545998d63711222f217cc916647611283934e769
[claws.git] / src / codeconv.c
1 /*
2  * Sylpheed -- a GTK+ based, lightweight, and fast e-mail client
3  * Copyright (C) 1999-2007 Hiroyuki Yamamoto and the Claws Mail team
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 3 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, see <http://www.gnu.org/licenses/>.
17  * 
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 "codeconv.h"
38 #include "unmime.h"
39 #include "base64.h"
40 #include "quoted-printable.h"
41 #include "utils.h"
42 #include "prefs_common.h"
43
44 /* For unknown reasons the inconv.m4 macro undefs that macro if no
45    const is needed.  This would break the code below so we define it. */
46 #ifndef ICONV_CONST
47 #define ICONV_CONST
48 #endif
49
50 typedef enum
51 {
52         JIS_ASCII,
53         JIS_KANJI,
54         JIS_HWKANA,
55         JIS_AUXKANJI
56 } JISState;
57
58 #define SUBST_CHAR      0x5f;
59 #define ESC             '\033'
60
61 #define iseuckanji(c) \
62         (((c) & 0xff) >= 0xa1 && ((c) & 0xff) <= 0xfe)
63 #define iseuchwkana1(c) \
64         (((c) & 0xff) == 0x8e)
65 #define iseuchwkana2(c) \
66         (((c) & 0xff) >= 0xa1 && ((c) & 0xff) <= 0xdf)
67 #define iseucaux(c) \
68         (((c) & 0xff) == 0x8f)
69 #define issjiskanji1(c) \
70         ((((c) & 0xff) >= 0x81 && ((c) & 0xff) <= 0x9f) || \
71          (((c) & 0xff) >= 0xe0 && ((c) & 0xff) <= 0xfc))
72 #define issjiskanji2(c) \
73         ((((c) & 0xff) >= 0x40 && ((c) & 0xff) <= 0x7e) || \
74          (((c) & 0xff) >= 0x80 && ((c) & 0xff) <= 0xfc))
75 #define issjishwkana(c) \
76         (((c) & 0xff) >= 0xa1 && ((c) & 0xff) <= 0xdf)
77
78 #define K_IN()                          \
79         if (state != JIS_KANJI) {       \
80                 *out++ = ESC;           \
81                 *out++ = '$';           \
82                 *out++ = 'B';           \
83                 state = JIS_KANJI;      \
84         }
85
86 #define K_OUT()                         \
87         if (state != JIS_ASCII) {       \
88                 *out++ = ESC;           \
89                 *out++ = '(';           \
90                 *out++ = 'B';           \
91                 state = JIS_ASCII;      \
92         }
93
94 #define HW_IN()                         \
95         if (state != JIS_HWKANA) {      \
96                 *out++ = ESC;           \
97                 *out++ = '(';           \
98                 *out++ = 'I';           \
99                 state = JIS_HWKANA;     \
100         }
101
102 #define AUX_IN()                        \
103         if (state != JIS_AUXKANJI) {    \
104                 *out++ = ESC;           \
105                 *out++ = '$';           \
106                 *out++ = '(';           \
107                 *out++ = 'D';           \
108                 state = JIS_AUXKANJI;   \
109         }
110
111 static CodeConvFunc conv_get_code_conv_func     (const gchar    *src_charset_str,
112                                          const gchar    *dest_charset_str);
113
114 static gchar *conv_iconv_strdup_with_cd (const gchar    *inbuf,
115                                          iconv_t         cd);
116
117 static gchar *conv_iconv_strdup         (const gchar    *inbuf,
118                                          const gchar    *src_code,
119                                          const gchar    *dest_code);
120
121 static CharSet conv_get_locale_charset                  (void);
122 static CharSet conv_get_outgoing_charset                (void);
123 static CharSet conv_guess_ja_encoding(const gchar *str);
124 static gboolean conv_is_ja_locale                       (void);
125
126 static void conv_jistoeuc(gchar *outbuf, gint outlen, const gchar *inbuf);
127 static void conv_euctojis(gchar *outbuf, gint outlen, const gchar *inbuf);
128 static void conv_sjistoeuc(gchar *outbuf, gint outlen, const gchar *inbuf);
129
130 static void conv_jistoutf8(gchar *outbuf, gint outlen, const gchar *inbuf);
131 static void conv_sjistoutf8(gchar *outbuf, gint outlen, const gchar *inbuf);
132 static void conv_euctoutf8(gchar *outbuf, gint outlen, const gchar *inbuf);
133 static void conv_anytoutf8(gchar *outbuf, gint outlen, const gchar *inbuf);
134
135 static void conv_utf8toeuc(gchar *outbuf, gint outlen, const gchar *inbuf);
136 static void conv_utf8tojis(gchar *outbuf, gint outlen, const gchar *inbuf);
137
138 static void conv_unreadable_8bit(gchar *str);
139
140 static void conv_jistodisp(gchar *outbuf, gint outlen, const gchar *inbuf);
141 static void conv_sjistodisp(gchar *outbuf, gint outlen, const gchar *inbuf);
142 static void conv_euctodisp(gchar *outbuf, gint outlen, const gchar *inbuf);
143
144 static void conv_anytodisp(gchar *outbuf, gint outlen, const gchar *inbuf);
145 static void conv_ustodisp(gchar *outbuf, gint outlen, const gchar *inbuf);
146 static void conv_noconv(gchar *outbuf, gint outlen, const gchar *inbuf);
147
148 static gboolean strict_mode = FALSE;
149
150 void codeconv_set_strict(gboolean mode)
151 {
152         strict_mode = mode;
153 }
154
155 static void conv_jistoeuc(gchar *outbuf, gint outlen, const gchar *inbuf)
156 {
157         const guchar *in = inbuf;
158         guchar *out = outbuf;
159         JISState state = JIS_ASCII;
160
161         while (*in != '\0') {
162                 if (*in == ESC) {
163                         in++;
164                         if (*in == '$') {
165                                 if (*(in + 1) == '@' || *(in + 1) == 'B') {
166                                         state = JIS_KANJI;
167                                         in += 2;
168                                 } else if (*(in + 1) == '(' &&
169                                            *(in + 2) == 'D') {
170                                         state = JIS_AUXKANJI;
171                                         in += 3;
172                                 } else {
173                                         /* unknown escape sequence */
174                                         state = JIS_ASCII;
175                                 }
176                         } else if (*in == '(') {
177                                 if (*(in + 1) == 'B' || *(in + 1) == 'J') {
178                                         state = JIS_ASCII;
179                                         in += 2;
180                                 } else if (*(in + 1) == 'I') {
181                                         state = JIS_HWKANA;
182                                         in += 2;
183                                 } else {
184                                         /* unknown escape sequence */
185                                         state = JIS_ASCII;
186                                 }
187                         } else {
188                                 /* unknown escape sequence */
189                                 state = JIS_ASCII;
190                         }
191                 } else if (*in == 0x0e) {
192                         state = JIS_HWKANA;
193                         in++;
194                 } else if (*in == 0x0f) {
195                         state = JIS_ASCII;
196                         in++;
197                 } else {
198                         switch (state) {
199                         case JIS_ASCII:
200                                 *out++ = *in++;
201                                 break;
202                         case JIS_KANJI:
203                                 *out++ = *in++ | 0x80;
204                                 if (*in == '\0') break;
205                                 *out++ = *in++ | 0x80;
206                                 break;
207                         case JIS_HWKANA:
208                                 *out++ = 0x8e;
209                                 *out++ = *in++ | 0x80;
210                                 break;
211                         case JIS_AUXKANJI:
212                                 *out++ = 0x8f;
213                                 *out++ = *in++ | 0x80;
214                                 if (*in == '\0') break;
215                                 *out++ = *in++ | 0x80;
216                                 break;
217                         }
218                 }
219         }
220
221         *out = '\0';
222 }
223
224 #define JIS_HWDAKUTEN           0x5e
225 #define JIS_HWHANDAKUTEN        0x5f
226
227 static gint conv_jis_hantozen(guchar *outbuf, guchar jis_code, guchar sound_sym)
228 {
229         static guint16 h2z_tbl[] = {
230                 /* 0x20 - 0x2f */
