2005-09-18 [colin] 1.9.14cvs39
[claws.git] / src / codeconv.c
1 /*
2  * Sylpheed -- a GTK+ based, lightweight, and fast e-mail client
3  * Copyright (C) 1999-2005 Hiroyuki Yamamoto
4  *
5  * This program is free software; you can redistribute it and/or modify
6  * it under the terms of the GNU General Public License as published by
7  * the Free Software Foundation; either version 2 of the License, or
8  * (at your option) any later version.
9  *
10  * This program is distributed in the hope that it will be useful,
11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
13  * GNU General Public License for more details.
14  *
15  * You should have received a copy of the GNU General Public License
16  * along with this program; if not, write to the Free Software
17  * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
18  */
19
20 #ifdef HAVE_CONFIG_H
21 #  include "config.h"
22 #endif
23
24 #include "defs.h"
25
26 #include <glib.h>
27 #include <glib/gi18n.h>
28 #include <string.h>
29 #include <ctype.h>
30 #include <stdlib.h>
31 #include <errno.h>
32
33 #if HAVE_LOCALE_H
34 #  include <locale.h>
35 #endif
36
37 #include <iconv.h>
38
39 #include "codeconv.h"
40 #include "unmime.h"
41 #include "base64.h"
42 #include "quoted-printable.h"
43 #include "utils.h"
44 #include "prefs_common.h"
45
46 typedef enum
47 {
48         JIS_ASCII,
49         JIS_KANJI,
50         JIS_HWKANA,
51         JIS_AUXKANJI
52 } JISState;
53
54 #define SUBST_CHAR      0x5f;
55 #define ESC             '\033'
56
57 #define iseuckanji(c) \
58         (((c) & 0xff) >= 0xa1 && ((c) & 0xff) <= 0xfe)
59 #define iseuchwkana1(c) \
60         (((c) & 0xff) == 0x8e)
61 #define iseuchwkana2(c) \
62         (((c) & 0xff) >= 0xa1 && ((c) & 0xff) <= 0xdf)
63 #define iseucaux(c) \
64         (((c) & 0xff) == 0x8f)
65 #define issjiskanji1(c) \
66         ((((c) & 0xff) >= 0x81 && ((c) & 0xff) <= 0x9f) || \
67          (((c) & 0xff) >= 0xe0 && ((c) & 0xff) <= 0xfc))
68 #define issjiskanji2(c) \
69         ((((c) & 0xff) >= 0x40 && ((c) & 0xff) <= 0x7e) || \
70          (((c) & 0xff) >= 0x80 && ((c) & 0xff) <= 0xfc))
71 #define issjishwkana(c) \
72         (((c) & 0xff) >= 0xa1 && ((c) & 0xff) <= 0xdf)
73
74 #define K_IN()                          \
75         if (state != JIS_KANJI) {       \
76                 *out++ = ESC;           \
77                 *out++ = '$';           \
78                 *out++ = 'B';           \
79                 state = JIS_KANJI;      \
80         }
81
82 #define K_OUT()                         \
83         if (state != JIS_ASCII) {       \
84                 *out++ = ESC;           \
85                 *out++ = '(';           \
86                 *out++ = 'B';           \
87                 state = JIS_ASCII;      \
88         }
89
90 #define HW_IN()                         \
91         if (state != JIS_HWKANA) {      \
92                 *out++ = ESC;           \
93                 *out++ = '(';           \
94                 *out++ = 'I';           \
95                 state = JIS_HWKANA;     \
96         }
97
98 #define AUX_IN()                        \
99         if (state != JIS_AUXKANJI) {    \
100                 *out++ = ESC;           \
101                 *out++ = '$';           \
102                 *out++ = '(';           \
103                 *out++ = 'D';           \
104                 state = JIS_AUXKANJI;   \
105         }
106
107 static void conv_jistoeuc(gchar *outbuf, gint outlen, const gchar *inbuf);
108 static void conv_euctojis(gchar *outbuf, gint outlen, const gchar *inbuf);
109 static void conv_sjistoeuc(gchar *outbuf, gint outlen, const gchar *inbuf);
110
111 static void conv_jistoutf8(gchar *outbuf, gint outlen, const gchar *inbuf);
112 static void conv_sjistoutf8(gchar *outbuf, gint outlen, const gchar *inbuf);
113 static void conv_euctoutf8(gchar *outbuf, gint outlen, const gchar *inbuf);
114 static void conv_anytoutf8(gchar *outbuf, gint outlen, const gchar *inbuf);
115
116 static void conv_utf8toeuc(gchar *outbuf, gint outlen, const gchar *inbuf);
117 static void conv_utf8tojis(gchar *outbuf, gint outlen, const gchar *inbuf);
118
119 static void conv_unreadable_8bit(gchar *str);
120
121 static void conv_jistodisp(gchar *outbuf, gint outlen, const gchar *inbuf);
122 static void conv_sjistodisp(gchar *outbuf, gint outlen, const gchar *inbuf);
123 static void conv_euctodisp(gchar *outbuf, gint outlen, const gchar *inbuf);
124
125 static void conv_anytodisp(gchar *outbuf, gint outlen, const gchar *inbuf);
126 static void conv_ustodisp(gchar *outbuf, gint outlen, const gchar *inbuf);
127 static void conv_noconv(gchar *outbuf, gint outlen, const gchar *inbuf);
128
129 static void conv_jistoeuc(gchar *outbuf, gint outlen, const gchar *inbuf)
130 {
131         const guchar *in = inbuf;
132         guchar *out = outbuf;
133         JISState state = JIS_ASCII;
134
135         while (*in != '\0') {
136                 if (*in == ESC) {
137                         in++;
138                         if (*in == '$') {
139                                 if (*(in + 1) == '@' || *(in + 1) == 'B') {
140                                         state = JIS_KANJI;
141                                         in += 2;
142                                 } else if (*(in + 1) == '(' &&
143                                            *(in + 2) == 'D') {
144                                         state = JIS_AUXKANJI;
145                                         in += 3;
146                                 } else {
147                                         /* unknown escape sequence */
148                                         state = JIS_ASCII;
149                                 }
150                         } else if (*in == '(') {
151                                 if (*(in + 1) == 'B' || *(in + 1) == 'J') {
152                                         state = JIS_ASCII;
153                                         in += 2;
154                                 } else if (*(in + 1) == 'I') {
155                                         state = JIS_HWKANA;
156                                         in += 2;
157                                 } else {
158                                         /* unknown escape sequence */
159                                         state = JIS_ASCII;
160                                 }
161                         } else {
162                                 /* unknown escape sequence */
163                                 state = JIS_ASCII;
164                         }
165                 } else if (*in == 0x0e) {
166                         state = JIS_HWKANA;
167                         in++;
168                 } else if (*in == 0x0f) {
169                         state = JIS_ASCII;
170                         in++;
171                 } else {
172                         switch (state) {
173                         case JIS_ASCII:
174                                 *out++ = *in++;
175                                 break;
176                         case JIS_KANJI:
177                                 *out++ = *in++ | 0x80;
178                                 if (*in == '\0') break;
179                                 *out++ = *in++ | 0x80;
180                                 break;
181                         case JIS_HWKANA:
182                                 *out++ = 0x8e;
183                                 *out++ = *in++ | 0x80;
184                                 break;
185                         case JIS_AUXKANJI:
186                                 *out++ = 0x8f;
187                                 *out++ = *in++ | 0x80;
188                                 if (*in == '\0') break;
189                                 *out++ = *in++ | 0x80;
190                                 break;
191                         }
192                 }
193         }
194
195         *out = '\0';
196 }
197
198 #define JIS_HWDAKUTEN           0x5e
199 #define JIS_HWHANDAKUTEN        0x5f
200
201 static gint conv_jis_hantozen(guchar *outbuf, guchar jis_code, guchar sound_sym)
202 {
203         static guint16 h2z_tbl[] = {
204                 /* 0x20 - 0x2f */
205                 0x0000, 0x2123, 0x2156, 0x2157, 0x2122, 0x2126, 0x2572, 0x2521,
206                 0x2523, 0x2525, 0x2527, 0x2529, 0x2563, 0x2565, 0x2567, 0x2543,
207                 /* 0x30 - 0x3f */
208                 0x213c, 0x2522, 0x2524, 0x2526, 0x2528, 0x252a, 0x252b, 0x252d,
209                 0x252f, 0x2531, 0x2533, 0x2535, 0x2537, 0x2539, 0x253b, 0x253d,
