <= 0xDFFF))
1528 return seterror(EILSEQ);
1529 *wbufsize = 2;
1530 return 4;
1531 }
1532 *wbufsize = 1;
1533 return 2;
1534 }
1535
1536 static int
1537 utf16_wctomb(csconv_t *cv, ushort *wbuf, int wbufsize, uchar *buf, int bufsize)
1538 {
1539 if ((cv->flags & FLAG_USE_BOM) && !(cv->mode & UNICODE_MODE_BOM_DONE))
1540 {
1541 int r;
1542
1543 cv->mode |= UNICODE_MODE_BOM_DONE;
1544 if (bufsize < 2)
1545 return seterror(E2BIG);
1546 if (cv->codepage == 1200) /* little endian */
1547 memcpy(buf, "\xFF\xFE", 2);
1548 else if (cv->codepage == 1201) /* big endian */
1549 memcpy(buf, "\xFE\xFF", 2);
1550
1551 r = utf16_wctomb(cv, wbuf, wbufsize, buf + 2, bufsize - 2);
1552 if (r == -1)
1553 return -1;
1554 return r + 2;
1555 }
1556
1557 if (bufsize < 2)
1558 return seterror(E2BIG);
1559 if (cv->codepage == 1200) /* little endian */
1560 {
1561 buf[0] = (wbuf[0] & 0x00FF);
1562 buf[1] = (wbuf[0] & 0xFF00) >> 8;
1563 }
1564 else if (cv->codepage == 1201) /* big endian */
1565 {
1566 buf[0] = (wbuf[0] & 0xFF00) >> 8;
1567 buf[1] = (wbuf[0] & 0x00FF);
1568 }
1569 if (0xD800 <= wbuf[0] && wbuf[0] <= 0xDBFF)
1570 {
1571 if (bufsize < 4)
1572 return seterror(E2BIG);
1573 if (cv->codepage == 1200) /* little endian */
1574 {
1575 buf[2] = (wbuf[1] & 0x00FF);
1576 buf[3] = (wbuf[1] & 0xFF00) >> 8;
1577 }
1578 else if (cv->codepage == 1201) /* big endian */
1579 {
1580 buf[2] = (wbuf[1] & 0xFF00) >> 8;
1581 buf[3] = (wbuf[1] & 0x00FF);
1582 }
1583 return 4;
1584 }
1585 return 2;
1586 }
1587
1588 static int
1589 utf32_mbtowc(csconv_t *cv, const uchar *buf, int bufsize, ushort *wbuf, int *wbufsize)
1590 {
1591 int codepage = cv->codepage;
1592 uint wc = 0xD800;
1593
1594 /* swap endian: 12000 <-> 12001 */
1595 if (cv->mode & UNICODE_MODE_SWAPPED)
1596 codepage ^= 1;
1597
1598 if (bufsize < 4)
1599 return seterror(EINVAL);
1600 if (codepage == 12000) /* little endian */
1601 wc = (buf[3] << 24) | (buf[2] << 16) | (buf[1] << 8) | buf[0];
1602 else if (codepage == 12001) /* big endian */
1603 wc = (buf[0] << 24) | (buf[1] << 16) | (buf[2] << 8) | buf[3];
1604
1605 if ((cv->flags & FLAG_USE_BOM) && !(cv->mode & UNICODE_MODE_BOM_DONE))
1606 {
1607 cv->mode |= UNICODE_MODE_BOM_DONE;
1608 if (wc == 0xFFFE0000)
1609 {
1610 cv->mode |= UNICODE_MODE_SWAPPED;
1611 *wbufsize = 0;
1612 return 4;
1613 }
1614 else if (wc == 0x0000FEFF)
1615 {
1616 *wbufsize = 0;
1617 return 4;
1618 }
1619 }
1620
1621 if ((0xD800 <= wc && wc <= 0xDFFF) || 0x10FFFF < wc)
1622 return seterror(EILSEQ);
1623 ucs4_to_utf16(wc, wbuf, wbufsize);
1624 return 4;
1625 }
1626
1627 static int
1628 utf32_wctomb(csconv_t *cv, ushort *wbuf, int wbufsize, uchar *buf, int bufsize)
1629 {
1630 uint wc;
1631
1632 if ((cv->flags & FLAG_USE_BOM) && !(cv->mode & UNICODE_MODE_BOM_DONE))
1633 {
1634 int r;
1635
1636 cv->mode |= UNICODE_MODE_BOM_DONE;
1637 if (bufsize < 4)
1638 return seterror(E2BIG);
1639 if (cv->codepage == 12000) /* little endian */
1640 memcpy(buf, "\xFF\xFE\x00\x00", 4);