231                 0x0000, 0x2123, 0x2156, 0x2157, 0x2122, 0x2126, 0x2572, 0x2521,
232                 0x2523, 0x2525, 0x2527, 0x2529, 0x2563, 0x2565, 0x2567, 0x2543,
233                 /* 0x30 - 0x3f */
234                 0x213c, 0x2522, 0x2524, 0x2526, 0x2528, 0x252a, 0x252b, 0x252d,
235                 0x252f, 0x2531, 0x2533, 0x2535, 0x2537, 0x2539, 0x253b, 0x253d,
236                 /* 0x40 - 0x4f */
237                 0x253f, 0x2541, 0x2544, 0x2546, 0x2548, 0x254a, 0x254b, 0x254c,
238                 0x254d, 0x254e, 0x254f, 0x2552, 0x2555, 0x2558, 0x255b, 0x255e,
239                 /* 0x50 - 0x5f */
240                 0x255f, 0x2560, 0x2561, 0x2562, 0x2564, 0x2566, 0x2568, 0x2569,
241                 0x256a, 0x256b, 0x256c, 0x256d, 0x256f, 0x2573, 0x212b, 0x212c
242         };
243
244         static guint16 dakuten_tbl[] = {
245                 /* 0x30 - 0x3f */
246                 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x252c, 0x252e,
247                 0x2530, 0x2532, 0x2534, 0x2536, 0x2538, 0x253a, 0x253c, 0x253e,
248                 /* 0x40 - 0x4f */
249                 0x2540, 0x2542, 0x2545, 0x2547, 0x2549, 0x0000, 0x0000, 0x0000,
250                 0x0000, 0x0000, 0x2550, 0x2553, 0x2556, 0x2559, 0x255c, 0x0000
251         };
252
253         static guint16 handakuten_tbl[] = {
254                 /* 0x4a - 0x4e */
255                 0x2551, 0x2554, 0x2557, 0x255a, 0x255d
256         };
257
258         guint16 out_code;
259
260         jis_code &= 0x7f;
261         sound_sym &= 0x7f;
262
263         if (jis_code < 0x21 || jis_code > 0x5f)
264                 return 0;
265
266         if (sound_sym == JIS_HWDAKUTEN &&
267             jis_code >= 0x36 && jis_code <= 0x4e) {
268                 out_code = dakuten_tbl[jis_code - 0x30];
269                 if (out_code != 0) {
270                         *outbuf = out_code >> 8;
271                         *(outbuf + 1) = out_code & 0xff;
272                         return 2;
273                 }
274         }
275
276         if (sound_sym == JIS_HWHANDAKUTEN &&
277             jis_code >= 0x4a && jis_code <= 0x4e) {
278                 out_code = handakuten_tbl[jis_code - 0x4a];
279                 *outbuf = out_code >> 8;
280                 *(outbuf + 1) = out_code & 0xff;
281                 return 2;
282         }
283
284         out_code = h2z_tbl[jis_code - 0x20];
285         *outbuf = out_code >> 8;
286         *(outbuf + 1) = out_code & 0xff;
287         return 1;
288 }
289
290 static void conv_euctojis(gchar *outbuf, gint outlen, const gchar *inbuf)
291 {
292         const guchar *in = inbuf;
293         guchar *out = outbuf;
294         JISState state = JIS_ASCII;
295
296         while (*in != '\0') {
297                 if (IS_ASCII(*in)) {
298                         K_OUT();
299                         *out++ = *in++;
300                 } else if (iseuckanji(*in)) {
301                         if (iseuckanji(*(in + 1))) {
302                                 K_IN();
303                                 *out++ = *in++ & 0x7f;
304                                 *out++ = *in++ & 0x7f;
305                         } else {
306                                 K_OUT();
307                                 *out++ = SUBST_CHAR;
308                                 in++;
309                                 if (*in != '\0' && !IS_ASCII(*in)) {
310                                         *out++ = SUBST_CHAR;
311                                         in++;
312                                 }
313                         }
314                 } else if (iseuchwkana1(*in)) {
315                         if (iseuchwkana2(*(in + 1))) {
316                                 if (prefs_common.allow_jisx0201_kana) {
317                                         HW_IN();
318                                         in++;
319                                         *out++ = *in++ & 0x7f;
320                                 } else {
321                                         guchar jis_ch[2];
322                                         gint len;
323
324                                         if (iseuchwkana1(*(in + 2)) &&
325                                             iseuchwkana2(*(in + 3)))
326                                                 len = conv_jis_hantozen
327                                                         (jis_ch,
328                                                          *(in + 1), *(in + 3));
329                                         else
330                                                 len = conv_jis_hantozen
331                                                         (jis_ch,
332                                                          *(in + 1), '\0');
333                                         if (len == 0)
334                                                 in += 2;
335                                         else {
336                                                 K_IN();
337                                                 in += len * 2;
338                                                 *out++ = jis_ch[0];
339                                                 *out++ = jis_ch[1];
340                                         }
341                                 }
342                         } else {
343                                 K_OUT();
344                                 in++;
345                                 if (*in != '\0' && !IS_ASCII(*in)) {
346                                         *out++ = SUBST_CHAR;
347                                         in++;
348                                 }
349                         }
350                 } else if (iseucaux(*in)) {
351                         in++;
352                         if (iseuckanji(*in) && iseuckanji(*(in + 1))) {
353                                 AUX_IN();
354                                 *out++ = *in++ & 0x7f;
355                                 *out++ = *in++ & 0x7f;
356                         } else {
357                                 K_OUT();
358                                 if (*in != '\0' && !IS_ASCII(*in)) {
359                                         *out++ = SUBST_CHAR;
360                                         in++;
361                                         if (*in != '\0' && !IS_ASCII(*in)) {
362                                                 *out++ = SUBST_CHAR;
363                                                 in++;
364                                         }
365                                 }
366                         }
367                 } else {
368                         K_OUT();
369                         *out++ = SUBST_CHAR;
370                         in++;
371                 }
372         }
373
374         K_OUT();
375         *out = '\0';
376 }
377
378 static void conv_sjistoeuc(gchar *outbuf, gint outlen, const gchar *inbuf)
379 {
380         const guchar *in = inbuf;
381         guchar *out = outbuf;
382
383         while (*in != '\0') {
384                 if (IS_ASCII(*in)) {
385                         *out++ = *in++;
386                 } else if (issjiskanji1(*in)) {
387                         if (issjiskanji2(*(in + 1))) {
388                                 guchar out1 = *in;
389                                 guchar out2 = *(in + 1);
390                                 guchar row;
391
392                                 row = out1 < 0xa0 ? 0x70 : 0xb0;
393                                 if (out2 < 0x9f) {
394                                         out1 = (out1 - row) * 2 - 1;
395                                         out2 -= out2 > 0x7f ? 0x20 : 0x1f;
396                                 } else {
397                                         out1 = (out1 - row) * 2;
398                                         out2 -= 0x7e;
399                                 }
400
401                                 *out++ = out1 | 0x80;
402                                 *out++ = out2 | 0x80;
403                                 in += 2;
404                         } else {
405                                 *out++ = SUBST_CHAR;
406                                 in++;
407                                 if (*in != '\0' && !IS_ASCII(*in)) {
408                                         *out++ = SUBST_CHAR;
409                                         in++;
410                                 }
411                         }
412                 } else if (issjishwkana(*in)) {
413                         *out++ = 0x8e;