210                 /* 0x40 - 0x4f */
211                 0x253f, 0x2541, 0x2544, 0x2546, 0x2548, 0x254a, 0x254b, 0x254c,
212                 0x254d, 0x254e, 0x254f, 0x2552, 0x2555, 0x2558, 0x255b, 0x255e,
213                 /* 0x50 - 0x5f */
214                 0x255f, 0x2560, 0x2561, 0x2562, 0x2564, 0x2566, 0x2568, 0x2569,
215                 0x256a, 0x256b, 0x256c, 0x256d, 0x256f, 0x2573, 0x212b, 0x212c
216         };
217
218         static guint16 dakuten_tbl[] = {
219                 /* 0x30 - 0x3f */
220                 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x252c, 0x252e,
221                 0x2530, 0x2532, 0x2534, 0x2536, 0x2538, 0x253a, 0x253c, 0x253e,
222                 /* 0x40 - 0x4f */
223                 0x2540, 0x2542, 0x2545, 0x2547, 0x2549, 0x0000, 0x0000, 0x0000,
224                 0x0000, 0x0000, 0x2550, 0x2553, 0x2556, 0x2559, 0x255c, 0x0000
225         };
226
227         static guint16 handakuten_tbl[] = {
228                 /* 0x4a - 0x4e */
229                 0x2551, 0x2554, 0x2557, 0x255a, 0x255d
230         };
231
232         guint16 out_code;
233
234         jis_code &= 0x7f;
235         sound_sym &= 0x7f;
236
237         if (jis_code < 0x21 || jis_code > 0x5f)
238                 return 0;
239
240         if (sound_sym == JIS_HWDAKUTEN &&
241             jis_code >= 0x36 && jis_code <= 0x4e) {
242                 out_code = dakuten_tbl[jis_code - 0x30];
243                 if (out_code != 0) {
244                         *outbuf = out_code >> 8;
245                         *(outbuf + 1) = out_code & 0xff;
246                         return 2;
247                 }
248         }
249
250         if (sound_sym == JIS_HWHANDAKUTEN &&
251             jis_code >= 0x4a && jis_code <= 0x4e) {
252                 out_code = handakuten_tbl[jis_code - 0x4a];
253                 *outbuf = out_code >> 8;
254                 *(outbuf + 1) = out_code & 0xff;
255                 return 2;
256         }
257
258         out_code = h2z_tbl[jis_code - 0x20];
259         *outbuf = out_code >> 8;
260         *(outbuf + 1) = out_code & 0xff;
261         return 1;
262 }
263
264 static void conv_euctojis(gchar *outbuf, gint outlen, const gchar *inbuf)
265 {
266         const guchar *in = inbuf;
267         guchar *out = outbuf;
268         JISState state = JIS_ASCII;
269
270         while (*in != '\0') {
271                 if (IS_ASCII(*in)) {
272                         K_OUT();
273                         *out++ = *in++;
274                 } else if (iseuckanji(*in)) {
275                         if (iseuckanji(*(in + 1))) {
276                                 K_IN();
277                                 *out++ = *in++ & 0x7f;
278                                 *out++ = *in++ & 0x7f;
279                         } else {
280                                 K_OUT();
281                                 *out++ = SUBST_CHAR;
282                                 in++;
283                                 if (*in != '\0' && !IS_ASCII(*in)) {
284                                         *out++ = SUBST_CHAR;
285                                         in++;
286                                 }
287                         }
288                 } else if (iseuchwkana1(*in)) {
289                         if (iseuchwkana2(*(in + 1))) {
290                                 if (prefs_common.allow_jisx0201_kana) {
291                                         HW_IN();
292                                         in++;
293                                         *out++ = *in++ & 0x7f;
294                                 } else {
295                                         guchar jis_ch[2];
296                                         gint len;
297
298                                         if (iseuchwkana1(*(in + 2)) &&
299                                             iseuchwkana2(*(in + 3)))
300                                                 len = conv_jis_hantozen
301                                                         (jis_ch,
302                                                          *(in + 1), *(in + 3));
303                                         else
304                                                 len = conv_jis_hantozen
305                                                         (jis_ch,
306                                                          *(in + 1), '\0');
307                                         if (len == 0)
308                                                 in += 2;
309                                         else {
310                                                 K_IN();
311                                                 in += len * 2;
312                                                 *out++ = jis_ch[0];
313                                                 *out++ = jis_ch[1];
314                                         }
315                                 }
316                         } else {
317                                 K_OUT();
318                                 in++;
319                                 if (*in != '\0' && !IS_ASCII(*in)) {
320                                         *out++ = SUBST_CHAR;
321                                         in++;
322                                 }
323                         }
324                 } else if (iseucaux(*in)) {
325                         in++;
326                         if (iseuckanji(*in) && iseuckanji(*(in + 1))) {
327                                 AUX_IN();
328                                 *out++ = *in++ & 0x7f;
329                                 *out++ = *in++ & 0x7f;
330                         } else {
331                                 K_OUT();
332                                 if (*in != '\0' && !IS_ASCII(*in)) {
333                                         *out++ = SUBST_CHAR;
334                                         in++;
335                                         if (*in != '\0' && !IS_ASCII(*in)) {
336                                                 *out++ = SUBST_CHAR;
337                                                 in++;
338                                         }
339                                 }
340                         }
341                 } else {
342                         K_OUT();
343                         *out++ = SUBST_CHAR;
344                         in++;
345                 }
346         }
347
348         K_OUT();
349         *out = '\0';
350 }
351
352 static void conv_sjistoeuc(gchar *outbuf, gint outlen, const gchar *inbuf)
353 {
354         const guchar *in = inbuf;
355         guchar *out = outbuf;
356
357         while (*in != '\0') {
358                 if (IS_ASCII(*in)) {
359                         *out++ = *in++;
360                 } else if (issjiskanji1(*in)) {
361                         if (issjiskanji2(*(in + 1))) {
362                                 guchar out1 = *in;
363                                 guchar out2 = *(in + 1);
364                                 guchar row;
365
366                                 row = out1 < 0xa0 ? 0x70 : 0xb0;
367                                 if (out2 < 0x9f) {
368                                         out1 = (out1 - row) * 2 - 1;
369                                         out2 -= out2 > 0x7f ? 0x20 : 0x1f;
370                                 } else {
371                                         out1 = (out1 - row) * 2;
372                                         out2 -= 0x7e;
373                                 }
374
375                                 *out++ = out1 | 0x80;
376                                 *out++ = out2 | 0x80;
377                                 in += 2;
378                         } else {
379                                 *out++ = SUBST_CHAR;
380                                 in++;
381                                 if (*in != '\0' && !IS_ASCII(*in)) {
382                                         *out++ = SUBST_CHAR;
383                                         in++;
384                                 }
385                         }
386                 } else if (issjishwkana(*in)) {
387                         *out++ = 0x8e;
388                         *out++ = *in++;
389                 } else {
390                         *out++ = SUBST_CHAR;
391                         in++;
392                 }
393         }
394
395         *out = '\0';
396 }
397
398 static void conv_jistoutf8(gchar *outbuf, gint outlen, const gchar *inbuf)
