2006-04-06 [mones] 2.1.0cvs7
[claws.git] / src / codeconv.c
1 /*
2  * Sylpheed -- a GTK+ based, lightweight, and fast e-mail client
3  * Copyright (C) 1999-2006 Hiroyuki Yamamoto and the Sylpheed-Claws 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 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., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, 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 "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 void conv_jistoeuc(gchar *outbuf, gint outlen, const gchar *inbuf);
112 static void conv_euctojis(gchar *outbuf, gint outlen, const gchar *inbuf);
113 static void conv_sjistoeuc(gchar *outbuf, gint outlen, const gchar *inbuf);
114
115 static void conv_jistoutf8(gchar *outbuf, gint outlen, const gchar *inbuf);
116 static void conv_sjistoutf8(gchar *outbuf, gint outlen, const gchar *inbuf);
117 static void conv_euctoutf8(gchar *outbuf, gint outlen, const gchar *inbuf);
118 static void conv_anytoutf8(gchar *outbuf, gint outlen, const gchar *inbuf);
119
120 static void conv_utf8toeuc(gchar *outbuf, gint outlen, const gchar *inbuf);
121 static void conv_utf8tojis(gchar *outbuf, gint outlen, const gchar *inbuf);
122
123 static void conv_unreadable_8bit(gchar *str);
124
125 static void conv_jistodisp(gchar *outbuf, gint outlen, const gchar *inbuf);
126 static void conv_sjistodisp(gchar *outbuf, gint outlen, const gchar *inbuf);
127 static void conv_euctodisp(gchar *outbuf, gint outlen, const gchar *inbuf);
128
129 static void conv_anytodisp(gchar *outbuf, gint outlen, const gchar *inbuf);
130 static void conv_ustodisp(gchar *outbuf, gint outlen, const gchar *inbuf);
131 static void conv_noconv(gchar *outbuf, gint outlen, const gchar *inbuf);
132
133 static gboolean strict_mode = FALSE;
134
135 void codeconv_set_strict(gboolean mode)
136 {
137         strict_mode = mode;
138 }
139
140 static void conv_jistoeuc(gchar *outbuf, gint outlen, const gchar *inbuf)
141 {
142         const guchar *in = inbuf;
143         guchar *out = outbuf;
144         JISState state = JIS_ASCII;
145
146         while (*in != '\0') {
147                 if (*in == ESC) {
148                         in++;
149                         if (*in == '$') {
150                                 if (*(in + 1) == '@' || *(in + 1) == 'B') {
151                                         state = JIS_KANJI;
152                                         in += 2;
153                                 } else if (*(in + 1) == '(' &&
154                                            *(in + 2) == 'D') {
155                                         state = JIS_AUXKANJI;
156                                         in += 3;
157                                 } else {
158                                         /* unknown escape sequence */
159                                         state = JIS_ASCII;
160                                 }
161                         } else if (*in == '(') {
162                                 if (*(in + 1) == 'B' || *(in + 1) == 'J') {
163                                         state = JIS_ASCII;
164                                         in += 2;
165                                 } else if (*(in + 1) == 'I') {
166                                         state = JIS_HWKANA;
167                                         in += 2;
168                                 } else {
169                                         /* unknown escape sequence */
170                                         state = JIS_ASCII;
171                                 }
172                         } else {
173                                 /* unknown escape sequence */
174                                 state = JIS_ASCII;
175                         }
176                 } else if (*in == 0x0e) {
177                         state = JIS_HWKANA;
178                         in++;
179                 } else if (*in == 0x0f) {
180                         state = JIS_ASCII;
181                         in++;
182                 } else {
183                         switch (state) {
184                         case JIS_ASCII:
185                                 *out++ = *in++;
186                                 break;
187                         case JIS_KANJI:
188                                 *out++ = *in++ | 0x80;
189                                 if (*in == '\0') break;
190                                 *out++ = *in++ | 0x80;
191                                 break;
192                         case JIS_HWKANA:
193                                 *out++ = 0x8e;
194                                 *out++ = *in++ | 0x80;
195                                 break;
196                         case JIS_AUXKANJI:
197                                 *out++ = 0x8f;
198                                 *out++ = *in++ | 0x80;
199                                 if (*in == '\0') break;
200                                 *out++ = *in++ | 0x80;
201                                 break;
202                         }
203                 }
204         }
205
206         *out = '\0';
207 }
208
209 #define JIS_HWDAKUTEN           0x5e
210 #define JIS_HWHANDAKUTEN        0x5f
211
212 static gint conv_jis_hantozen(guchar *outbuf, guchar jis_code, guchar sound_sym)
213 {
214         static guint16 h2z_tbl[] = {
215                 /* 0x20 - 0x2f */
216                 0x0000, 0x2123, 0x2156, 0x2157, 0x2122, 0x2126, 0x2572, 0x2521,
217                 0x2523, 0x2525, 0x2527, 0x2529, 0x2563, 0x2565, 0x2567, 0x2543,
218                 /* 0x30 - 0x3f */
219                 0x213c, 0x2522, 0x2524, 0x2526, 0x2528, 0x252a, 0x252b, 0x252d,
220                 0x252f, 0x2531, 0x2533, 0x2535, 0x2537, 0x2539, 0x253b, 0x253d,
221                 /* 0x40 - 0x4f */
222                 0x253f, 0x2541, 0x2544, 0x2546, 0x2548, 0x254a, 0x254b, 0x254c,
223                 0x254d, 0x254e, 0x254f, 0x2552, 0x2555, 0x2558, 0x255b, 0x255e,
224                 /* 0x50 - 0x5f */
225                 0x255f, 0x2560, 0x2561, 0x2562, 0x2564, 0x2566, 0x2568, 0x2569,
226                 0x256a, 0x256b, 0x256c, 0x256d, 0x256f, 0x2573, 0x212b, 0x212c
227         };
228
229         static guint16 dakuten_tbl[] = {
230                 /* 0x30 - 0x3f */
231                 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x252c, 0x252e,
232                 0x2530, 0x2532, 0x2534, 0x2536, 0x2538, 0x253a, 0x253c, 0x253e,
233                 /* 0x40 - 0x4f */
234                 0x2540, 0x2542, 0x2545, 0x2547, 0x2549, 0x0000, 0x0000, 0x0000,
235                 0x0000, 0x0000, 0x2550, 0x2553, 0x2556, 0x2559, 0x255c, 0x0000
236         };
237
238         static guint16 handakuten_tbl[] = {
239                 /* 0x4a - 0x4e */
240                 0x2551, 0x2554, 0x2557, 0x255a, 0x255d
241         };
242
243         guint16 out_code;
244
245         jis_code &= 0x7f;
246         sound_sym &= 0x7f;
247
248         if (jis_code < 0x21 || jis_code > 0x5f)
249                 return 0;
250
251         if (sound_sym == JIS_HWDAKUTEN &&
252             jis_code >= 0x36 && jis_code <= 0x4e) {
253                 out_code = dakuten_tbl[jis_code - 0x30];
254                 if (out_code != 0) {
255                         *outbuf = out_code >> 8;
256                         *(outbuf + 1) = out_code & 0xff;
257                         return 2;
258                 }
259         }
260
261         if (sound_sym == JIS_HWHANDAKUTEN &&
262             jis_code >= 0x4a && jis_code <= 0x4e) {
263                 out_code = handakuten_tbl[jis_code - 0x4a];
264                 *outbuf = out_code >> 8;
265                 *(outbuf + 1) = out_code & 0xff;
266                 return 2;
267         }
268
269         out_code = h2z_tbl[jis_code - 0x20];
270         *outbuf = out_code >> 8;
271         *(outbuf + 1) = out_code & 0xff;
272         return 1;
273 }
274
275 static void conv_euctojis(gchar *outbuf, gint outlen, const gchar *inbuf)
276 {
277         const guchar *in = inbuf;
278         guchar *out = outbuf;
279         JISState state = JIS_ASCII;
280
281         while (*in != '\0') {
282                 if (IS_ASCII(*in)) {
283                         K_OUT();
284                         *out++ = *in++;
285                 } else if (iseuckanji(*in)) {
286                         if (iseuckanji(*(in + 1))) {
287                                 K_IN();
288                                 *out++ = *in++ & 0x7f;
289                                 *out++ = *in++ & 0x7f;
290                         } else {