414                         *out++ = *in++;
415                 } else {
416                         *out++ = SUBST_CHAR;
417                         in++;
418                 }
419         }
420
421         *out = '\0';
422 }
423
424 static void conv_jistoutf8(gchar *outbuf, gint outlen, const gchar *inbuf)
425 {
426         gchar *eucstr;
427
428         Xalloca(eucstr, outlen, return);
429
430         conv_jistoeuc(eucstr, outlen, inbuf);
431         conv_euctoutf8(outbuf, outlen, eucstr);
432 }
433
434 static void conv_sjistoutf8(gchar *outbuf, gint outlen, const gchar *inbuf)
435 {
436         gchar *tmpstr;
437
438         tmpstr = conv_iconv_strdup(inbuf, CS_SHIFT_JIS, CS_UTF_8);
439         if (tmpstr) {
440                 strncpy2(outbuf, tmpstr, outlen);
441                 g_free(tmpstr);
442         } else
443                 strncpy2(outbuf, inbuf, outlen);
444 }
445
446 static void conv_euctoutf8(gchar *outbuf, gint outlen, const gchar *inbuf)
447 {
448         static iconv_t cd = (iconv_t)-1;
449         static gboolean iconv_ok = TRUE;
450         gchar *tmpstr;
451
452         if (cd == (iconv_t)-1) {
453                 if (!iconv_ok) {
454                         strncpy2(outbuf, inbuf, outlen);
455                         return;
456                 }
457                 cd = iconv_open(CS_UTF_8, CS_EUC_JP_MS);
458                 if (cd == (iconv_t)-1) {
459                         cd = iconv_open(CS_UTF_8, CS_EUC_JP);
460                         if (cd == (iconv_t)-1) {
461                                 g_warning("conv_euctoutf8(): %s\n",
462                                           g_strerror(errno));
463                                 iconv_ok = FALSE;
464                                 strncpy2(outbuf, inbuf, outlen);
465                                 return;
466                         }
467                 }
468         }
469
470         tmpstr = conv_iconv_strdup_with_cd(inbuf, cd);
471         if (tmpstr) {
472                 strncpy2(outbuf, tmpstr, outlen);
473                 g_free(tmpstr);
474         } else
475                 strncpy2(outbuf, inbuf, outlen);
476 }
477
478 static void conv_anytoutf8(gchar *outbuf, gint outlen, const gchar *inbuf)
479 {
480         switch (conv_guess_ja_encoding(inbuf)) {
481         case C_ISO_2022_JP:
482                 conv_jistoutf8(outbuf, outlen, inbuf);
483                 break;
484         case C_SHIFT_JIS:
485                 conv_sjistoutf8(outbuf, outlen, inbuf);
486                 break;
487         case C_EUC_JP:
488                 conv_euctoutf8(outbuf, outlen, inbuf);
489                 break;
490         default:
491                 strncpy2(outbuf, inbuf, outlen);
492                 break;
493         }
494 }
495
496 static void conv_utf8toeuc(gchar *outbuf, gint outlen, const gchar *inbuf)
497 {
498         static iconv_t cd = (iconv_t)-1;
499         static gboolean iconv_ok = TRUE;
500         gchar *tmpstr;
501
502         if (cd == (iconv_t)-1) {
503                 if (!iconv_ok) {
504                         strncpy2(outbuf, inbuf, outlen);
505                         return;
506                 }
507                 cd = iconv_open(CS_EUC_JP_MS, CS_UTF_8);
508                 if (cd == (iconv_t)-1) {
509                         cd = iconv_open(CS_EUC_JP, CS_UTF_8);
510                         if (cd == (iconv_t)-1) {
511                                 g_warning("conv_utf8toeuc(): %s\n",
512                                           g_strerror(errno));
513                                 iconv_ok = FALSE;
514                                 strncpy2(outbuf, inbuf, outlen);
515                                 return;
516                         }
517                 }
518         }
519
520         tmpstr = conv_iconv_strdup_with_cd(inbuf, cd);
521         if (tmpstr) {
522                 strncpy2(outbuf, tmpstr, outlen);
523                 g_free(tmpstr);
524         } else
525                 strncpy2(outbuf, inbuf, outlen);
526 }
527
528 static void conv_utf8tojis(gchar *outbuf, gint outlen, const gchar *inbuf)
529 {
530         gchar *eucstr;
531
532         Xalloca(eucstr, outlen, return);
533
534         conv_utf8toeuc(eucstr, outlen, inbuf);
535         conv_euctojis(outbuf, outlen, eucstr);
536 }
537
538 static void conv_unreadable_8bit(gchar *str)
539 {
540         register guchar *p = str;
541
542         while (*p != '\0') {
543                 /* convert CR+LF -> LF */
544                 if (*p == '\r' && *(p + 1) == '\n')
545                         memmove(p, p + 1, strlen(p));
546                 else if (!IS_ASCII(*p)) *p = SUBST_CHAR;
547                 p++;
548         }
549 }
550
551 static CharSet conv_guess_ja_encoding(const gchar *str)
552 {
553         const guchar *p = str;
554         CharSet guessed = C_US_ASCII;
555
556         while (*p != '\0') {
557                 if (*p == ESC && (*(p + 1) == '$' || *(p + 1) == '(')) {
558                         if (guessed == C_US_ASCII)
559                                 return C_ISO_2022_JP;
560                         p += 2;
561                 } else if (IS_ASCII(*p)) {
562                         p++;
563                 } else if (iseuckanji(*p) && iseuckanji(*(p + 1))) {
564                         if (*p >= 0xfd && *p <= 0xfe)
565                                 return C_EUC_JP;
566                         else if (guessed == C_SHIFT_JIS) {
567                                 if ((issjiskanji1(*p) &&
568                                      issjiskanji2(*(p + 1))) ||
569                                     issjishwkana(*p))
570                                         guessed = C_SHIFT_JIS;
571                                 else
572                                         guessed = C_EUC_JP;
573                         } else
574                                 guessed = C_EUC_JP;
575                         p += 2;
576                 } else if (issjiskanji1(*p) && issjiskanji2(*(p + 1))) {
577                         if (iseuchwkana1(*p) && iseuchwkana2(*(p + 1)))
578                                 guessed = C_SHIFT_JIS;
579                         else
580                                 return C_SHIFT_JIS;
581                         p += 2;
582                 } else if (issjishwkana(*p)) {
583                         guessed = C_SHIFT_JIS;
584                         p++;
585                 } else {
586                         p++;
587                 }
588         }
589
590         return guessed;
591 }
592
593 static void conv_jistodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
594 {
595         conv_jistoutf8(outbuf, outlen, inbuf);
596 }
597
598 static void conv_sjistodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
599 {
600         conv_sjistoutf8(outbuf, outlen, inbuf);
601 }
602
603 static void conv_euctodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
604 {
605         conv_euctoutf8(outbuf, outlen, inbuf);
606 }
607
608 void conv_utf8todisp(gchar *outbuf, gint outlen, const gchar *inbuf)
609 {
610         if (g_utf8_validate(inbuf, -1, NULL) == TRUE)
611                 strncpy2(outbuf, inbuf, outlen);
612         else
613                 conv_ustodisp(outbuf, outlen, inbuf);
614 }
615
616 static void conv_anytodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
617 {
618         conv_anytoutf8(outbuf, outlen, inbuf);
619         if (g_utf8_validate(outbuf, -1, NULL) != TRUE)
620                 conv_unreadable_8bit(outbuf);
621 }
622
623 static void conv_ustodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
624 {
625         strncpy2(outbuf, inbuf, outlen);
626         conv_unreadable_8bit(outbuf);
627 }
628
629 void conv_localetodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
630 {
631         gchar *tmpstr;
632
633         codeconv_set_strict(TRUE);
634         tmpstr = conv_iconv_strdup(inbuf, conv_get_locale_charset_str(),
635                                    CS_INTERNAL);
636         codeconv_set_strict(FALSE);
637         if (tmpstr && g_utf8_validate(tmpstr, -1, NULL)) {
638                 strncpy2(outbuf, tmpstr, outlen);
639                 g_free(tmpstr);
640                 return;
641         } else if (tmpstr && !g_utf8_validate(tmpstr, -1, NULL)) {
642                 g_free(tmpstr);
643                 codeconv_set_strict(TRUE);
644                 tmpstr = conv_iconv_strdup(inbuf, 
645                                 conv_get_locale_charset_str_no_utf8(),
646                                 CS_INTERNAL);