399 {
400         gchar *eucstr;
401
402         Xalloca(eucstr, outlen, return);
403
404         conv_jistoeuc(eucstr, outlen, inbuf);
405         conv_euctoutf8(outbuf, outlen, eucstr);
406 }
407
408 static void conv_sjistoutf8(gchar *outbuf, gint outlen, const gchar *inbuf)
409 {
410         gchar *tmpstr;
411
412         tmpstr = conv_iconv_strdup(inbuf, CS_SHIFT_JIS, CS_UTF_8);
413         if (tmpstr) {
414                 strncpy2(outbuf, tmpstr, outlen);
415                 g_free(tmpstr);
416         } else
417                 strncpy2(outbuf, inbuf, outlen);
418 }
419
420 static void conv_euctoutf8(gchar *outbuf, gint outlen, const gchar *inbuf)
421 {
422         static iconv_t cd = (iconv_t)-1;
423         static gboolean iconv_ok = TRUE;
424         gchar *tmpstr;
425
426         if (cd == (iconv_t)-1) {
427                 if (!iconv_ok) {
428                         strncpy2(outbuf, inbuf, outlen);
429                         return;
430                 }
431                 cd = iconv_open(CS_UTF_8, CS_EUC_JP_MS);
432                 if (cd == (iconv_t)-1) {
433                         cd = iconv_open(CS_UTF_8, CS_EUC_JP);
434                         if (cd == (iconv_t)-1) {
435                                 g_warning("conv_euctoutf8(): %s\n",
436                                           g_strerror(errno));
437                                 iconv_ok = FALSE;
438                                 strncpy2(outbuf, inbuf, outlen);
439                                 return;
440                         }
441                 }
442         }
443
444         tmpstr = conv_iconv_strdup_with_cd(inbuf, cd);
445         if (tmpstr) {
446                 strncpy2(outbuf, tmpstr, outlen);
447                 g_free(tmpstr);
448         } else
449                 strncpy2(outbuf, inbuf, outlen);
450 }
451
452 static void conv_anytoutf8(gchar *outbuf, gint outlen, const gchar *inbuf)
453 {
454         switch (conv_guess_ja_encoding(inbuf)) {
455         case C_ISO_2022_JP:
456                 conv_jistoutf8(outbuf, outlen, inbuf);
457                 break;
458         case C_SHIFT_JIS:
459                 conv_sjistoutf8(outbuf, outlen, inbuf);
460                 break;
461         case C_EUC_JP:
462                 conv_euctoutf8(outbuf, outlen, inbuf);
463                 break;
464         default:
465                 strncpy2(outbuf, inbuf, outlen);
466                 break;
467         }
468 }
469
470 static void conv_utf8toeuc(gchar *outbuf, gint outlen, const gchar *inbuf)
471 {
472         static iconv_t cd = (iconv_t)-1;
473         static gboolean iconv_ok = TRUE;
474         gchar *tmpstr;
475
476         if (cd == (iconv_t)-1) {
477                 if (!iconv_ok) {
478                         strncpy2(outbuf, inbuf, outlen);
479                         return;
480                 }
481                 cd = iconv_open(CS_EUC_JP_MS, CS_UTF_8);
482                 if (cd == (iconv_t)-1) {
483                         cd = iconv_open(CS_EUC_JP, CS_UTF_8);
484                         if (cd == (iconv_t)-1) {
485                                 g_warning("conv_utf8toeuc(): %s\n",
486                                           g_strerror(errno));
487                                 iconv_ok = FALSE;
488                                 strncpy2(outbuf, inbuf, outlen);
489                                 return;
490                         }
491                 }
492         }
493
494         tmpstr = conv_iconv_strdup_with_cd(inbuf, cd);
495         if (tmpstr) {
496                 strncpy2(outbuf, tmpstr, outlen);
497                 g_free(tmpstr);
498         } else
499                 strncpy2(outbuf, inbuf, outlen);
500 }
501
502 static void conv_utf8tojis(gchar *outbuf, gint outlen, const gchar *inbuf)
503 {
504         gchar *eucstr;
505
506         Xalloca(eucstr, outlen, return);
507
508         conv_utf8toeuc(eucstr, outlen, inbuf);
509         conv_euctojis(outbuf, outlen, eucstr);
510 }
511
512 static void conv_unreadable_8bit(gchar *str)
513 {
514         register guchar *p = str;
515
516         while (*p != '\0') {
517                 /* convert CR+LF -> LF */
518                 if (*p == '\r' && *(p + 1) == '\n')
519                         memmove(p, p + 1, strlen(p));
520                 else if (!IS_ASCII(*p)) *p = SUBST_CHAR;
521                 p++;
522         }
523 }
524
525 #define NCV     '\0'
526
527 void conv_mb_alnum(gchar *str)
528 {
529         static guchar char_tbl[] = {
530                 /* 0xa0 - 0xaf */
531                 NCV, ' ', NCV, NCV, ',', '.', NCV, ':',
532                 ';', '?', '!', NCV, NCV, NCV, NCV, NCV,
533                 /* 0xb0 - 0xbf */
534                 NCV, NCV, NCV, NCV, NCV, NCV, NCV, NCV,
535                 NCV, NCV, NCV, NCV, NCV, NCV, NCV, NCV,
536                 /* 0xc0 - 0xcf */
537                 NCV, NCV, NCV, NCV, NCV, NCV, NCV, NCV,
538                 NCV, NCV, '(', ')', NCV, NCV, '[', ']',
539                 /* 0xd0 - 0xdf */
540                 '{', '}', NCV, NCV, NCV, NCV, NCV, NCV,
541                 NCV, NCV, NCV, NCV, '+', '-', NCV, NCV,
542                 /* 0xe0 - 0xef */
543                 NCV, '=', NCV, '<', '>', NCV, NCV, NCV,
544                 NCV, NCV, NCV, NCV, NCV, NCV, NCV, NCV
545         };
546
547         register guchar *p = str;
548         register gint len;
549
550         len = strlen(str);
551
552         while (len > 1) {
553                 if (*p == 0xa3) {
554                         register guchar ch = *(p + 1);
555
556                         if (ch >= 0xb0 && ch <= 0xfa) {
557                                 /* [a-zA-Z] */
558                                 *p = ch & 0x7f;
559                                 p++;
560                                 len--;
561                                 memmove(p, p + 1, len);
562                                 len--;
563                         } else  {
564                                 p += 2;
565                                 len -= 2;
566                         }
567                 } else if (*p == 0xa1) {
568                         register guchar ch = *(p + 1);
569
570                         if (ch >= 0xa0 && ch <= 0xef &&
571                             NCV != char_tbl[ch - 0xa0]) {
572                                 *p = char_tbl[ch - 0xa0];
573                                 p++;
574                                 len--;
575                                 memmove(p, p + 1, len);
576                                 len--;
577                         } else {
578                                 p += 2;
579                                 len -= 2;
580                         }
581                 } else if (iseuckanji(*p)) {
582                         p += 2;
583                         len -= 2;
584                 } else {
585                         p++;
586                         len--;
587                 }
588         }
589 }
590
591 CharSet conv_guess_ja_encoding(const gchar *str)
592 {
593         const guchar *p = str;
594         CharSet guessed = C_US_ASCII;
595
596         while (*p != '\0') {
597                 if (*p == ESC && (*(p + 1) == '$' || *(p + 1) == '(')) {
598                         if (guessed == C_US_ASCII)
599                                 return C_ISO_2022_JP;
600                         p += 2;
601                 } else if (IS_ASCII(*p)) {
602                         p++;
603                 } else if (iseuckanji(*p) && iseuckanji(*(p + 1))) {
604                         if (*p >= 0xfd && *p <= 0xfe)
605                                 return C_EUC_JP;
606                         else if (guessed == C_SHIFT_JIS) {
607                                 if ((issjiskanji1(*p) &&
608                                      issjiskanji2(*(p + 1))) ||
609                                     issjishwkana(*p))
610                                         guessed = C_SHIFT_JIS;
611                                 else
612                                         guessed = C_EUC_JP;
613                         } else
614                                 guessed = C_EUC_JP;
615                         p += 2;
616                 } else if (issjiskanji1(*p) && issjiskanji2(*(p + 1))) {
617                         if (iseuchwkana1(*p) && iseuchwkana2(*(p + 1)))
618                                 guessed = C_SHIFT_JIS;
619                         else
620                                 return C_SHIFT_JIS;