291                                 K_OUT();
292                                 *out++ = SUBST_CHAR;
293                                 in++;
294                                 if (*in != '\0' && !IS_ASCII(*in)) {
295                                         *out++ = SUBST_CHAR;
296                                         in++;
297                                 }
298                         }
299                 } else if (iseuchwkana1(*in)) {
300                         if (iseuchwkana2(*(in + 1))) {
301                                 if (prefs_common.allow_jisx0201_kana) {
302                                         HW_IN();
303                                         in++;
304                                         *out++ = *in++ & 0x7f;
305                                 } else {
306                                         guchar jis_ch[2];
307                                         gint len;
308
309                                         if (iseuchwkana1(*(in + 2)) &&
310                                             iseuchwkana2(*(in + 3)))
311                                                 len = conv_jis_hantozen
312                                                         (jis_ch,
313                                                          *(in + 1), *(in + 3));
314                                         else
315                                                 len = conv_jis_hantozen
316                                                         (jis_ch,
317                                                          *(in + 1), '\0');
318                                         if (len == 0)
319                                                 in += 2;
320                                         else {
321                                                 K_IN();
322                                                 in += len * 2;
323                                                 *out++ = jis_ch[0];
324                                                 *out++ = jis_ch[1];
325                                         }
326                                 }
327                         } else {
328                                 K_OUT();
329                                 in++;
330                                 if (*in != '\0' && !IS_ASCII(*in)) {
331                                         *out++ = SUBST_CHAR;
332                                         in++;
333                                 }
334                         }
335                 } else if (iseucaux(*in)) {
336                         in++;
337                         if (iseuckanji(*in) && iseuckanji(*(in + 1))) {
338                                 AUX_IN();
339                                 *out++ = *in++ & 0x7f;
340                                 *out++ = *in++ & 0x7f;
341                         } else {
342                                 K_OUT();
343                                 if (*in != '\0' && !IS_ASCII(*in)) {
344                                         *out++ = SUBST_CHAR;
345                                         in++;
346                                         if (*in != '\0' && !IS_ASCII(*in)) {
347                                                 *out++ = SUBST_CHAR;
348                                                 in++;
349                                         }
350                                 }
351                         }
352                 } else {
353                         K_OUT();
354                         *out++ = SUBST_CHAR;
355                         in++;
356                 }
357         }
358
359         K_OUT();
360         *out = '\0';
361 }
362
363 static void conv_sjistoeuc(gchar *outbuf, gint outlen, const gchar *inbuf)
364 {
365         const guchar *in = inbuf;
366         guchar *out = outbuf;
367
368         while (*in != '\0') {
369                 if (IS_ASCII(*in)) {
370                         *out++ = *in++;
371                 } else if (issjiskanji1(*in)) {
372                         if (issjiskanji2(*(in + 1))) {
373                                 guchar out1 = *in;
374                                 guchar out2 = *(in + 1);
375                                 guchar row;
376
377                                 row = out1 < 0xa0 ? 0x70 : 0xb0;