647                 codeconv_set_strict(FALSE);
648         }
649         if (tmpstr && g_utf8_validate(tmpstr, -1, NULL)) {
650                 strncpy2(outbuf, tmpstr, outlen);
651                 g_free(tmpstr);
652                 return;
653         } else {
654                 g_free(tmpstr);
655                 conv_utf8todisp(outbuf, outlen, inbuf);
656         }
657 }
658
659 static void conv_noconv(gchar *outbuf, gint outlen, const gchar *inbuf)
660 {
661         strncpy2(outbuf, inbuf, outlen);
662 }
663
664 static const gchar *
665 conv_get_fallback_for_private_encoding(const gchar *encoding)
666 {
667         if (encoding && (encoding[0] == 'X' || encoding[0] == 'x') &&
668             encoding[1] == '-') {
669                 if (!g_ascii_strcasecmp(encoding, CS_X_GBK))
670                         return CS_GBK;
671         }
672
673         return encoding;
674 }
675
676 CodeConverter *conv_code_converter_new(const gchar *src_charset)
677 {
678         CodeConverter *conv;
679
680         src_charset = conv_get_fallback_for_private_encoding(src_charset);
681
682         conv = g_new0(CodeConverter, 1);
683         conv->code_conv_func = conv_get_code_conv_func(src_charset, NULL);
684         conv->charset_str = g_strdup(src_charset);
685         conv->charset = conv_get_charset_from_str(src_charset);
686
687         return conv;
688 }
689
690 void conv_code_converter_destroy(CodeConverter *conv)
691 {
692         g_free(conv->charset_str);
693         g_free(conv);
694 }
695
696 gint conv_convert(CodeConverter *conv, gchar *outbuf, gint outlen,
697                   const gchar *inbuf)
698 {
699         if (conv->code_conv_func != conv_noconv)
700                 conv->code_conv_func(outbuf, outlen, inbuf);
701         else {
702                 gchar *str;
703
704                 str = conv_iconv_strdup(inbuf, conv->charset_str, NULL);
705                 if (!str)
706                         return -1;
707                 else {
708                         strncpy2(outbuf, str, outlen);
709                         g_free(str);
710                 }
711         }
712
713         return 0;
714 }
715
716 gchar *conv_codeset_strdup(const gchar *inbuf,
717                            const gchar *src_code, const gchar *dest_code)
718 {
719         gchar *buf;
720         size_t len;
721         CodeConvFunc conv_func;
722
723         if (!strcmp2(src_code, dest_code))
724                 return g_strdup(inbuf);
725
726         src_code = conv_get_fallback_for_private_encoding(src_code);
727         conv_func = conv_get_code_conv_func(src_code, dest_code);
728         if (conv_func != conv_noconv) {
729                 len = (strlen(inbuf) + 1) * 3;
730                 buf = g_malloc(len);
731                 if (!buf) return NULL;
732
733                 conv_func(buf, len, inbuf);
734                 return g_realloc(buf, strlen(buf) + 1);
735         }
736
737         return conv_iconv_strdup(inbuf, src_code, dest_code);
738 }
739
740 static CodeConvFunc conv_get_code_conv_func(const gchar *src_charset_str,
741                                      const gchar *dest_charset_str)
742 {
743         CodeConvFunc code_conv = conv_noconv;
744         CharSet src_charset;
745         CharSet dest_charset;
746
747         if (!src_charset_str)
748                 src_charset = conv_get_locale_charset();
749         else
750                 src_charset = conv_get_charset_from_str(src_charset_str);
751
752         /* auto detection mode */
753         if (!src_charset_str && !dest_charset_str) {
754                 if (conv_is_ja_locale())
755                         return conv_anytodisp;
756                 else
757                         return conv_noconv;
758         }
759
760         dest_charset = conv_get_charset_from_str(dest_charset_str);
761
762         if (dest_charset == C_US_ASCII)
763                 return conv_ustodisp;
764
765         switch (src_charset) {
766         case C_US_ASCII:
767         case C_ISO_8859_1:
768         case C_ISO_8859_2:
769         case C_ISO_8859_3:
770         case C_ISO_8859_4:
771         case C_ISO_8859_5:
772         case C_ISO_8859_6:
773         case C_ISO_8859_7:
774         case C_ISO_8859_8:
775         case C_ISO_8859_9:
776         case C_ISO_8859_10:
777         case C_ISO_8859_11:
778         case C_ISO_8859_13:
779         case C_ISO_8859_14:
780         case C_ISO_8859_15:
781                 break;
782         case C_ISO_2022_JP:
783         case C_ISO_2022_JP_2:
784         case C_ISO_2022_JP_3:
785                 if (dest_charset == C_AUTO)
786                         code_conv = conv_jistodisp;
787                 else if (dest_charset == C_EUC_JP)
788                         code_conv = conv_jistoeuc;
789                 else if (dest_charset == C_UTF_8)
790                         code_conv = conv_jistoutf8;
791                 break;
792         case C_SHIFT_JIS:
793                 if (dest_charset == C_AUTO)
794                         code_conv = conv_sjistodisp;
795                 else if (dest_charset == C_EUC_JP)
796                         code_conv = conv_sjistoeuc;
797                 else if (dest_charset == C_UTF_8)
798                         code_conv = conv_sjistoutf8;
799                 break;
800         case C_EUC_JP:
801                 if (dest_charset == C_AUTO)
802                         code_conv = conv_euctodisp;
803                 else if (dest_charset == C_ISO_2022_JP   ||
804                          dest_charset == C_ISO_2022_JP_2 ||
805                          dest_charset == C_ISO_2022_JP_3)
806                         code_conv = conv_euctojis;
807                 else if (dest_charset == C_UTF_8)
808                         code_conv = conv_euctoutf8;
809                 break;
810         case C_UTF_8:
811                 if (dest_charset == C_EUC_JP)
812                         code_conv = conv_utf8toeuc;
813                 else if (dest_charset == C_ISO_2022_JP   ||
814                          dest_charset == C_ISO_2022_JP_2 ||
815                          dest_charset == C_ISO_2022_JP_3)
816                         code_conv = conv_utf8tojis;
817                 break;
818         default:
819                 break;
820         }
821
822         return code_conv;
823 }
824
825 static gchar *conv_iconv_strdup(const gchar *inbuf,
826                          const gchar *src_code, const gchar *dest_code)
827 {
828         iconv_t cd;
829         gchar *outbuf;
830
831         if (!src_code && !dest_code && 
832             g_utf8_validate(inbuf, -1, NULL))
833                 return g_strdup(inbuf);
834
835         if (!src_code)
836                 src_code = conv_get_outgoing_charset_str();
837         if (!dest_code)
838                 dest_code = CS_INTERNAL;
839
840         /* don't convert if src and dest codeset are identical */
841         if (!strcasecmp(src_code, dest_code))
842                 return g_strdup(inbuf);
843
844         /* don't convert if dest codeset is US-ASCII */
845         if (!strcasecmp(src_code, CS_US_ASCII))
846                 return g_strdup(inbuf);
847
848         /* don't convert if dest codeset is US-ASCII */
849         if (!strcasecmp(dest_code, CS_US_ASCII))
850                 return g_strdup(inbuf);
851
852         cd = iconv_open(dest_code, src_code);
853         if (cd == (iconv_t)-1)
854                 return NULL;
855
856         outbuf = conv_iconv_strdup_with_cd(inbuf, cd);
857
858         iconv_close(cd);
859
860         return outbuf;
861 }
862
863 gchar *conv_iconv_strdup_with_cd(const gchar *inbuf, iconv_t cd)
864 {
865         const gchar *inbuf_p;
866         gchar *outbuf;
867         gchar *outbuf_p;
868         size_t in_size;
869         size_t in_left;
870         size_t out_size;
871         size_t out_left;
872         size_t n_conv;
873         size_t len;
874
875         inbuf_p = inbuf;
876         in_size = strlen(inbuf);
877         in_left = in_size;
878         out_size = (in_size + 1) * 2;
879         outbuf = g_malloc(out_size);
880         outbuf_p = outbuf;
881         out_left = out_size;
882
883 #define EXPAND_BUF()                            \
884 {                                               \
885         len = outbuf_p - outbuf;                \
886         out_size *= 2;                          \
887         outbuf = g_realloc(outbuf, out_size);   \
888         outbuf_p = outbuf + len;                \
889         out_left = out_size - len;              \
890 }
891