621                         p += 2;
622                 } else if (issjishwkana(*p)) {
623                         guessed = C_SHIFT_JIS;
624                         p++;
625                 } else {
626                         p++;
627                 }
628         }
629
630         return guessed;
631 }
632
633 static void conv_jistodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
634 {
635         conv_jistoutf8(outbuf, outlen, inbuf);
636 }
637
638 static void conv_sjistodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
639 {
640         conv_sjistoutf8(outbuf, outlen, inbuf);
641 }
642
643 static void conv_euctodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
644 {
645         conv_euctoutf8(outbuf, outlen, inbuf);
646 }
647
648 void conv_utf8todisp(gchar *outbuf, gint outlen, const gchar *inbuf)
649 {
650         if (g_utf8_validate(inbuf, -1, NULL) == TRUE)
651                 strncpy2(outbuf, inbuf, outlen);
652         else
653                 conv_ustodisp(outbuf, outlen, inbuf);
654 }
655
656 static void conv_anytodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
657 {
658         conv_anytoutf8(outbuf, outlen, inbuf);
659         if (g_utf8_validate(outbuf, -1, NULL) != TRUE)
660                 conv_unreadable_8bit(outbuf);
661 }
662
663 static void conv_ustodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
664 {
665         strncpy2(outbuf, inbuf, outlen);
666         conv_unreadable_8bit(outbuf);
667 }
668
669 void conv_localetodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
670 {
671         gchar *tmpstr;
672
673         tmpstr = conv_iconv_strdup(inbuf, conv_get_locale_charset_str(),
674                                    CS_INTERNAL);
675         if (tmpstr && g_utf8_validate(tmpstr, -1, NULL)) {
676                 strncpy2(outbuf, tmpstr, outlen);
677                 g_free(tmpstr);
678         } else if (tmpstr && !g_utf8_validate(tmpstr, -1, NULL)) {
679                 g_free(tmpstr);
680                 conv_utf8todisp(outbuf, outlen, inbuf);
681         } else
682                 conv_utf8todisp(outbuf, outlen, inbuf);
683 }
684
685 static void conv_noconv(gchar *outbuf, gint outlen, const gchar *inbuf)
686 {
687         strncpy2(outbuf, inbuf, outlen);
688 }
689
690 static const gchar *
691 conv_get_fallback_for_private_encoding(const gchar *encoding)
692 {
693         if (encoding && (encoding[0] == 'X' || encoding[0] == 'x') &&
694             encoding[1] == '-') {
695                 if (!g_ascii_strcasecmp(encoding, CS_X_GBK))
696                         return CS_GBK;
697         }
698
699         return encoding;
700 }
701
702 CodeConverter *conv_code_converter_new(const gchar *src_charset)
703 {
704         CodeConverter *conv;
705
706         src_charset = conv_get_fallback_for_private_encoding(src_charset);
707
708         conv = g_new0(CodeConverter, 1);
709         conv->code_conv_func = conv_get_code_conv_func(src_charset, NULL);
710         conv->charset_str = g_strdup(src_charset);
711         conv->charset = conv_get_charset_from_str(src_charset);
712
713         return conv;
714 }
715
716 void conv_code_converter_destroy(CodeConverter *conv)
717 {
718         g_free(conv->charset_str);
719         g_free(conv);
720 }
721
722 gint conv_convert(CodeConverter *conv, gchar *outbuf, gint outlen,
723                   const gchar *inbuf)
724 {
725         if (conv->code_conv_func != conv_noconv)
726                 conv->code_conv_func(outbuf, outlen, inbuf);
727         else {
728                 gchar *str;
729
730                 str = conv_iconv_strdup(inbuf, conv->charset_str, NULL);
731                 if (!str)
732                         return -1;
733                 else {
734                         strncpy2(outbuf, str, outlen);
735                         g_free(str);
736                 }
737         }
738
739         return 0;
740 }
741
742 gchar *conv_codeset_strdup(const gchar *inbuf,
743                            const gchar *src_code, const gchar *dest_code)
744 {
745         gchar *buf;
746         size_t len;
747         CodeConvFunc conv_func;
748
749         src_code = conv_get_fallback_for_private_encoding(src_code);
750         conv_func = conv_get_code_conv_func(src_code, dest_code);
751         if (conv_func != conv_noconv) {
752                 len = (strlen(inbuf) + 1) * 3;
753                 buf = g_malloc(len);
754                 if (!buf) return NULL;
755
756                 conv_func(buf, len, inbuf);
757                 return g_realloc(buf, strlen(buf) + 1);
758         }
759
760         return conv_iconv_strdup(inbuf, src_code, dest_code);
761 }
762
763 CodeConvFunc conv_get_code_conv_func(const gchar *src_charset_str,
764                                      const gchar *dest_charset_str)
765 {
766         CodeConvFunc code_conv = conv_noconv;
767         CharSet src_charset;
768         CharSet dest_charset;
769
770         if (!src_charset_str)
771                 src_charset = conv_get_locale_charset();
772         else
773                 src_charset = conv_get_charset_from_str(src_charset_str);
774
775         /* auto detection mode */
776         if (!src_charset_str && !dest_charset_str) {
777                 if (conv_is_ja_locale())
778                         return conv_anytodisp;
779                 else
780                         return conv_noconv;
781         }
782
783         dest_charset = conv_get_charset_from_str(dest_charset_str);
784
785         if (dest_charset == C_US_ASCII)
786                 return conv_ustodisp;
787
788         switch (src_charset) {
789         case C_US_ASCII:
790         case C_ISO_8859_1:
791         case C_ISO_8859_2:
792         case C_ISO_8859_3:
793         case C_ISO_8859_4:
794         case C_ISO_8859_5:
795         case C_ISO_8859_6:
796         case C_ISO_8859_7:
797         case C_ISO_8859_8:
798         case C_ISO_8859_9:
799         case C_ISO_8859_10:
800         case C_ISO_8859_11:
801         case C_ISO_8859_13:
802         case C_ISO_8859_14:
803         case C_ISO_8859_15:
804                 break;
805         case C_ISO_2022_JP:
806         case C_ISO_2022_JP_2:
807         case C_ISO_2022_JP_3:
808                 if (dest_charset == C_AUTO)
809                         code_conv = conv_jistodisp;
810                 else if (dest_charset == C_EUC_JP)
811                         code_conv = conv_jistoeuc;
812                 else if (dest_charset == C_UTF_8)
813                         code_conv = conv_jistoutf8;
814                 break;
815         case C_SHIFT_JIS:
816                 if (dest_charset == C_AUTO)
817                         code_conv = conv_sjistodisp;
818                 else if (dest_charset == C_EUC_JP)
819                         code_conv = conv_sjistoeuc;
820                 else if (dest_charset == C_UTF_8)
821                         code_conv = conv_sjistoutf8;
822                 break;
823         case C_EUC_JP:
824                 if (dest_charset == C_AUTO)
825                         code_conv = conv_euctodisp;
826                 else if (dest_charset == C_ISO_2022_JP   ||
827                          dest_charset == C_ISO_2022_JP_2 ||
828                          dest_charset == C_ISO_2022_JP_3)
829                         code_conv = conv_euctojis;
830                 else if (dest_charset == C_UTF_8)
831                         code_conv = conv_euctoutf8;
832                 break;
833         case C_UTF_8:
834                 if (dest_charset == C_EUC_JP)
835                         code_conv = conv_utf8toeuc;
836                 else if (dest_charset == C_ISO_2022_JP   ||
837                          dest_charset == C_ISO_2022_JP_2 ||
838                          dest_charset == C_ISO_2022_JP_3)
839                         code_conv = conv_utf8tojis;
840                 break;
841         default:
842                 break;
843         }
844
845         return code_conv;
846 }
847
848 gchar *conv_iconv_strdup(const gchar *inbuf,
849                          const gchar *src_code, const gchar *dest_code)
850 {
851         iconv_t cd;
852         gchar *outbuf;
853
854         if (!src_code)
855                 src_code = conv_get_outgoing_charset_str();
856         if (!dest_code)
857                 dest_code = CS_INTERNAL;
858
859         /* don't convert if src and dest codeset are identical */
860         if (!strcasecmp(src_code, dest_code))
861                 return g_strdup(inbuf);
862