1641 else if (cv->codepage == 12001) /* big endian */
1642 memcpy(buf, "\x00\x00\xFE\xFF", 4);
1643
1644 r = utf32_wctomb(cv, wbuf, wbufsize, buf + 4, bufsize - 4);
1645 if (r == -1)
1646 return -1;
1647 return r + 4;
1648 }
1649
1650 if (bufsize < 4)
1651 return seterror(E2BIG);
1652 wc = utf16_to_ucs4(wbuf);
1653 if (cv->codepage == 12000) /* little endian */
1654 {
1655 buf[0] = wc & 0x000000FF;
1656 buf[1] = (wc & 0x0000FF00) >> 8;
1657 buf[2] = (wc & 0x00FF0000) >> 16;
1658 buf[3] = (wc & 0xFF000000) >> 24;
1659 }
1660 else if (cv->codepage == 12001) /* big endian */
1661 {
1662 buf[0] = (wc & 0xFF000000) >> 24;
1663 buf[1] = (wc & 0x00FF0000) >> 16;
1664 buf[2] = (wc & 0x0000FF00) >> 8;
1665 buf[3] = wc & 0x000000FF;
1666 }
1667 return 4;
1668 }
1669
1670 /*
1671 * 50220: ISO 2022 Japanese with no halfwidth Katakana; Japanese (JIS)
1672 * 50221: ISO 2022 Japanese with halfwidth Katakana; Japanese (JIS-Allow
1673 * 1 byte Kana)
1674 * 50222: ISO 2022 Japanese JIS X 0201-1989; Japanese (JIS-Allow 1 byte
1675 * Kana - SO/SI)
1676 *
1677 * MultiByteToWideChar() and WideCharToMultiByte() behave differently
1678 * depending on Windows version. On XP, WideCharToMultiByte() doesn't
1679 * terminate result sequence with ascii escape. But Vista does.
1680 * Use MLang instead.
1681 */
1682
1683 #define ISO2022_MODE(cs, shift) (((cs) << 8) | (shift))
1684 #define ISO2022_MODE_CS(mode) (((mode) >> 8) & 0xFF)
1685 #define ISO2022_MODE_SHIFT(mode) ((mode) & 0xFF)
1686
1687 #define ISO2022_SI 0
1688 #define ISO2022_SO 1
1689
1690 /* shift in */
1691 static const char iso2022_SI_seq[] = "\x0F";
1692 /* shift out */
1693 static const char iso2022_SO_seq[] = "\x0E";
1694
1695 typedef struct iso2022_esc_t iso2022_esc_t;
1696 struct iso2022_esc_t {
1697 const char *esc;
1698 int esc_len;
1699 int len;
1700 int cs;
1701 };
1702
1703 #define ISO2022JP_CS_ASCII 0
1704 #define ISO2022JP_CS_JISX0201_ROMAN 1
1705 #define ISO2022JP_CS_JISX0201_KANA 2
1706 #define ISO2022JP_CS_JISX0208_1978 3
1707 #define ISO2022JP_CS_JISX0208_1983 4
1708 #define ISO2022JP_CS_JISX0212 5
1709
1710 static iso2022_esc_t iso2022jp_esc[] = {
1711 {"\x1B\x28\x42", 3, 1, ISO2022JP_CS_ASCII},
1712 {"\x1B\x28\x4A", 3, 1, ISO2022JP_CS_JISX0201_ROMAN},
1713 {"\x1B\x28\x49", 3, 1, ISO2022JP_CS_JISX0201_KANA},
1714 {"\x1B\x24\x40", 3, 2, ISO2022JP_CS_JISX0208_1983}, /* unify 1978 with 1983 */
1715 {"\x1B\x24\x42", 3, 2, ISO2022JP_CS_JISX0208_1983},
1716 {"\x1B\x24\x28\x44", 4, 2, ISO2022JP_CS_JISX0212},
1717 {NULL, 0, 0, 0}
1718 };
1719
1720 static int
1721 iso2022jp_mbtowc(csconv_t *cv, const uchar *buf, int bufsize, ushort *wbuf, int *wbufsize)
1722 {
1723 iso2022_esc_t *iesc = iso2022jp_esc;
1724 char tmp[MB_CHAR_MAX];
1725 int insize;
1726 HRESULT hr;
1727 DWORD dummy = 0;
1728 int len;
1729 int esc_len;
1730 int cs;
1731 int shift;
1732 int i;
1733
1734 if (buf[0] == 0x1B)
1735 {
1736 for (i = 0; iesc[i].esc != NULL; ++i)
1737 {
1738 esc_len = iesc[i].esc_len;