892         while ((n_conv = iconv(cd, (ICONV_CONST gchar **)&inbuf_p, &in_left,
893                                &outbuf_p, &out_left)) == (size_t)-1) {
894                 if (EILSEQ == errno) {
895                         if (strict_mode) {
896                                 g_free(outbuf);
897                                 return NULL;
898                         }
899                         //g_print("iconv(): at %d: %s\n", in_size - in_left, g_strerror(errno));
900                         inbuf_p++;
901                         in_left--;
902                         if (out_left == 0) {
903                                 EXPAND_BUF();
904                         }
905                         *outbuf_p++ = SUBST_CHAR;
906                         out_left--;
907                 } else if (EINVAL == errno) {
908                         break;
909                 } else if (E2BIG == errno) {
910                         EXPAND_BUF();
911                 } else {
912                         g_warning("conv_iconv_strdup(): %s\n",
913                                   g_strerror(errno));
914                         break;
915                 }
916         }
917
918         while ((n_conv = iconv(cd, NULL, NULL, &outbuf_p, &out_left)) ==
919                (size_t)-1) {
920                 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 #undef EXPAND_BUF
930
931         len = outbuf_p - outbuf;
932         outbuf = g_realloc(outbuf, len + 1);
933         outbuf[len] = '\0';
934
935         return outbuf;
936 }
937
938 static const struct {
939         CharSet charset;
940         gchar *const name;
941 } charsets[] = {
942         {C_US_ASCII,            CS_US_ASCII},
943         {C_US_ASCII,            CS_ANSI_X3_4_1968},
944         {C_UTF_8,               CS_UTF_8},
945         {C_UTF_7,               CS_UTF_7},
946         {C_ISO_8859_1,          CS_ISO_8859_1},
947         {C_ISO_8859_2,          CS_ISO_8859_2},
948         {C_ISO_8859_3,          CS_ISO_8859_3},
949         {C_ISO_8859_4,          CS_ISO_8859_4},
950         {C_ISO_8859_5,          CS_ISO_8859_5},
951         {C_ISO_8859_6,          CS_ISO_8859_6},
952         {C_ISO_8859_7,          CS_ISO_8859_7},
953         {C_ISO_8859_8,          CS_ISO_8859_8},
954         {C_ISO_8859_9,          CS_ISO_8859_9},
955         {C_ISO_8859_10,         CS_ISO_8859_10},
956         {C_ISO_8859_11,         CS_ISO_8859_11},
957         {C_ISO_8859_13,         CS_ISO_8859_13},
958         {C_ISO_8859_14,         CS_ISO_8859_14},
959         {C_ISO_8859_15,         CS_ISO_8859_15},
960         {C_BALTIC,              CS_BALTIC},
961         {C_CP1250,              CS_CP1250},
962         {C_CP1251,              CS_CP1251},
963         {C_CP1252,              CS_CP1252},
964         {C_CP1253,              CS_CP1253},
965         {C_CP1254,              CS_CP1254},
966         {C_CP1255,              CS_CP1255},
967         {C_CP1256,              CS_CP1256},
968         {C_CP1257,              CS_CP1257},
969         {C_CP1258,              CS_CP1258},
970         {C_WINDOWS_1250,        CS_WINDOWS_1250},
971         {C_WINDOWS_1251,        CS_WINDOWS_1251},
972         {C_WINDOWS_1252,        CS_WINDOWS_1252},
973         {C_WINDOWS_1253,        CS_WINDOWS_1253},
974         {C_WINDOWS_1254,        CS_WINDOWS_1254},
975         {C_WINDOWS_1255,        CS_WINDOWS_1255},
976         {C_WINDOWS_1256,        CS_WINDOWS_1256},
977         {C_WINDOWS_1257,        CS_WINDOWS_1257},
978         {C_WINDOWS_1258,        CS_WINDOWS_1258},
979         {C_KOI8_R,              CS_KOI8_R},
980         {C_KOI8_T,              CS_KOI8_T},
981         {C_KOI8_U,              CS_KOI8_U},
982         {C_ISO_2022_JP,         CS_ISO_2022_JP},
983         {C_ISO_2022_JP_2,       CS_ISO_2022_JP_2},
984         {C_ISO_2022_JP_3,       CS_ISO_2022_JP_3},
985         {C_EUC_JP,              CS_EUC_JP},
986         {C_EUC_JP,              CS_EUCJP},
987         {C_EUC_JP_MS,           CS_EUC_JP_MS},
988         {C_SHIFT_JIS,           CS_SHIFT_JIS},
989         {C_SHIFT_JIS,           CS_SHIFT__JIS},
990         {C_SHIFT_JIS,           CS_SJIS},
991         {C_ISO_2022_KR,         CS_ISO_2022_KR},
992         {C_EUC_KR,              CS_EUC_KR},
993         {C_ISO_2022_CN,         CS_ISO_2022_CN},
994         {C_EUC_CN,              CS_EUC_CN},
995         {C_GB2312,              CS_GB2312},
996         {C_GBK,                 CS_GBK},
997         {C_EUC_TW,              CS_EUC_TW},
998         {C_BIG5,                CS_BIG5},
999         {C_BIG5_HKSCS,          CS_BIG5_HKSCS},
1000         {C_TIS_620,             CS_TIS_620},
1001         {C_WINDOWS_874,         CS_WINDOWS_874},
1002         {C_GEORGIAN_PS,         CS_GEORGIAN_PS},
1003         {C_TCVN5712_1,          CS_TCVN5712_1},
1004 };
1005
1006 static const struct {
1007         gchar *const locale;
1008         CharSet charset;
1009         CharSet out_charset;
1010 } locale_table[] = {
1011         {"ja_JP.eucJP"  , C_EUC_JP      , C_ISO_2022_JP},
1012         {"ja_JP.EUC-JP" , C_EUC_JP      , C_ISO_2022_JP},
1013         {"ja_JP.EUC"    , C_EUC_JP      , C_ISO_2022_JP},
1014         {"ja_JP.ujis"   , C_EUC_JP      , C_ISO_2022_JP},
1015         {"ja_JP.SJIS"   , C_SHIFT_JIS   , C_ISO_2022_JP},
1016         {"ja_JP.JIS"    , C_ISO_2022_JP , C_ISO_2022_JP},
1017 #ifdef G_OS_WIN32
1018         {"ja_JP"        , C_SHIFT_JIS   , C_ISO_2022_JP},
1019 #else
1020         {"ja_JP"        , C_EUC_JP      , C_ISO_2022_JP},
1021 #endif
1022         {"ko_KR.EUC-KR" , C_EUC_KR      , C_EUC_KR},
1023         {"ko_KR"        , C_EUC_KR      , C_EUC_KR},
1024         {"zh_CN.GB2312" , C_GB2312      , C_GB2312},
1025         {"zh_CN.GBK"    , C_GBK         , C_GBK},
1026         {"zh_CN"        , C_GB2312      , C_GB2312},
1027         {"zh_HK"        , C_BIG5_HKSCS  , C_BIG5_HKSCS},
1028         {"zh_TW.eucTW"  , C_EUC_TW      , C_BIG5},
1029         {"zh_TW.EUC-TW" , C_EUC_TW      , C_BIG5},
1030         {"zh_TW.Big5"   , C_BIG5        , C_BIG5},
1031         {"zh_TW"        , C_BIG5        , C_BIG5},
1032
1033         {"ru_RU.KOI8-R" , C_KOI8_R      , C_KOI8_R},
1034         {"ru_RU.KOI8R"  , C_KOI8_R      , C_KOI8_R},
1035         {"ru_RU.CP1251" , C_WINDOWS_1251, C_KOI8_R},
1036         {"ru_RU"        , C_ISO_8859_5  , C_KOI8_R},
1037         {"tg_TJ"        , C_KOI8_T      , C_KOI8_T},
1038         {"ru_UA"        , C_KOI8_U      , C_KOI8_U},
1039         {"uk_UA.CP1251" , C_WINDOWS_1251, C_KOI8_U},
1040         {"uk_UA"        , C_KOI8_U      , C_KOI8_U},
1041
1042         {"be_BY"        , C_WINDOWS_1251, C_WINDOWS_1251},
1043         {"bg_BG"        , C_WINDOWS_1251, C_WINDOWS_1251},
1044
1045         {"yi_US"        , C_WINDOWS_1255, C_WINDOWS_1255},
1046
1047         {"af_ZA"        , C_ISO_8859_1  , C_ISO_8859_1},
1048         {"br_FR"        , C_ISO_8859_1  , C_ISO_8859_1},
1049         {"ca_ES"        , C_ISO_8859_1  , C_ISO_8859_1},
1050         {"da_DK"        , C_ISO_8859_1  , C_ISO_8859_1},
1051         {"de_AT"        , C_ISO_8859_1  , C_ISO_8859_1},
1052         {"de_BE"        , C_ISO_8859_1  , C_ISO_8859_1},
1053         {"de_CH"        , C_ISO_8859_1  , C_ISO_8859_1},
1054         {"de_DE"        , C_ISO_8859_1  , C_ISO_8859_1},
1055         {"de_LU"        , C_ISO_8859_1  , C_ISO_8859_1},
1056         {"en_AU"        , C_ISO_8859_1  , C_ISO_8859_1},
1057         {"en_BW"        , C_ISO_8859_1  , C_ISO_8859_1},
1058         {"en_CA"        , C_ISO_8859_1  , C_ISO_8859_1},
1059         {"en_DK"        , C_ISO_8859_1  , C_ISO_8859_1},
1060         {"en_GB"        , C_ISO_8859_1  , C_ISO_8859_1},
1061         {"en_HK"        , C_ISO_8859_1  , C_ISO_8859_1},
1062         {"en_IE"        , C_ISO_8859_1  , C_ISO_8859_1},
1063         {"en_NZ"        , C_ISO_8859_1  , C_ISO_8859_1},
1064         {"en_PH"        , C_ISO_8859_1  , C_ISO_8859_1},
1065         {"en_SG"        , C_ISO_8859_1  , C_ISO_8859_1},
1066         {"en_US"        , C_ISO_8859_1  , C_ISO_8859_1},
1067         {"en_ZA"        , C_ISO_8859_1  , C_ISO_8859_1},
1068         {"en_ZW"        , C_ISO_8859_1  , C_ISO_8859_1},
1069         {"es_AR"        , C_ISO_8859_1  , C_ISO_8859_1},
1070         {"es_BO"        , C_ISO_8859_1  , C_ISO_8859_1},
1071         {"es_CL"        , C_ISO_8859_1  , C_ISO_8859_1},
1072         {"es_CO"        , C_ISO_8859_1  , C_ISO_8859_1},
1073         {"es_CR"        , C_ISO_8859_1  , C_ISO_8859_1},
1074         {"es_DO"        , C_ISO_8859_1  , C_ISO_8859_1},
1075         {"es_EC"        , C_ISO_8859_1  , C_ISO_8859_1},