863         /* don't convert if current codeset is US-ASCII */
864         if (!strcasecmp(dest_code, CS_US_ASCII))
865                 return g_strdup(inbuf);
866
867         cd = iconv_open(dest_code, src_code);
868         if (cd == (iconv_t)-1)
869                 return NULL;
870
871         outbuf = conv_iconv_strdup_with_cd(inbuf, cd);
872
873         iconv_close(cd);
874
875         return outbuf;
876 }
877
878 gchar *conv_iconv_strdup_with_cd(const gchar *inbuf, iconv_t cd)
879 {
880         const gchar *inbuf_p;
881         gchar *outbuf;
882         gchar *outbuf_p;
883         size_t in_size;
884         size_t in_left;
885         size_t out_size;
886         size_t out_left;
887         size_t n_conv;
888         size_t len;
889
890         inbuf_p = inbuf;
891         in_size = strlen(inbuf);
892         in_left = in_size;
893         out_size = (in_size + 1) * 2;
894         outbuf = g_malloc(out_size);
895         outbuf_p = outbuf;
896         out_left = out_size;
897
898 #define EXPAND_BUF()                            \
899 {                                               \
900         len = outbuf_p - outbuf;                \
901         out_size *= 2;                          \
902         outbuf = g_realloc(outbuf, out_size);   \
903         outbuf_p = outbuf + len;                \
904         out_left = out_size - len;              \
905 }
906
907         while ((n_conv = iconv(cd, (ICONV_CONST gchar **)&inbuf_p, &in_left,
908                                &outbuf_p, &out_left)) == (size_t)-1) {
909                 if (EILSEQ == errno) {
910                         //g_print("iconv(): at %d: %s\n", in_size - in_left, g_strerror(errno));
911                         inbuf_p++;
912                         in_left--;
913                         if (out_left == 0) {
914                                 EXPAND_BUF();
915                         }
916                         *outbuf_p++ = SUBST_CHAR;
917                         out_left--;
918                 } else if (EINVAL == errno) {
919                         break;
920                 } else if (E2BIG == errno) {
921                         EXPAND_BUF();
922                 } else {
923                         g_warning("conv_iconv_strdup(): %s\n",
924                                   g_strerror(errno));
925                         break;
926                 }
927         }
928
929         while ((n_conv = iconv(cd, NULL, NULL, &outbuf_p, &out_left)) ==
930                (size_t)-1) {
931                 if (E2BIG == errno) {
932                         EXPAND_BUF();
933                 } else {
934                         g_warning("conv_iconv_strdup(): %s\n",
935                                   g_strerror(errno));
936                         break;
937                 }
938         }
939
940 #undef EXPAND_BUF
941
942         len = outbuf_p - outbuf;
943         outbuf = g_realloc(outbuf, len + 1);
944         outbuf[len] = '\0';
945
946         return outbuf;
947 }
948
949 static const struct {
950         CharSet charset;
951         gchar *const name;
952 } charsets[] = {
953         {C_US_ASCII,            CS_US_ASCII},
954         {C_US_ASCII,            CS_ANSI_X3_4_1968},
955         {C_UTF_8,               CS_UTF_8},
956         {C_UTF_7,               CS_UTF_7},
957         {C_ISO_8859_1,          CS_ISO_8859_1},
958         {C_ISO_8859_2,          CS_ISO_8859_2},
959         {C_ISO_8859_3,          CS_ISO_8859_3},
960         {C_ISO_8859_4,          CS_ISO_8859_4},
961         {C_ISO_8859_5,          CS_ISO_8859_5},
962         {C_ISO_8859_6,          CS_ISO_8859_6},
963         {C_ISO_8859_7,          CS_ISO_8859_7},
964         {C_ISO_8859_8,          CS_ISO_8859_8},
965         {C_ISO_8859_9,          CS_ISO_8859_9},
966         {C_ISO_8859_10,         CS_ISO_8859_10},
967         {C_ISO_8859_11,         CS_ISO_8859_11},
968         {C_ISO_8859_13,         CS_ISO_8859_13},
969         {C_ISO_8859_14,         CS_ISO_8859_14},
970         {C_ISO_8859_15,         CS_ISO_8859_15},
971         {C_BALTIC,              CS_BALTIC},
972         {C_CP1250,              CS_CP1250},
973         {C_CP1251,              CS_CP1251},
974         {C_CP1252,              CS_CP1252},
975         {C_CP1253,              CS_CP1253},
976         {C_CP1254,              CS_CP1254},
977         {C_CP1255,              CS_CP1255},
978         {C_CP1256,              CS_CP1256},
979         {C_CP1257,              CS_CP1257},
980         {C_CP1258,              CS_CP1258},
981         {C_WINDOWS_1250,        CS_WINDOWS_1250},
982         {C_WINDOWS_1251,        CS_WINDOWS_1251},
983         {C_WINDOWS_1252,        CS_WINDOWS_1252},
984         {C_WINDOWS_1253,        CS_WINDOWS_1253},
985         {C_WINDOWS_1254,        CS_WINDOWS_1254},
986         {C_WINDOWS_1255,        CS_WINDOWS_1255},
987         {C_WINDOWS_1256,        CS_WINDOWS_1256},
988         {C_WINDOWS_1257,        CS_WINDOWS_1257},
989         {C_WINDOWS_1258,        CS_WINDOWS_1258},
990         {C_KOI8_R,              CS_KOI8_R},
991         {C_KOI8_T,              CS_KOI8_T},
992         {C_KOI8_U,              CS_KOI8_U},
993         {C_ISO_2022_JP,         CS_ISO_2022_JP},
994         {C_ISO_2022_JP_2,       CS_ISO_2022_JP_2},
995         {C_ISO_2022_JP_3,       CS_ISO_2022_JP_3},
996         {C_EUC_JP,              CS_EUC_JP},
997         {C_EUC_JP,              CS_EUCJP},
998         {C_EUC_JP_MS,           CS_EUC_JP_MS},
999         {C_SHIFT_JIS,           CS_SHIFT_JIS},
1000         {C_SHIFT_JIS,           CS_SHIFT__JIS},
1001         {C_SHIFT_JIS,           CS_SJIS},
1002         {C_ISO_2022_KR,         CS_ISO_2022_KR},
1003         {C_EUC_KR,              CS_EUC_KR},
1004         {C_ISO_2022_CN,         CS_ISO_2022_CN},
1005         {C_EUC_CN,              CS_EUC_CN},
1006         {C_GB2312,              CS_GB2312},
1007         {C_GBK,                 CS_GBK},
1008         {C_EUC_TW,              CS_EUC_TW},
1009         {C_BIG5,                CS_BIG5},
1010         {C_BIG5_HKSCS,          CS_BIG5_HKSCS},
1011         {C_TIS_620,             CS_TIS_620},
1012         {C_WINDOWS_874,         CS_WINDOWS_874},
1013         {C_GEORGIAN_PS,         CS_GEORGIAN_PS},
1014         {C_TCVN5712_1,          CS_TCVN5712_1},
1015 };
1016
1017 static const struct {
1018         gchar *const locale;
1019         CharSet charset;
1020         CharSet out_charset;
1021 } locale_table[] = {
1022         {"ja_JP.eucJP"  , C_EUC_JP      , C_ISO_2022_JP},
1023         {"ja_JP.EUC-JP" , C_EUC_JP      , C_ISO_2022_JP},
1024         {"ja_JP.EUC"    , C_EUC_JP      , C_ISO_2022_JP},
1025         {"ja_JP.ujis"   , C_EUC_JP      , C_ISO_2022_JP},
1026         {"ja_JP.SJIS"   , C_SHIFT_JIS   , C_ISO_2022_JP},
1027         {"ja_JP.JIS"    , C_ISO_2022_JP , C_ISO_2022_JP},
1028 #ifdef G_OS_WIN32
1029         {"ja_JP"        , C_SHIFT_JIS   , C_ISO_2022_JP},
1030 #else
1031         {"ja_JP"        , C_EUC_JP      , C_ISO_2022_JP},
1032 #endif
1033         {"ko_KR.EUC-KR" , C_EUC_KR      , C_EUC_KR},
1034         {"ko_KR"        , C_EUC_KR      , C_EUC_KR},
1035         {"zh_CN.GB2312" , C_GB2312      , C_GB2312},
1036         {"zh_CN.GBK"    , C_GBK         , C_GBK},
1037         {"zh_CN"        , C_GB2312      , C_GB2312},
1038         {"zh_HK"        , C_BIG5_HKSCS  , C_BIG5_HKSCS},
1039         {"zh_TW.eucTW"  , C_EUC_TW      , C_BIG5},
1040         {"zh_TW.EUC-TW" , C_EUC_TW      , C_BIG5},
1041         {"zh_TW.Big5"   , C_BIG5        , C_BIG5},
1042         {"zh_TW"        , C_BIG5        , C_BIG5},
1043
1044         {"ru_RU.KOI8-R" , C_KOI8_R      , C_KOI8_R},
1045         {"ru_RU.KOI8R"  , C_KOI8_R      , C_KOI8_R},
1046         {"ru_RU.CP1251" , C_WINDOWS_1251, C_KOI8_R},
1047         {"ru_RU"        , C_ISO_8859_5  , C_KOI8_R},
1048         {"tg_TJ"        , C_KOI8_T      , C_KOI8_T},
1049         {"ru_UA"        , C_KOI8_U      , C_KOI8_U},
1050         {"uk_UA.CP1251" , C_WINDOWS_1251, C_KOI8_U},
1051         {"uk_UA"        , C_KOI8_U      , C_KOI8_U},
1052
1053         {"be_BY"        , C_WINDOWS_1251, C_WINDOWS_1251},
1054         {"bg_BG"        , C_WINDOWS_1251, C_WINDOWS_1251},
1055
1056         {"yi_US"        , C_WINDOWS_1255, C_WINDOWS_1255},
1057
1058         {"af_ZA"        , C_ISO_8859_1  , C_ISO_8859_1},
1059         {"br_FR"        , C_ISO_8859_1  , C_ISO_8859_1},
1060         {"ca_ES"        , C_ISO_8859_1  , C_ISO_8859_1},
1061         {"da_DK"        , C_ISO_8859_1  , C_ISO_8859_1},