1739 if (bufsize < esc_len)
1740 {
1741 if (strncmp((char *)buf, iesc[i].esc, bufsize) == 0)
1742 return seterror(EINVAL);
1743 }
1744 else
1745 {
1746 if (strncmp((char *)buf, iesc[i].esc, esc_len) == 0)
1747 {
1748 cv->mode = ISO2022_MODE(iesc[i].cs, ISO2022_SI);
1749 *wbufsize = 0;
1750 return esc_len;
1751 }
1752 }
1753 }
1754 /* not supported escape sequence */
1755 return seterror(EILSEQ);
1756 }
1757 else if (buf[0] == iso2022_SO_seq[0])
1758 {
1759 cv->mode = ISO2022_MODE(ISO2022_MODE_CS(cv->mode), ISO2022_SO);
1760 *wbufsize = 0;
1761 return 1;
1762 }
1763 else if (buf[0] == iso2022_SI_seq[0])
1764 {
1765 cv->mode = ISO2022_MODE(ISO2022_MODE_CS(cv->mode), ISO2022_SI);
1766 *wbufsize = 0;
1767 return 1;
1768 }
1769
1770 cs = ISO2022_MODE_CS(cv->mode);
1771 shift = ISO2022_MODE_SHIFT(cv->mode);
1772
1773 /* reset the mode for informal sequence */
1774 if (buf[0] < 0x20)
1775 {
1776 cs = ISO2022JP_CS_ASCII;
1777 shift = ISO2022_SI;
1778 }
1779
1780 len = iesc[cs].len;
1781 if (bufsize < len)
1782 return seterror(EINVAL);
1783 for (i = 0; i < len; ++i)
1784 if (!(buf[i] < 0x80))
1785 return seterror(EILSEQ);
1786 esc_len = iesc[cs].esc_len;
1787 memcpy(tmp, iesc[cs].esc, esc_len);
1788 if (shift == ISO2022_SO)
1789 {
1790 memcpy(tmp + esc_len, iso2022_SO_seq, 1);
1791 esc_len += 1;
1792 }
1793 memcpy(tmp + esc_len, buf, len);
1794
1795 if ((cv->codepage == 50220 || cv->codepage == 50221
1796 || cv->codepage == 50222) && shift == ISO2022_SO)
1797 {
1798 /* XXX: shift-out cannot be used for mbtowc (both kernel and
1799 * mlang) */
1800 esc_len = iesc[ISO2022JP_CS_JISX0201_KANA].esc_len;
1801 memcpy(tmp, iesc[ISO2022JP_CS_JISX0201_KANA].esc, esc_len);
1802 memcpy(tmp + esc_len, buf, len);
1803 }
1804
1805 insize = len + esc_len;
1806 hr = ConvertINetMultiByteToUnicode(&dummy, cv->codepage,
1807 (const char *)tmp, &insize, (wchar_t *)wbuf, wbufsize);
1808 if (hr != S_OK || insize != len + esc_len)
1809 return seterror(EILSEQ);
1810
1811 /* Check for conversion error. Assuming defaultChar is 0x3F. */
1812 /* ascii should be converted from ascii */
1813 if (wbuf[0] == buf[0]
1814 && cv->mode != ISO2022_MODE(ISO2022JP_CS_ASCII, ISO2022_SI))
1815 return seterror(EILSEQ);
1816
1817 /* reset the mode for informal sequence */
1818 if (cv->mode != ISO2022_MODE(cs, shift))
1819 cv->mode = ISO2022_MODE(cs, shift);
1820
1821 return len;
1822 }
1823
1824 static int
1825 iso2022jp_wctomb(csconv_t *cv, ushort *wbuf, int wbufsize, uchar *buf, int bufsize)
1826 {
1827 iso2022_esc_t *iesc = iso2022jp_esc;
1828 char tmp[MB_CHAR_MAX];
1829 int tmpsize = MB_CHAR_MAX;
1830 int insize = wbufsize;
1831 HRESULT hr;
1832 DWORD dummy = 0;
1833 int len;
1834 int esc_len;
1835 int cs;
1836 int shift;
1837 int i;
1838
1839 /*
1840 * MultiByte = [escape sequence] + character + [escape sequence]
1841 *
1842 * Whether trailing escape sequence is added depends on which API is
1843 * used (kernel or MLang, and its version).