1076         {"es_ES"        , C_ISO_8859_1  , C_ISO_8859_1},
1077         {"es_GT"        , C_ISO_8859_1  , C_ISO_8859_1},
1078         {"es_HN"        , C_ISO_8859_1  , C_ISO_8859_1},
1079         {"es_MX"        , C_ISO_8859_1  , C_ISO_8859_1},
1080         {"es_NI"        , C_ISO_8859_1  , C_ISO_8859_1},
1081         {"es_PA"        , C_ISO_8859_1  , C_ISO_8859_1},
1082         {"es_PE"        , C_ISO_8859_1  , C_ISO_8859_1},
1083         {"es_PR"        , C_ISO_8859_1  , C_ISO_8859_1},
1084         {"es_PY"        , C_ISO_8859_1  , C_ISO_8859_1},
1085         {"es_SV"        , C_ISO_8859_1  , C_ISO_8859_1},
1086         {"es_US"        , C_ISO_8859_1  , C_ISO_8859_1},
1087         {"es_UY"        , C_ISO_8859_1  , C_ISO_8859_1},
1088         {"es_VE"        , C_ISO_8859_1  , C_ISO_8859_1},
1089         {"et_EE"        , C_ISO_8859_1  , C_ISO_8859_1},
1090         {"eu_ES"        , C_ISO_8859_1  , C_ISO_8859_1},
1091         {"fi_FI"        , C_ISO_8859_1  , C_ISO_8859_1},
1092         {"fo_FO"        , C_ISO_8859_1  , C_ISO_8859_1},
1093         {"fr_BE"        , C_ISO_8859_1  , C_ISO_8859_1},
1094         {"fr_CA"        , C_ISO_8859_1  , C_ISO_8859_1},
1095         {"fr_CH"        , C_ISO_8859_1  , C_ISO_8859_1},
1096         {"fr_FR"        , C_ISO_8859_1  , C_ISO_8859_1},
1097         {"fr_LU"        , C_ISO_8859_1  , C_ISO_8859_1},
1098         {"ga_IE"        , C_ISO_8859_1  , C_ISO_8859_1},
1099         {"gl_ES"        , C_ISO_8859_1  , C_ISO_8859_1},
1100         {"gv_GB"        , C_ISO_8859_1  , C_ISO_8859_1},
1101         {"id_ID"        , C_ISO_8859_1  , C_ISO_8859_1},
1102         {"is_IS"        , C_ISO_8859_1  , C_ISO_8859_1},
1103         {"it_CH"        , C_ISO_8859_1  , C_ISO_8859_1},
1104         {"it_IT"        , C_ISO_8859_1  , C_ISO_8859_1},
1105         {"kl_GL"        , C_ISO_8859_1  , C_ISO_8859_1},
1106         {"kw_GB"        , C_ISO_8859_1  , C_ISO_8859_1},
1107         {"ms_MY"        , C_ISO_8859_1  , C_ISO_8859_1},
1108         {"nl_BE"        , C_ISO_8859_1  , C_ISO_8859_1},
1109         {"nl_NL"        , C_ISO_8859_1  , C_ISO_8859_1},
1110         {"nn_NO"        , C_ISO_8859_1  , C_ISO_8859_1},
1111         {"no_NO"        , C_ISO_8859_1  , C_ISO_8859_1},
1112         {"oc_FR"        , C_ISO_8859_1  , C_ISO_8859_1},
1113         {"pt_BR"        , C_ISO_8859_1  , C_ISO_8859_1},
1114         {"pt_PT"        , C_ISO_8859_1  , C_ISO_8859_1},
1115         {"sq_AL"        , C_ISO_8859_1  , C_ISO_8859_1},
1116         {"sv_FI"        , C_ISO_8859_1  , C_ISO_8859_1},
1117         {"sv_SE"        , C_ISO_8859_1  , C_ISO_8859_1},
1118         {"tl_PH"        , C_ISO_8859_1  , C_ISO_8859_1},
1119         {"uz_UZ"        , C_ISO_8859_1  , C_ISO_8859_1},
1120         {"wa_BE"        , C_ISO_8859_1  , C_ISO_8859_1},
1121
1122         {"bs_BA"        , C_ISO_8859_2  , C_ISO_8859_2},
1123         {"cs_CZ"        , C_ISO_8859_2  , C_ISO_8859_2},
1124         {"hr_HR"        , C_ISO_8859_2  , C_ISO_8859_2},
1125         {"hu_HU"        , C_ISO_8859_2  , C_ISO_8859_2},
1126         {"pl_PL"        , C_ISO_8859_2  , C_ISO_8859_2},
1127         {"ro_RO"        , C_ISO_8859_2  , C_ISO_8859_2},
1128         {"sk_SK"        , C_ISO_8859_2  , C_ISO_8859_2},
1129         {"sl_SI"        , C_ISO_8859_2  , C_ISO_8859_2},
1130
1131         {"sr_YU@cyrillic"       , C_ISO_8859_5  , C_ISO_8859_5},
1132         {"sr_YU"                , C_ISO_8859_2  , C_ISO_8859_2},
1133
1134         {"mt_MT"                , C_ISO_8859_3  , C_ISO_8859_3},
1135
1136         {"lt_LT.iso88594"       , C_ISO_8859_4  , C_ISO_8859_4},
1137         {"lt_LT.ISO8859-4"      , C_ISO_8859_4  , C_ISO_8859_4},
1138         {"lt_LT.ISO_8859-4"     , C_ISO_8859_4  , C_ISO_8859_4},
1139         {"lt_LT"                , C_ISO_8859_13 , C_ISO_8859_13},
1140
1141         {"mk_MK"        , C_ISO_8859_5  , C_ISO_8859_5},
1142
1143         {"ar_AE"        , C_ISO_8859_6  , C_ISO_8859_6},
1144         {"ar_BH"        , C_ISO_8859_6  , C_ISO_8859_6},
1145         {"ar_DZ"        , C_ISO_8859_6  , C_ISO_8859_6},
1146         {"ar_EG"        , C_ISO_8859_6  , C_ISO_8859_6},
1147         {"ar_IQ"        , C_ISO_8859_6  , C_ISO_8859_6},
1148         {"ar_JO"        , C_ISO_8859_6  , C_ISO_8859_6},
1149         {"ar_KW"        , C_ISO_8859_6  , C_ISO_8859_6},
1150         {"ar_LB"        , C_ISO_8859_6  , C_ISO_8859_6},
1151         {"ar_LY"        , C_ISO_8859_6  , C_ISO_8859_6},
1152         {"ar_MA"        , C_ISO_8859_6  , C_ISO_8859_6},
1153         {"ar_OM"        , C_ISO_8859_6  , C_ISO_8859_6},
1154         {"ar_QA"        , C_ISO_8859_6  , C_ISO_8859_6},
1155         {"ar_SA"        , C_ISO_8859_6  , C_ISO_8859_6},
1156         {"ar_SD"        , C_ISO_8859_6  , C_ISO_8859_6},
1157         {"ar_SY"        , C_ISO_8859_6  , C_ISO_8859_6},
1158         {"ar_TN"        , C_ISO_8859_6  , C_ISO_8859_6},
1159         {"ar_YE"        , C_ISO_8859_6  , C_ISO_8859_6},
1160
1161         {"el_GR"        , C_ISO_8859_7  , C_ISO_8859_7},
1162         {"he_IL"        , C_ISO_8859_8  , C_ISO_8859_8},
1163         {"iw_IL"        , C_ISO_8859_8  , C_ISO_8859_8},
1164         {"tr_TR"        , C_ISO_8859_9  , C_ISO_8859_9},
1165
1166         {"lv_LV"        , C_ISO_8859_13 , C_ISO_8859_13},
1167         {"mi_NZ"        , C_ISO_8859_13 , C_ISO_8859_13},
1168
1169         {"cy_GB"        , C_ISO_8859_14 , C_ISO_8859_14},
1170
1171         {"ar_IN"        , C_UTF_8       , C_UTF_8},
1172         {"en_IN"        , C_UTF_8       , C_UTF_8},
1173         {"se_NO"        , C_UTF_8       , C_UTF_8},
1174         {"ta_IN"        , C_UTF_8       , C_UTF_8},
1175         {"te_IN"        , C_UTF_8       , C_UTF_8},
1176         {"ur_PK"        , C_UTF_8       , C_UTF_8},
1177
1178         {"th_TH"        , C_TIS_620     , C_TIS_620},
1179         /* {"th_TH"     , C_WINDOWS_874}, */
1180         /* {"th_TH"     , C_ISO_8859_11}, */
1181
1182         {"ka_GE"        , C_GEORGIAN_PS , C_GEORGIAN_PS},
1183         {"vi_VN.TCVN"   , C_TCVN5712_1  , C_TCVN5712_1},
1184
1185         {"C"                    , C_US_ASCII    , C_US_ASCII},
1186         {"POSIX"                , C_US_ASCII    , C_US_ASCII},
1187         {"ANSI_X3.4-1968"       , C_US_ASCII    , C_US_ASCII},
1188 };
1189
1190 static GHashTable *conv_get_charset_to_str_table(void)
1191 {
1192         static GHashTable *table;
1193         gint i;
1194
1195         if (table)
1196                 return table;
1197
1198         table = g_hash_table_new(NULL, g_direct_equal);
1199
1200         for (i = 0; i < sizeof(charsets) / sizeof(charsets[0]); i++) {
1201                 if (g_hash_table_lookup(table, GUINT_TO_POINTER(charsets[i].charset))
1202                     == NULL) {
1203                         g_hash_table_insert
1204                                 (table, GUINT_TO_POINTER(charsets[i].charset),
1205                                  charsets[i].name);
1206                 }
1207         }
1208
1209         return table;
1210 }
1211
1212 static GHashTable *conv_get_charset_from_str_table(void)
1213 {
1214         static GHashTable *table;
1215         gint i;
1216
1217         if (table)
1218                 return table;
1219
1220         table = g_hash_table_new(str_case_hash, str_case_equal);
1221
1222         for (i = 0; i < sizeof(charsets) / sizeof(charsets[0]); i++) {
1223                 g_hash_table_insert(table, charsets[i].name,
1224                                     GUINT_TO_POINTER(charsets[i].charset));
1225         }
1226
1227         return table;
1228 }
1229
1230 const gchar *conv_get_charset_str(CharSet charset)
1231 {
1232         GHashTable *table;
1233
1234         table = conv_get_charset_to_str_table();
1235         return g_hash_table_lookup(table, GUINT_TO_POINTER(charset));
1236 }
1237
1238 CharSet conv_get_charset_from_str(const gchar *charset)
1239 {
1240         GHashTable *table;
1241
1242         if (!charset) return C_AUTO;
1243
1244         table = conv_get_charset_from_str_table();
1245         return GPOINTER_TO_UINT(g_hash_table_lookup(table, charset));
1246 }
1247
1248 static CharSet conv_get_locale_charset(void)
1249 {
1250         static CharSet cur_charset = -1;
1251         const gchar *cur_locale;
1252         const gchar *p;
1253         gint i;
1254
1255         if (cur_charset != -1)
1256                 return cur_charset;
1257
1258         cur_locale = conv_get_current_locale();
1259         if (!cur_locale) {
1260                 cur_charset = C_US_ASCII;
1261                 return cur_charset;