1062         {"de_AT"        , C_ISO_8859_1  , C_ISO_8859_1},
1063         {"de_BE"        , C_ISO_8859_1  , C_ISO_8859_1},
1064         {"de_CH"        , C_ISO_8859_1  , C_ISO_8859_1},
1065         {"de_DE"        , C_ISO_8859_1  , C_ISO_8859_1},
1066         {"de_LU"        , C_ISO_8859_1  , C_ISO_8859_1},
1067         {"en_AU"        , C_ISO_8859_1  , C_ISO_8859_1},
1068         {"en_BW"        , C_ISO_8859_1  , C_ISO_8859_1},
1069         {"en_CA"        , C_ISO_8859_1  , C_ISO_8859_1},
1070         {"en_DK"        , C_ISO_8859_1  , C_ISO_8859_1},
1071         {"en_GB"        , C_ISO_8859_1  , C_ISO_8859_1},
1072         {"en_HK"        , C_ISO_8859_1  , C_ISO_8859_1},
1073         {"en_IE"        , C_ISO_8859_1  , C_ISO_8859_1},
1074         {"en_NZ"        , C_ISO_8859_1  , C_ISO_8859_1},
1075         {"en_PH"        , C_ISO_8859_1  , C_ISO_8859_1},
1076         {"en_SG"        , C_ISO_8859_1  , C_ISO_8859_1},
1077         {"en_US"        , C_ISO_8859_1  , C_ISO_8859_1},
1078         {"en_ZA"        , C_ISO_8859_1  , C_ISO_8859_1},
1079         {"en_ZW"        , C_ISO_8859_1  , C_ISO_8859_1},
1080         {"es_AR"        , C_ISO_8859_1  , C_ISO_8859_1},
1081         {"es_BO"        , C_ISO_8859_1  , C_ISO_8859_1},
1082         {"es_CL"        , C_ISO_8859_1  , C_ISO_8859_1},
1083         {"es_CO"        , C_ISO_8859_1  , C_ISO_8859_1},
1084         {"es_CR"        , C_ISO_8859_1  , C_ISO_8859_1},
1085         {"es_DO"        , C_ISO_8859_1  , C_ISO_8859_1},
1086         {"es_EC"        , C_ISO_8859_1  , C_ISO_8859_1},
1087         {"es_ES"        , C_ISO_8859_1  , C_ISO_8859_1},
1088         {"es_GT"        , C_ISO_8859_1  , C_ISO_8859_1},
1089         {"es_HN"        , C_ISO_8859_1  , C_ISO_8859_1},
1090         {"es_MX"        , C_ISO_8859_1  , C_ISO_8859_1},
1091         {"es_NI"        , C_ISO_8859_1  , C_ISO_8859_1},
1092         {"es_PA"        , C_ISO_8859_1  , C_ISO_8859_1},
1093         {"es_PE"        , C_ISO_8859_1  , C_ISO_8859_1},
1094         {"es_PR"        , C_ISO_8859_1  , C_ISO_8859_1},
1095         {"es_PY"        , C_ISO_8859_1  , C_ISO_8859_1},
1096         {"es_SV"        , C_ISO_8859_1  , C_ISO_8859_1},
1097         {"es_US"        , C_ISO_8859_1  , C_ISO_8859_1},
1098         {"es_UY"        , C_ISO_8859_1  , C_ISO_8859_1},
1099         {"es_VE"        , C_ISO_8859_1  , C_ISO_8859_1},
1100         {"et_EE"        , C_ISO_8859_1  , C_ISO_8859_1},
1101         {"eu_ES"        , C_ISO_8859_1  , C_ISO_8859_1},
1102         {"fi_FI"        , C_ISO_8859_1  , C_ISO_8859_1},
1103         {"fo_FO"        , C_ISO_8859_1  , C_ISO_8859_1},
1104         {"fr_BE"        , C_ISO_8859_1  , C_ISO_8859_1},
1105         {"fr_CA"        , C_ISO_8859_1  , C_ISO_8859_1},
1106         {"fr_CH"        , C_ISO_8859_1  , C_ISO_8859_1},
1107         {"fr_FR"        , C_ISO_8859_1  , C_ISO_8859_1},
1108         {"fr_LU"        , C_ISO_8859_1  , C_ISO_8859_1},
1109         {"ga_IE"        , C_ISO_8859_1  , C_ISO_8859_1},
1110         {"gl_ES"        , C_ISO_8859_1  , C_ISO_8859_1},
1111         {"gv_GB"        , C_ISO_8859_1  , C_ISO_8859_1},
1112         {"id_ID"        , C_ISO_8859_1  , C_ISO_8859_1},
1113         {"is_IS"        , C_ISO_8859_1  , C_ISO_8859_1},
1114         {"it_CH"        , C_ISO_8859_1  , C_ISO_8859_1},
1115         {"it_IT"        , C_ISO_8859_1  , C_ISO_8859_1},
1116         {"kl_GL"        , C_ISO_8859_1  , C_ISO_8859_1},
1117         {"kw_GB"        , C_ISO_8859_1  , C_ISO_8859_1},
1118         {"ms_MY"        , C_ISO_8859_1  , C_ISO_8859_1},
1119         {"nl_BE"        , C_ISO_8859_1  , C_ISO_8859_1},
1120         {"nl_NL"        , C_ISO_8859_1  , C_ISO_8859_1},
1121         {"nn_NO"        , C_ISO_8859_1  , C_ISO_8859_1},
1122         {"no_NO"        , C_ISO_8859_1  , C_ISO_8859_1},
1123         {"oc_FR"        , C_ISO_8859_1  , C_ISO_8859_1},
1124         {"pt_BR"        , C_ISO_8859_1  , C_ISO_8859_1},
1125         {"pt_PT"        , C_ISO_8859_1  , C_ISO_8859_1},
1126         {"sq_AL"        , C_ISO_8859_1  , C_ISO_8859_1},
1127         {"sv_FI"        , C_ISO_8859_1  , C_ISO_8859_1},
1128         {"sv_SE"        , C_ISO_8859_1  , C_ISO_8859_1},
1129         {"tl_PH"        , C_ISO_8859_1  , C_ISO_8859_1},
1130         {"uz_UZ"        , C_ISO_8859_1  , C_ISO_8859_1},
1131         {"wa_BE"        , C_ISO_8859_1  , C_ISO_8859_1},
1132
1133         {"bs_BA"        , C_ISO_8859_2  , C_ISO_8859_2},
1134         {"cs_CZ"        , C_ISO_8859_2  , C_ISO_8859_2},
1135         {"hr_HR"        , C_ISO_8859_2  , C_ISO_8859_2},
1136         {"hu_HU"        , C_ISO_8859_2  , C_ISO_8859_2},
1137         {"pl_PL"        , C_ISO_8859_2  , C_ISO_8859_2},
1138         {"ro_RO"        , C_ISO_8859_2  , C_ISO_8859_2},
1139         {"sk_SK"        , C_ISO_8859_2  , C_ISO_8859_2},
1140         {"sl_SI"        , C_ISO_8859_2  , C_ISO_8859_2},
1141
1142         {"sr_YU@cyrillic"       , C_ISO_8859_5  , C_ISO_8859_5},
1143         {"sr_YU"                , C_ISO_8859_2  , C_ISO_8859_2},
1144
1145         {"mt_MT"                , C_ISO_8859_3  , C_ISO_8859_3},
1146
1147         {"lt_LT.iso88594"       , C_ISO_8859_4  , C_ISO_8859_4},
1148         {"lt_LT.ISO8859-4"      , C_ISO_8859_4  , C_ISO_8859_4},
1149         {"lt_LT.ISO_8859-4"     , C_ISO_8859_4  , C_ISO_8859_4},
1150         {"lt_LT"                , C_ISO_8859_13 , C_ISO_8859_13},
1151
1152         {"mk_MK"        , C_ISO_8859_5  , C_ISO_8859_5},
1153
1154         {"ar_AE"        , C_ISO_8859_6  , C_ISO_8859_6},
1155         {"ar_BH"        , C_ISO_8859_6  , C_ISO_8859_6},
1156         {"ar_DZ"        , C_ISO_8859_6  , C_ISO_8859_6},
1157         {"ar_EG"        , C_ISO_8859_6  , C_ISO_8859_6},
1158         {"ar_IQ"        , C_ISO_8859_6  , C_ISO_8859_6},
1159         {"ar_JO"        , C_ISO_8859_6  , C_ISO_8859_6},
1160         {"ar_KW"        , C_ISO_8859_6  , C_ISO_8859_6},
1161         {"ar_LB"        , C_ISO_8859_6  , C_ISO_8859_6},
1162         {"ar_LY"        , C_ISO_8859_6  , C_ISO_8859_6},
1163         {"ar_MA"        , C_ISO_8859_6  , C_ISO_8859_6},
1164         {"ar_OM"        , C_ISO_8859_6  , C_ISO_8859_6},
1165         {"ar_QA"        , C_ISO_8859_6  , C_ISO_8859_6},
1166         {"ar_SA"        , C_ISO_8859_6  , C_ISO_8859_6},
1167         {"ar_SD"        , C_ISO_8859_6  , C_ISO_8859_6},
1168         {"ar_SY"        , C_ISO_8859_6  , C_ISO_8859_6},
1169         {"ar_TN"        , C_ISO_8859_6  , C_ISO_8859_6},
1170         {"ar_YE"        , C_ISO_8859_6  , C_ISO_8859_6},
1171
1172         {"el_GR"        , C_ISO_8859_7  , C_ISO_8859_7},
1173         {"he_IL"        , C_ISO_8859_8  , C_ISO_8859_8},
1174         {"iw_IL"        , C_ISO_8859_8  , C_ISO_8859_8},
1175         {"tr_TR"        , C_ISO_8859_9  , C_ISO_8859_9},
1176
1177         {"lv_LV"        , C_ISO_8859_13 , C_ISO_8859_13},
1178         {"mi_NZ"        , C_ISO_8859_13 , C_ISO_8859_13},
1179
1180         {"cy_GB"        , C_ISO_8859_14 , C_ISO_8859_14},
1181
1182         {"ar_IN"        , C_UTF_8       , C_UTF_8},
1183         {"en_IN"        , C_UTF_8       , C_UTF_8},
1184         {"se_NO"        , C_UTF_8       , C_UTF_8},
1185         {"ta_IN"        , C_UTF_8       , C_UTF_8},
1186         {"te_IN"        , C_UTF_8       , C_UTF_8},
1187         {"ur_PK"        , C_UTF_8       , C_UTF_8},
1188
1189         {"th_TH"        , C_TIS_620     , C_TIS_620},
1190         /* {"th_TH"     , C_WINDOWS_874}, */
1191         /* {"th_TH"     , C_ISO_8859_11}, */
1192
1193         {"ka_GE"        , C_GEORGIAN_PS , C_GEORGIAN_PS},
1194         {"vi_VN.TCVN"   , C_TCVN5712_1  , C_TCVN5712_1},
1195
1196         {"C"                    , C_US_ASCII    , C_US_ASCII},
1197         {"POSIX"                , C_US_ASCII    , C_US_ASCII},
1198         {"ANSI_X3.4-1968"       , C_US_ASCII    , C_US_ASCII},
1199 };
1200
1201 static GHashTable *conv_get_charset_to_str_table(void)
1202 {
1203         static GHashTable *table;
1204         gint i;
1205
1206         if (table)
1207                 return table;
1208
1209         table = g_hash_table_new(NULL, g_direct_equal);
1210
1211         for (i = 0; i < sizeof(charsets) / sizeof(charsets[0]); i++) {
1212                 if (g_hash_table_lookup(table, GUINT_TO_POINTER(charsets[i].charset))
1213                     == NULL) {
1214                         g_hash_table_insert
1215                                 (table, GUINT_TO_POINTER(charsets[i].charset),
1216                                  charsets[i].name);
1217                 }
1218         }
1219