1844 */
1845 hr = ConvertINetUnicodeToMultiByte(&dummy, cv->codepage,
1846 (const wchar_t *)wbuf, &wbufsize, tmp, &tmpsize);
1847 if (hr != S_OK || insize != wbufsize)
1848 return seterror(EILSEQ);
1849 else if (bufsize < tmpsize)
1850 return seterror(E2BIG);
1851
1852 if (tmpsize == 1)
1853 {
1854 cs = ISO2022JP_CS_ASCII;
1855 esc_len = 0;
1856 }
1857 else
1858 {
1859 for (i = 1; iesc[i].esc != NULL; ++i)
1860 {
1861 esc_len = iesc[i].esc_len;
1862 if (strncmp(tmp, iesc[i].esc, esc_len) == 0)
1863 {
1864 cs = iesc[i].cs;
1865 break;
1866 }
1867 }
1868 if (iesc[i].esc == NULL)
1869 /* not supported escape sequence */
1870 return seterror(EILSEQ);
1871 }
1872
1873 shift = ISO2022_SI;
1874 if (tmp[esc_len] == iso2022_SO_seq[0])
1875 {
1876 shift = ISO2022_SO;
1877 esc_len += 1;
1878 }
1879
1880 len = iesc[cs].len;
1881
1882 /* Check for converting error. Assuming defaultChar is 0x3F. */
1883 /* ascii should be converted from ascii */
1884 if (cs == ISO2022JP_CS_ASCII && !(wbuf[0] < 0x80))
1885 return seterror(EILSEQ);
1886 else if (tmpsize < esc_len + len)
1887 return seterror(EILSEQ);
1888
1889 if (cv->mode == ISO2022_MODE(cs, shift))
1890 {
1891 /* remove escape sequence */
1892 if (esc_len != 0)
1893 memmove(tmp, tmp + esc_len, len);
1894 esc_len = 0;
1895 }
1896 else
1897 {
1898 if (cs == ISO2022JP_CS_ASCII)
1899 {
1900 esc_len = iesc[ISO2022JP_CS_ASCII].esc_len;
1901 memmove(tmp + esc_len, tmp, len);
1902 memcpy(tmp, iesc[ISO2022JP_CS_ASCII].esc, esc_len);
1903 }
1904 if (ISO2022_MODE_SHIFT(cv->mode) == ISO2022_SO)
1905 {
1906 /* shift-in before changing to other mode */
1907 memmove(tmp + 1, tmp, len + esc_len);
1908 memcpy(tmp, iso2022_SI_seq, 1);
1909 esc_len += 1;
1910 }
1911 }
1912
1913 if (bufsize < len + esc_len)
1914 return seterror(E2BIG);
1915 memcpy(buf, tmp, len + esc_len);
1916 cv->mode = ISO2022_MODE(cs, shift);
1917 return len + esc_len;
1918 }
1919
1920 static int
1921 iso2022jp_flush(csconv_t *cv, uchar *buf, int bufsize)
1922 {
1923 iso2022_esc_t *iesc = iso2022jp_esc;
1924 int esc_len;
1925
1926 if (cv->mode != ISO2022_MODE(ISO2022JP_CS_ASCII, ISO2022_SI))
1927 {
1928 esc_len = 0;
1929 if (ISO2022_MODE_SHIFT(cv->mode) != ISO2022_SI)
1930 esc_len += 1;
1931 if (ISO2022_MODE_CS(cv->mode) != ISO2022JP_CS_ASCII)
1932 esc_len += iesc[ISO2022JP_CS_ASCII].esc_len;
1933 if (bufsize < esc_len)
1934 return seterror(E2BIG);
1935
1936 esc_len = 0;
1937 if (ISO2022_MODE_SHIFT(cv->mode) != ISO2022_SI)
1938 {
1939 memcpy(buf, iso2022_SI_seq, 1);
1940 esc_len += 1;
1941 }
1942 if (ISO2022_MODE_CS(cv->mode) != ISO2022JP_CS_ASCII)
1943 {
1944 memcpy(buf + esc_len, iesc[ISO2022JP_CS_ASCII].esc,
1945 iesc[ISO2022JP_CS_ASCII].esc_len);
1946 esc_len += iesc[ISO2022JP_CS_ASCII].esc_len;
1947 }
1948 return esc_len;
1949 }
1950 return 0;
1951 }
1952
1953 #if defined(MAKE_DLL) && defined(USE_LIBICONV_DLL)
1954 BOOL WINAPI
1955 DllMain(HINSTANCE hinstDLL, DWORD fdwReason, LPVOID lpReserved)
1956 {
1957 switch( fdwReason )
1958 {
1959 case DLL_PROCESS_ATTACH:
1960 hwiniconv = (HMODULE)hinstDLL;
1961 break;
1962 case DLL_THREAD_ATTACH:
1963 case DLL_THREAD_DETACH:
1964 case DLL_PROCESS_DETACH:
1965 break;
1966 }
1967 return TRUE;
1968 }
1969 #endif
1970
1971 #if defined(MAKE_EXE)
1972 #include <stdio.h>
1973 #include <fcntl.h>
1974 #include <io.h>
1975 int
1976 main(int argc, char **argv)
1977 {
1978 char *fromcode = NULL;
1979 char *tocode = NULL;
1980 int i;
1981 char inbuf[BUFSIZ];
1982 char outbuf[BUFSIZ];
1983 const char *pin;
1984 char *pout;
1985 size_t inbytesleft;
1986 size_t outbytesleft;
1987 size_t rest = 0;
1988 iconv_t cd;
1989 size_t r;
1990 FILE *in = stdin;
1991 FILE *out = stdout;
1992 FILE *in_allocated = NULL, *out_allocated = NULL;
1993 int ignore = 0;
1994 char *p;
1995
1996 _setmode(_fileno(stdin), _O_BINARY);
1997 _setmode(_fileno(stdout), _O_BINARY);
1998
1999 for (i = 1; i < argc; ++i)
2000 {
2001 if (strcmp(argv[i], "-l") == 0)
2002 {
2003 for (i = 0; codepage_alias[i].name != NULL; ++i)
2004 printf("%s\n", codepage_alias[i].name);
2005 return 0;
2006 }
2007
2008 if (strcmp(argv[i], "-f") == 0)
2009 fromcode = argv[++i];
2010 else if (strcmp(argv[i], "-t") == 0)
2011 tocode = argv[++i];
2012 else if (strcmp(argv[i], "-c") == 0)
2013 ignore = 1;
2014 else if (strcmp(argv[i], "--output") == 0)
2015 {
2016 out_allocated = out = fopen(argv[++i], "wb");
2017 if(out == NULL)
2018 {
2019 fprintf(stderr, "cannot open %s\n", argv[i]);
2020 return 1;
2021 }
2022 }
2023 else
2024 {
2025 in_allocated = in = fopen(argv[i], "rb");
2026 if (in == NULL)
2027 {
2028 fprintf(stderr, "cannot open %s\n", argv[i]);
2029 return 1;
2030 }
2031 break;
2032 }
2033 }
2034
2035 if (fromcode == NULL || tocode == NULL)
2036 {
2037 printf("usage: %s [-c] -f from-enc -t to-enc [file]\n", argv[0]);
2038 return 0;
2039 }
2040
2041 if (ignore)
2042 {
2043 p = tocode;
2044 tocode = (char *)malloc(strlen(p) + strlen("//IGNORE") + 1);
2045 if (tocode == NULL)
2046 {
2047 perror("fatal error");
2048 return 1;
2049 }
2050 strcpy(tocode, p);
2051 strcat(tocode, "//IGNORE");
2052 }
2053
2054 cd = iconv_open(tocode, fromcode);
2055 if (cd == (iconv_t)(-1))
2056 {
2057 perror("iconv_open error");
2058 return 1;
2059 }
2060
2061 while ((inbytesleft = fread(inbuf + rest, 1, sizeof(inbuf) - rest, in)) != 0
2062 || rest != 0)
2063 {
2064 inbytesleft += rest;
2065 pin = inbuf;
2066 pout = outbuf;
2067 outbytesleft = sizeof(outbuf);
2068 r = iconv(cd, &pin, &inbytesleft, &pout, &outbytesleft);
2069 fwrite(outbuf, 1, sizeof(outbuf) - outbytesleft, out);
2070 if (r == (size_t)(-1) && errno != E2BIG && (errno != EINVAL || feof(in)))
2071 {
2072 perror("conversion error");
2073 return 1;
2074 }
2075 memmove(inbuf, pin, inbytesleft);
2076 rest = inbytesleft;
2077 }
2078 pout = outbuf;
2079 outbytesleft = sizeof(outbuf);
2080 r = iconv(cd, NULL, NULL, &pout, &outbytesleft);
2081 fwrite(outbuf, 1, sizeof(outbuf) - outbytesleft, out);
2082 if (r == (size_t)(-1))
2083 {
2084 perror("conversion error");
2085 return 1;
2086 }
2087
2088 iconv_close(cd);
2089
2090 if (in_allocated != NULL)
2091 fclose (in_allocated);
2092 if (out_allocated != NULL)
2093 fclose (out_allocated);
2094
2095 return 0;
2096 }
2097 #endif
glib mirror