1262         }
1263
1264         if (strcasestr(cur_locale, ".UTF-8") ||
1265             strcasestr(cur_locale, ".utf8")) {
1266                 cur_charset = C_UTF_8;
1267                 return cur_charset;
1268         }
1269
1270         if ((p = strcasestr(cur_locale, "@euro")) && p[5] == '\0') {
1271                 cur_charset = C_ISO_8859_15;
1272                 return cur_charset;
1273         }
1274
1275         for (i = 0; i < sizeof(locale_table) / sizeof(locale_table[0]); i++) {
1276                 const gchar *p;
1277
1278                 /* "ja_JP.EUC" matches with "ja_JP.eucJP", "ja_JP.EUC" and
1279                    "ja_JP". "ja_JP" matches with "ja_JP.xxxx" and "ja" */
1280                 if (!g_ascii_strncasecmp(cur_locale, locale_table[i].locale,
1281                                  strlen(locale_table[i].locale))) {
1282                         cur_charset = locale_table[i].charset;
1283                         return cur_charset;
1284                 } else if ((p = strchr(locale_table[i].locale, '_')) &&
1285                          !strchr(p + 1, '.')) {
1286                         if (strlen(cur_locale) == 2 &&
1287                             !g_ascii_strncasecmp(cur_locale, locale_table[i].locale, 2)) {
1288                                 cur_charset = locale_table[i].charset;
1289                                 return cur_charset;
1290                         }
1291                 }
1292         }
1293
1294         cur_charset = C_AUTO;
1295         return cur_charset;
1296 }
1297
1298 static CharSet conv_get_locale_charset_no_utf8(void)
1299 {
1300         static CharSet cur_charset = -1;
1301         const gchar *cur_locale;
1302         const gchar *p;
1303         gchar *tmp;
1304         gint i;
1305
1306         if (prefs_common.broken_are_utf8)
1307                 return conv_get_locale_charset();
1308
1309         if (cur_charset != -1)
1310                 return cur_charset;
1311
1312         cur_locale = conv_get_current_locale();
1313         if (!cur_locale) {
1314                 cur_charset = C_US_ASCII;
1315                 return cur_charset;
1316         }
1317
1318         if (strcasestr(cur_locale, "UTF-8")) {
1319                 tmp = g_strdup(cur_locale);
1320                 *(strcasestr(tmp, ".UTF-8")) = '\0';
1321                 cur_locale = tmp;
1322         }
1323
1324         if ((p = strcasestr(cur_locale, "@euro")) && p[5] == '\0') {
1325                 cur_charset = C_ISO_8859_15;
1326                 return cur_charset;
1327         }
1328
1329         for (i = 0; i < sizeof(locale_table) / sizeof(locale_table[0]); i++) {
1330                 const gchar *p;
1331
1332                 /* "ja_JP.EUC" matches with "ja_JP.eucJP", "ja_JP.EUC" and
1333                    "ja_JP". "ja_JP" matches with "ja_JP.xxxx" and "ja" */
1334                 if (!g_ascii_strncasecmp(cur_locale, locale_table[i].locale,
1335                                  strlen(locale_table[i].locale))) {
1336                         cur_charset = locale_table[i].charset;
1337                         return cur_charset;
1338                 } else if ((p = strchr(locale_table[i].locale, '_')) &&
1339                          !strchr(p + 1, '.')) {
1340                         if (strlen(cur_locale) == 2 &&
1341                             !g_ascii_strncasecmp(cur_locale, locale_table[i].locale, 2)) {
1342                                 cur_charset = locale_table[i].charset;
1343                                 return cur_charset;
1344                         }
1345                 }
1346         }
1347
1348         cur_charset = C_AUTO;
1349         return cur_charset;
1350 }
1351
1352 const gchar *conv_get_locale_charset_str(void)
1353 {
1354         static const gchar *codeset = NULL;
1355
1356         if (!codeset)
1357                 codeset = conv_get_charset_str(conv_get_locale_charset());
1358
1359         return codeset ? codeset : CS_INTERNAL;
1360 }
1361
1362 const gchar *conv_get_locale_charset_str_no_utf8(void)
1363 {
1364         static const gchar *codeset = NULL;
1365
1366         if (!codeset)
1367                 codeset = conv_get_charset_str(conv_get_locale_charset_no_utf8());
1368
1369         return codeset ? codeset : CS_INTERNAL;
1370 }
1371
1372 static CharSet conv_get_outgoing_charset(void)
1373 {
1374         static CharSet out_charset = -1;
1375         const gchar *cur_locale;
1376         const gchar *p;
1377         gint i;
1378
1379         if (out_charset != -1)
1380                 return out_charset;
1381
1382         cur_locale = conv_get_current_locale();
1383         if (!cur_locale) {
1384                 out_charset = C_AUTO;
1385                 return out_charset;
1386         }
1387
1388         if (strcasestr(cur_locale, "UTF-8")) {
1389                 out_charset = C_UTF_8;
1390                 return out_charset;
1391         }
1392
1393         if ((p = strcasestr(cur_locale, "@euro")) && p[5] == '\0') {
1394                 out_charset = C_ISO_8859_15;
1395                 return out_charset;
1396         }
1397
1398         for (i = 0; i < sizeof(locale_table) / sizeof(locale_table[0]); i++) {
1399                 const gchar *p;
1400
1401                 if (!g_ascii_strncasecmp(cur_locale, locale_table[i].locale,
1402                                  strlen(locale_table[i].locale))) {
1403                         out_charset = locale_table[i].out_charset;
1404                         break;
1405                 } else if ((p = strchr(locale_table[i].locale, '_')) &&
1406                          !strchr(p + 1, '.')) {
1407                         if (strlen(cur_locale) == 2 &&
1408                             !g_ascii_strncasecmp(cur_locale, locale_table[i].locale, 2)) {
1409                                 out_charset = locale_table[i].out_charset;
1410                                 break;
1411                         }
1412                 }
1413         }
1414
1415         return out_charset;
1416 }
1417
1418 const gchar *conv_get_outgoing_charset_str(void)
1419 {
1420         CharSet out_charset;
1421         const gchar *str;
1422
1423         out_charset = conv_get_outgoing_charset();
1424         str = conv_get_charset_str(out_charset);
1425
1426         return str ? str : CS_UTF_8;
1427 }
1428
1429 const gchar *conv_get_current_locale(void)
1430 {
1431         const gchar *cur_locale;
1432
1433 #ifdef G_OS_WIN32
1434         cur_locale = g_win32_getlocale();
1435 #else
1436         cur_locale = g_getenv("LC_ALL");
1437         if (!cur_locale) cur_locale = g_getenv("LC_CTYPE");
1438         if (!cur_locale) cur_locale = g_getenv("LANG");
1439         if (!cur_locale) cur_locale = setlocale(LC_CTYPE, NULL);
1440 #endif /* G_OS_WIN32 */
1441
1442         debug_print("current locale: %s\n",
1443                     cur_locale ? cur_locale : "(none)");
1444
1445         return cur_locale;
1446 }
1447
1448 static gboolean conv_is_ja_locale(void)
1449 {
1450         static gint is_ja_locale = -1;
1451         const gchar *cur_locale;
1452
1453         if (is_ja_locale != -1)
1454                 return is_ja_locale != 0;
1455
1456         is_ja_locale = 0;
1457         cur_locale = conv_get_current_locale();
1458         if (cur_locale) {
1459                 if (g_ascii_strncasecmp(cur_locale, "ja", 2) == 0)
1460                         is_ja_locale = 1;
1461         }
1462
1463         return is_ja_locale != 0;
1464 }
1465
1466 gchar *conv_unmime_header(const gchar *str, const gchar *default_encoding)
1467 {
1468         gchar buf[BUFFSIZE];
1469
1470         if (is_ascii_str(str))
1471                 return unmime_header(str);
1472
1473         if (default_encoding) {
1474                 gchar *utf8_buf;
1475
1476                 utf8_buf = conv_codeset_strdup
1477                         (str, default_encoding, CS_INTERNAL);
1478                 if (utf8_buf) {
1479                         gchar *decoded_str;
1480
1481                         decoded_str = unmime_header(utf8_buf);
1482                         g_free(utf8_buf);
1483                         return decoded_str;
1484                 }
1485         }
1486
1487         if (conv_is_ja_locale())
1488                 conv_anytodisp(buf, sizeof(buf), str);
1489         else
1490                 conv_localetodisp(buf, sizeof(buf), str);
1491
1492         return unmime_header(buf);
1493 }
1494
1495 #define MAX_LINELEN             76
1496 #define MAX_HARD_LINELEN        996
1497 #define MIMESEP_BEGIN           "=?"