1220         return table;
1221 }
1222
1223 static GHashTable *conv_get_charset_from_str_table(void)
1224 {
1225         static GHashTable *table;
1226         gint i;
1227
1228         if (table)
1229                 return table;
1230
1231         table = g_hash_table_new(str_case_hash, str_case_equal);
1232
1233         for (i = 0; i < sizeof(charsets) / sizeof(charsets[0]); i++) {
1234                 g_hash_table_insert(table, charsets[i].name,
1235                                     GUINT_TO_POINTER(charsets[i].charset));
1236         }
1237
1238         return table;
1239 }
1240
1241 const gchar *conv_get_charset_str(CharSet charset)
1242 {
1243         GHashTable *table;
1244
1245         table = conv_get_charset_to_str_table();
1246         return g_hash_table_lookup(table, GUINT_TO_POINTER(charset));
1247 }
1248
1249 CharSet conv_get_charset_from_str(const gchar *charset)
1250 {
1251         GHashTable *table;
1252
1253         if (!charset) return C_AUTO;
1254
1255         table = conv_get_charset_from_str_table();
1256         return GPOINTER_TO_UINT(g_hash_table_lookup(table, charset));
1257 }
1258
1259 CharSet conv_get_locale_charset(void)
1260 {
1261         static CharSet cur_charset = -1;
1262         const gchar *cur_locale;
1263         const gchar *p;
1264         gint i;
1265
1266         if (cur_charset != -1)
1267                 return cur_charset;
1268
1269         cur_locale = conv_get_current_locale();
1270         if (!cur_locale) {
1271                 cur_charset = C_US_ASCII;
1272                 return cur_charset;
1273         }
1274
1275         if (strcasestr(cur_locale, "UTF-8")) {
1276                 cur_charset = C_UTF_8;
1277                 return cur_charset;
1278         }
1279
1280         if ((p = strcasestr(cur_locale, "@euro")) && p[5] == '\0') {
1281                 cur_charset = C_ISO_8859_15;
1282                 return cur_charset;
1283         }
1284
1285         for (i = 0; i < sizeof(locale_table) / sizeof(locale_table[0]); i++) {
1286                 const gchar *p;
1287
1288                 /* "ja_JP.EUC" matches with "ja_JP.eucJP", "ja_JP.EUC" and
1289                    "ja_JP". "ja_JP" matches with "ja_JP.xxxx" and "ja" */
1290                 if (!g_ascii_strncasecmp(cur_locale, locale_table[i].locale,
1291                                  strlen(locale_table[i].locale))) {
1292                         cur_charset = locale_table[i].charset;
1293                         return cur_charset;
1294                 } else if ((p = strchr(locale_table[i].locale, '_')) &&
1295                          !strchr(p + 1, '.')) {
1296                         if (strlen(cur_locale) == 2 &&
1297                             !g_ascii_strncasecmp(cur_locale, locale_table[i].locale, 2)) {
1298                                 cur_charset = locale_table[i].charset;
1299                                 return cur_charset;
1300                         }
1301                 }
1302         }
1303
1304         cur_charset = C_AUTO;
1305         return cur_charset;
1306 }
1307
1308 const gchar *conv_get_locale_charset_str(void)
1309 {
1310         static const gchar *codeset = NULL;
1311
1312         if (!codeset)
1313                 codeset = conv_get_charset_str(conv_get_locale_charset());
1314
1315         return codeset ? codeset : CS_INTERNAL;
1316 }
1317
1318 CharSet conv_get_internal_charset(void)
1319 {
1320         return C_INTERNAL;
1321 }
1322
1323 const gchar *conv_get_internal_charset_str(void)
1324 {
1325         return CS_INTERNAL;
1326 }
1327
1328 CharSet conv_get_outgoing_charset(void)
1329 {
1330         static CharSet out_charset = -1;
1331         const gchar *cur_locale;
1332         const gchar *p;
1333         gint i;
1334
1335         if (out_charset != -1)
1336                 return out_charset;
1337
1338         cur_locale = conv_get_current_locale();
1339         if (!cur_locale) {
1340                 out_charset = C_AUTO;
1341                 return out_charset;
1342         }
1343
1344         if (strcasestr(cur_locale, "UTF-8")) {
1345                 out_charset = C_UTF_8;
1346                 return out_charset;
1347         }
1348
1349         if ((p = strcasestr(cur_locale, "@euro")) && p[5] == '\0') {
1350                 out_charset = C_ISO_8859_15;
1351                 return out_charset;
1352         }
1353
1354         for (i = 0; i < sizeof(locale_table) / sizeof(locale_table[0]); i++) {
1355                 const gchar *p;
1356
1357                 if (!g_ascii_strncasecmp(cur_locale, locale_table[i].locale,
1358                                  strlen(locale_table[i].locale))) {
1359                         out_charset = locale_table[i].out_charset;
1360                         break;
1361                 } else if ((p = strchr(locale_table[i].locale, '_')) &&
1362                          !strchr(p + 1, '.')) {
1363                         if (strlen(cur_locale) == 2 &&
1364                             !g_ascii_strncasecmp(cur_locale, locale_table[i].locale, 2)) {
1365                                 out_charset = locale_table[i].out_charset;
1366                                 break;
1367                         }
1368                 }
1369         }
1370
1371         return out_charset;
1372 }
1373
1374 const gchar *conv_get_outgoing_charset_str(void)
1375 {
1376         CharSet out_charset;
1377         const gchar *str;
1378
1379         out_charset = conv_get_outgoing_charset();
1380         str = conv_get_charset_str(out_charset);
1381
1382         return str ? str : CS_UTF_8;
1383 }
1384
1385 gboolean conv_is_multibyte_encoding(CharSet encoding)
1386 {
1387         switch (encoding) {
1388         case C_EUC_JP:
1389         case C_EUC_JP_MS:
1390         case C_EUC_KR:
1391         case C_EUC_TW:
1392         case C_EUC_CN:
1393         case C_ISO_2022_JP:
1394         case C_ISO_2022_JP_2:
1395         case C_ISO_2022_JP_3:
1396         case C_ISO_2022_KR:
1397         case C_ISO_2022_CN:
1398         case C_SHIFT_JIS:
1399         case C_GB2312:
1400         case C_GBK:
1401         case C_BIG5:
1402         case C_UTF_8:
1403         case C_UTF_7:
1404                 return TRUE;
1405         default:
1406                 return FALSE;
1407         }
1408 }
1409
1410 const gchar *conv_get_current_locale(void)
1411 {
1412         const gchar *cur_locale;
1413
1414 #ifdef G_OS_WIN32
1415         cur_locale = g_win32_getlocale();
1416 #else
1417         cur_locale = g_getenv("LC_ALL");
1418         if (!cur_locale) cur_locale = g_getenv("LC_CTYPE");
1419         if (!cur_locale) cur_locale = g_getenv("LANG");
1420         if (!cur_locale) cur_locale = setlocale(LC_CTYPE, NULL);
1421 #endif /* G_OS_WIN32 */
1422
1423         debug_print("current locale: %s\n",
1424                     cur_locale ? cur_locale : "(none)");
1425
1426         return cur_locale;
1427 }
1428
1429 gboolean conv_is_ja_locale(void)
1430 {
1431         static gint is_ja_locale = -1;
1432         const gchar *cur_locale;
1433
1434         if (is_ja_locale != -1)
1435                 return is_ja_locale != 0;
1436
1437         is_ja_locale = 0;
1438         cur_locale = conv_get_current_locale();
1439         if (cur_locale) {
1440                 if (g_ascii_strncasecmp(cur_locale, "ja", 2) == 0)
1441                         is_ja_locale = 1;
1442         }
1443
1444         return is_ja_locale != 0;
1445 }
1446
1447 gchar *conv_unmime_header(const gchar *str, const gchar *default_encoding)
1448 {
1449         gchar buf[BUFFSIZE];
1450
1451         if (is_ascii_str(str))
1452                 return unmime_header(str);
1453
1454         if (default_encoding) {
1455                 gchar *utf8_buf;
1456
1457                 utf8_buf = conv_codeset_strdup
1458                         (str, default_encoding, CS_INTERNAL);
1459                 if (utf8_buf) {
1460                         gchar *decoded_str;
1461
1462                         decoded_str = unmime_header(utf8_buf);
1463                         g_free(utf8_buf);
1464                         return decoded_str;
1465                 }
1466         }
1467
1468         if (conv_is_ja_locale())
1469                 conv_anytodisp(buf, sizeof(buf), str);
1470         else
1471                 conv_localetodisp(buf, sizeof(buf), str);
1472
1473         return unmime_header(buf);
1474 }
1475
1476 #define MAX_LINELEN             76
1477 #define MAX_HARD_LINELEN        996
1478 #define MIMESEP_BEGIN           "=?"