1498 #define MIMESEP_END             "?="
1499
1500 #define LBREAK_IF_REQUIRED(cond, is_plain_text)                         \
1501 {                                                                       \
1502         if (len - (destp - (guchar *)dest) < MAX_LINELEN + 2) {         \
1503                 *destp = '\0';                                          \
1504                 return;                                                 \
1505         }                                                               \
1506                                                                         \
1507         if ((cond) && *srcp) {                                          \
1508                 if (destp > (guchar *)dest && left < MAX_LINELEN - 1) { \
1509                         if (isspace(*(destp - 1)))                      \
1510                                 destp--;                                \
1511                         else if (is_plain_text && isspace(*srcp))       \
1512                                 srcp++;                                 \
1513                         if (*srcp) {                                    \
1514                                 *destp++ = '\n';                        \
1515                                 *destp++ = ' ';                         \
1516                                 left = MAX_LINELEN - 1;                 \
1517                         }                                               \
1518                 }                                                       \
1519         }                                                               \
1520 }
1521
1522 void conv_encode_header_full(gchar *dest, gint len, const gchar *src,
1523                         gint header_len, gboolean addr_field,
1524                         const gchar *out_encoding_)
1525 {
1526         const gchar *cur_encoding;
1527         const gchar *out_encoding;
1528         gint mimestr_len;
1529         gchar *mimesep_enc;
1530         gint left;
1531         const guchar *srcp = src;
1532         guchar *destp = dest;
1533         gboolean use_base64;
1534
1535         g_return_if_fail(g_utf8_validate(src, -1, NULL) == TRUE);
1536
1537         if (MB_CUR_MAX > 1) {
1538                 use_base64 = TRUE;
1539                 mimesep_enc = "?B?";
1540         } else {
1541                 use_base64 = FALSE;
1542                 mimesep_enc = "?Q?";
1543         }
1544
1545         cur_encoding = CS_INTERNAL;
1546
1547         if (out_encoding_)
1548                 out_encoding = out_encoding_;
1549         else
1550                 out_encoding = conv_get_outgoing_charset_str();
1551
1552         if (!strcmp(out_encoding, CS_US_ASCII))
1553                 out_encoding = CS_ISO_8859_1;
1554
1555         mimestr_len = strlen(MIMESEP_BEGIN) + strlen(out_encoding) +
1556                 strlen(mimesep_enc) + strlen(MIMESEP_END);
1557
1558         left = MAX_LINELEN - header_len;
1559
1560         while (*srcp) {
1561                 LBREAK_IF_REQUIRED(left <= 0, TRUE);
1562
1563                 while (isspace(*srcp)) {
1564                         *destp++ = *srcp++;
1565                         left--;
1566                         LBREAK_IF_REQUIRED(left <= 0, TRUE);
1567                 }
1568
1569                 /* output as it is if the next word is ASCII string */
1570                 if (!is_next_nonascii(srcp)) {
1571                         gint word_len;
1572
1573                         word_len = get_next_word_len(srcp);
1574                         LBREAK_IF_REQUIRED(left < word_len, TRUE);
1575                         while (word_len > 0) {
1576                                 LBREAK_IF_REQUIRED(left + (MAX_HARD_LINELEN - MAX_LINELEN) <= 0, TRUE)
1577                                 *destp++ = *srcp++;
1578                                 left--;
1579                                 word_len--;
1580                         }
1581
1582                         continue;
1583                 }
1584
1585                 /* don't include parentheses and quotes in encoded strings */
1586                 if (addr_field && (*srcp == '(' || *srcp == ')' || *srcp == '"')) {
1587                         LBREAK_IF_REQUIRED(left < 2, FALSE);
1588                         *destp++ = *srcp++;
1589                         left--;
1590                 }
1591
1592                 while (1) {
1593                         gint mb_len = 0;
1594                         gint cur_len = 0;
1595                         gchar *part_str;
1596                         gchar *out_str;
1597                         gchar *enc_str;
1598                         const guchar *p = srcp;
1599                         gint out_str_len;
1600                         gint out_enc_str_len;
1601                         gint mime_block_len;
1602                         gboolean cont = FALSE;
1603
1604                         while (*p != '\0') {
1605                                 if (isspace(*p) && !is_next_nonascii(p + 1))
1606                                         break;
1607                                 /* don't include parentheses in encoded
1608                                    strings */
1609                                 if (addr_field && (*p == '(' || *p == ')' || *p == '"'))
1610                                         break;
1611
1612                                 mb_len = g_utf8_skip[*p];
1613
1614                                 Xstrndup_a(part_str, srcp, cur_len + mb_len, );
1615                                 out_str = conv_codeset_strdup
1616                                         (part_str, cur_encoding, out_encoding);
1617                                 if (!out_str) {
1618                                         if (strict_mode) {
1619                                                 *dest = '\0';
1620                                                 return;
1621                                         } else {
1622                                                 g_warning("conv_encode_header(): code conversion failed\n");
1623                                                 conv_unreadable_8bit(part_str);
1624                                                 out_str = g_strdup(part_str);
1625                                         }
1626                                 }
1627                                 out_str_len = strlen(out_str);
1628
1629                                 if (use_base64)
1630                                         out_enc_str_len = B64LEN(out_str_len);
1631                                 else
1632                                         out_enc_str_len =
1633                                                 qp_get_q_encoding_len(out_str);
1634
1635                                 g_free(out_str);
1636
1637                                 if (mimestr_len + out_enc_str_len <= left) {
1638                                         cur_len += mb_len;
1639                                         p += mb_len;
1640                                 } else if (cur_len == 0) {
1641                                         LBREAK_IF_REQUIRED(1, FALSE);
1642                                         continue;
1643                                 } else {
1644                                         cont = TRUE;
1645                                         break;
1646                                 }
1647                         }
1648
1649                         if (cur_len > 0) {
1650                                 Xstrndup_a(part_str, srcp, cur_len, );
1651                                 out_str = conv_codeset_strdup
1652                                         (part_str, cur_encoding, out_encoding);
1653                                 if (!out_str) {
1654                                         g_warning("conv_encode_header(): code conversion failed\n");
1655                                         conv_unreadable_8bit(part_str);
1656                                         out_str = g_strdup(part_str);
1657                                 }
1658                                 out_str_len = strlen(out_str);
1659
1660                                 if (use_base64)
1661                                         out_enc_str_len = B64LEN(out_str_len);
1662                                 else
1663                                         out_enc_str_len =
1664                                                 qp_get_q_encoding_len(out_str);
1665
1666                                 Xalloca(enc_str, out_enc_str_len + 1, );
1667                                 if (use_base64)
1668                                         base64_encode(enc_str, out_str, out_str_len);
1669                                 else
1670                                         qp_q_encode(enc_str, out_str);
1671
1672                                 g_free(out_str);
1673
1674                                 /* output MIME-encoded string block */
1675                                 mime_block_len = mimestr_len + strlen(enc_str);
1676                                 g_snprintf(destp, mime_block_len + 1,
1677                                            MIMESEP_BEGIN "%s%s%s" MIMESEP_END,
1678                                            out_encoding, mimesep_enc, enc_str);
1679                                 destp += mime_block_len;
1680                                 srcp += cur_len;
1681
1682                                 left -= mime_block_len;
1683                         }
1684
1685                         LBREAK_IF_REQUIRED(cont, FALSE);
1686
1687                         if (cur_len == 0)
1688                                 break;
1689                 }
1690         }
1691
1692         *destp = '\0';
1693 }
1694
1695 void conv_encode_header(gchar *dest, gint len, const gchar *src,
1696                         gint header_len, gboolean addr_field)
1697 {
1698         conv_encode_header_full(dest,len,src,header_len,addr_field,NULL);
1699 }
1700
1701 #undef LBREAK_IF_REQUIRED
1702 gchar *conv_filename_from_utf8(const gchar *utf8_file)
1703 {
1704         gchar *fs_file;
1705         GError *error = NULL;
1706
1707         fs_file = g_filename_from_utf8(utf8_file, -1, NULL, NULL, &error);
1708         if (error) {
1709                 g_warning("failed to convert encoding of file name: %s\n",
1710                           error->message);
1711                 g_error_free(error);
1712         }
1713         if (!fs_file)
1714                 fs_file = g_strdup(utf8_file);
1715
1716         return fs_file;
1717 }
1718
1719 gchar *conv_filename_to_utf8(const gchar *fs_file)
1720 {
1721         gchar *utf8_file = NULL;
1722         GError *error = NULL;
1723
1724         utf8_file = g_filename_to_utf8(fs_file, -1, NULL, NULL, &error);
1725         if (error) {
1726                 g_warning("failed to convert encoding of file name: %s\n",
1727                           error->message);
1728                 g_error_free(error);
1729         }
1730
1731         if (!utf8_file || !g_utf8_validate(utf8_file, -1, NULL)) {
1732                 g_free(utf8_file);
1733                 utf8_file = g_strdup(fs_file);
1734                 conv_unreadable_8bit(utf8_file);
1735         }
1736
1737         return utf8_file;
1738 }