1479 #define MIMESEP_END             "?="
1480
1481 #define LBREAK_IF_REQUIRED(cond, is_plain_text)                         \
1482 {                                                                       \
1483         if (len - (destp - (guchar *)dest) < MAX_LINELEN + 2) {         \
1484                 *destp = '\0';                                          \
1485                 return;                                                 \
1486         }                                                               \
1487                                                                         \
1488         if ((cond) && *srcp) {                                          \
1489                 if (destp > (guchar *)dest && left < MAX_LINELEN - 1) { \
1490                         if (isspace(*(destp - 1)))                      \
1491                                 destp--;                                \
1492                         else if (is_plain_text && isspace(*srcp))       \
1493                                 srcp++;                                 \
1494                         if (*srcp) {                                    \
1495                                 *destp++ = '\n';                        \
1496                                 *destp++ = ' ';                         \
1497                                 left = MAX_LINELEN - 1;                 \
1498                         }                                               \
1499                 }                                                       \
1500         }                                                               \
1501 }
1502
1503 void conv_encode_header_full(gchar *dest, gint len, const gchar *src,
1504                         gint header_len, gboolean addr_field,
1505                         const gchar *out_encoding_)
1506 {
1507         const gchar *cur_encoding;
1508         const gchar *out_encoding;
1509         gint mimestr_len;
1510         gchar *mimesep_enc;
1511         gint left;
1512         const guchar *srcp = src;
1513         guchar *destp = dest;
1514         gboolean use_base64;
1515
1516         g_return_if_fail(g_utf8_validate(src, -1, NULL) == TRUE);
1517
1518         if (MB_CUR_MAX > 1) {
1519                 use_base64 = TRUE;
1520                 mimesep_enc = "?B?";
1521         } else {
1522                 use_base64 = FALSE;
1523                 mimesep_enc = "?Q?";
1524         }
1525
1526         cur_encoding = CS_INTERNAL;
1527
1528         if (out_encoding_)
1529                 out_encoding = out_encoding_;
1530         else
1531                 out_encoding = conv_get_outgoing_charset_str();
1532
1533         if (!strcmp(out_encoding, CS_US_ASCII))
1534                 out_encoding = CS_ISO_8859_1;
1535
1536         mimestr_len = strlen(MIMESEP_BEGIN) + strlen(out_encoding) +
1537                 strlen(mimesep_enc) + strlen(MIMESEP_END);
1538
1539         left = MAX_LINELEN - header_len;
1540
1541         while (*srcp) {
1542                 LBREAK_IF_REQUIRED(left <= 0, TRUE);
1543
1544                 while (isspace(*srcp)) {
1545                         *destp++ = *srcp++;
1546                         left--;
1547                         LBREAK_IF_REQUIRED(left <= 0, TRUE);
1548                 }
1549
1550                 /* output as it is if the next word is ASCII string */
1551                 if (!is_next_nonascii(srcp)) {
1552                         gint word_len;
1553
1554                         word_len = get_next_word_len(srcp);
1555                         LBREAK_IF_REQUIRED(left < word_len, TRUE);
1556                         while (word_len > 0) {
1557                                 LBREAK_IF_REQUIRED(left + (MAX_HARD_LINELEN - MAX_LINELEN) <= 0, TRUE)
1558                                 *destp++ = *srcp++;
1559                                 left--;
1560                                 word_len--;
1561                         }
1562
1563                         continue;
1564                 }
1565
1566                 /* don't include parentheses in encoded strings */
1567                 if (addr_field && (*srcp == '(' || *srcp == ')')) {
1568                         LBREAK_IF_REQUIRED(left < 2, FALSE);
1569                         *destp++ = *srcp++;
1570                         left--;
1571                 }
1572
1573                 while (1) {
1574                         gint mb_len = 0;
1575                         gint cur_len = 0;
1576                         gchar *part_str;
1577                         gchar *out_str;
1578                         gchar *enc_str;
1579                         const guchar *p = srcp;
1580                         gint out_str_len;
1581                         gint out_enc_str_len;
1582                         gint mime_block_len;
1583                         gboolean cont = FALSE;
1584
1585                         while (*p != '\0') {
1586                                 if (isspace(*p) && !is_next_nonascii(p + 1))
1587                                         break;
1588                                 /* don't include parentheses in encoded
1589                                    strings */
1590                                 if (addr_field && (*p == '(' || *p == ')'))
1591                                         break;
1592
1593                                 mb_len = g_utf8_skip[*p];
1594
1595                                 Xstrndup_a(part_str, srcp, cur_len + mb_len, );
1596                                 out_str = conv_codeset_strdup
1597                                         (part_str, cur_encoding, out_encoding);
1598                                 if (!out_str) {
1599                                         g_warning("conv_encode_header(): code conversion failed\n");
1600                                         conv_unreadable_8bit(part_str);
1601                                         out_str = g_strdup(part_str);
1602                                 }
1603                                 out_str_len = strlen(out_str);
1604
1605                                 if (use_base64)
1606                                         out_enc_str_len = B64LEN(out_str_len);
1607                                 else
1608                                         out_enc_str_len =
1609                                                 qp_get_q_encoding_len(out_str);
1610
1611                                 g_free(out_str);
1612
1613                                 if (mimestr_len + out_enc_str_len <= left) {
1614                                         cur_len += mb_len;
1615                                         p += mb_len;
1616                                 } else if (cur_len == 0) {
1617                                         LBREAK_IF_REQUIRED(1, FALSE);
1618                                         continue;
1619                                 } else {
1620                                         cont = TRUE;
1621                                         break;
1622                                 }
1623                         }
1624
1625                         if (cur_len > 0) {
1626                                 Xstrndup_a(part_str, srcp, cur_len, );
1627                                 out_str = conv_codeset_strdup
1628                                         (part_str, cur_encoding, out_encoding);
1629                                 if (!out_str) {
1630                                         g_warning("conv_encode_header(): code conversion failed\n");
1631                                         conv_unreadable_8bit(part_str);
1632                                         out_str = g_strdup(part_str);
1633                                 }
1634                                 out_str_len = strlen(out_str);
1635
1636                                 if (use_base64)
1637                                         out_enc_str_len = B64LEN(out_str_len);
1638                                 else
1639                                         out_enc_str_len =
1640                                                 qp_get_q_encoding_len(out_str);
1641
1642                                 Xalloca(enc_str, out_enc_str_len + 1, );
1643                                 if (use_base64)
1644                                         base64_encode(enc_str, out_str, out_str_len);
1645                                 else
1646                                         qp_q_encode(enc_str, out_str);
1647
1648                                 g_free(out_str);
1649
1650                                 /* output MIME-encoded string block */
1651                                 mime_block_len = mimestr_len + strlen(enc_str);
1652                                 g_snprintf(destp, mime_block_len + 1,
1653                                            MIMESEP_BEGIN "%s%s%s" MIMESEP_END,
1654                                            out_encoding, mimesep_enc, enc_str);
1655                                 destp += mime_block_len;
1656                                 srcp += cur_len;
1657
1658                                 left -= mime_block_len;
1659                         }
1660
1661                         LBREAK_IF_REQUIRED(cont, FALSE);
1662
1663                         if (cur_len == 0)
1664                                 break;
1665                 }
1666         }
1667
1668         *destp = '\0';
1669 }
1670
1671 void conv_encode_header(gchar *dest, gint len, const gchar *src,
1672                         gint header_len, gboolean addr_field)
1673 {
1674         conv_encode_header_full(dest,len,src,header_len,addr_field,NULL);
1675 }
1676
1677 #undef LBREAK_IF_REQUIRED
1678 gchar *conv_filename_from_utf8(const gchar *utf8_file)
1679 {
1680         gchar *fs_file;
1681         GError *error = NULL;
1682
1683         fs_file = g_filename_from_utf8(utf8_file, -1, NULL, NULL, &error);
1684         if (error) {
1685                 g_warning("failed to convert encoding of file name: %s\n",
1686                           error->message);
1687                 g_error_free(error);
1688         }
1689         if (!fs_file)
1690                 fs_file = g_strdup(utf8_file);
1691
1692         return fs_file;
1693 }
1694
1695 gchar *conv_filename_to_utf8(const gchar *fs_file)
1696 {
1697         gchar *utf8_file = NULL;
1698         GError *error = NULL;
1699
1700         utf8_file = g_filename_to_utf8(fs_file, -1, NULL, NULL, &error);
1701         if (error) {
1702                 g_warning("failed to convert encoding of file name: %s\n",
1703                           error->message);
1704                 g_error_free(error);
1705         }
1706
1707         if (!utf8_file || !g_utf8_validate(utf8_file, -1, NULL)) {
1708                 g_free(utf8_file);
1709                 utf8_file = g_strdup(fs_file);
1710                 conv_unreadable_8bit(utf8_file);
1711         }
1712
1713         return utf8_file;
1714 }