378                                 if (out2 < 0x9f) {
379                                         out1 = (out1 - row) * 2 - 1;
380                                         out2 -= out2 > 0x7f ? 0x20 : 0x1f;
381                                 } else {
382                                         out1 = (out1 - row) * 2;
383                                         out2 -= 0x7e;
384                                 }
385
386                                 *out++ = out1 | 0x80;
387                                 *out++ = out2 | 0x80;
388                                 in += 2;
389                         } else {
390                                 *out++ = SUBST_CHAR;
391                                 in++;
392                                 if (*in != '\0' && !IS_ASCII(*in)) {
393                                         *out++ = SUBST_CHAR;
394                                         in++;
395                                 }
396                         }
397                 } else if (issjishwkana(*in)) {
398                         *out++ = 0x8e;
399                         *out++ = *in++;
400                 } else {
401                         *out++ = SUBST_CHAR;
402                         in++;
403                 }
404         }
405
406         *out = '\0';
407 }
408
409 static void conv_jistoutf8(gchar *outbuf, gint outlen, const gchar *inbuf)
410 {
411         gchar *eucstr;
412
413         Xalloca(eucstr, outlen, return);
414
415         conv_jistoeuc(eucstr, outlen, inbuf);
416         conv_euctoutf8(outbuf, outlen, eucstr);
417 }
418
419 static void conv_sjistoutf8(gchar *outbuf, gint outlen, const gchar *inbuf)
420 {
421         gchar *tmpstr;
422
423         tmpstr = conv_iconv_strdup(inbuf, CS_SHIFT_JIS, CS_UTF_8);
424         if (tmpstr) {
425                 strncpy2(outbuf, tmpstr, outlen);
426                 g_free(tmpstr);
427         } else
428                 strncpy2(outbuf, inbuf, outlen);
429 }
430
431 static void conv_euctoutf8(gchar *outbuf, gint outlen, const gchar *inbuf)
432 {
433         static iconv_t cd = (iconv_t)-1;
434         static gboolean iconv_ok = TRUE;
435         gchar *tmpstr;
436
437         if (cd == (iconv_t)-1) {
438                 if (!iconv_ok) {
439                         strncpy2(outbuf, inbuf, outlen);
440                         return;
441                 }
442                 cd = iconv_open(CS_UTF_8, CS_EUC_JP_MS);
443                 if (cd == (iconv_t)-1) {
444                         cd = iconv_open(CS_UTF_8, CS_EUC_JP);
445                         if (cd == (iconv_t)-1) {
446                                 g_warning("conv_euctoutf8(): %s\n",
447                                           g_strerror(errno));
448                                 iconv_ok = FALSE;
449                                 strncpy2(outbuf, inbuf, outlen);
450                                 return;
451                         }
452                 }
453         }
454
455         tmpstr = conv_iconv_strdup_with_cd(inbuf, cd);
456         if (tmpstr) {
457                 strncpy2(outbuf, tmpstr, outlen);
458                 g_free(tmpstr);
459         } else
460                 strncpy2(outbuf, inbuf, outlen);
461 }
462
463 static void conv_anytoutf8(gchar *outbuf, gint outlen, const gchar *inbuf)
464 {
465         switch (conv_guess_ja_encoding(inbuf)) {
466         case C_ISO_2022_JP:
467                 conv_jistoutf8(outbuf, outlen, inbuf);
468                 break;
469         case C_SHIFT_JIS:
470                 conv_sjistoutf8(outbuf, outlen, inbuf);
471                 break;
472         case C_EUC_JP:
473                 conv_euctoutf8(outbuf, outlen, inbuf);
474                 break;
475         default:
476                 strncpy2(outbuf, inbuf, outlen);
477                 break;
478         }
479 }
480
481 static void conv_utf8toeuc(gchar *outbuf, gint outlen, const gchar *inbuf)
482 {
483         static iconv_t cd = (iconv_t)-1;
484         static gboolean iconv_ok = TRUE;
485         gchar *tmpstr;
486
487         if (cd == (iconv_t)-1) {
488                 if (!iconv_ok) {
489                         strncpy2(outbuf, inbuf, outlen);
490                         return;
491                 }
492                 cd = iconv_open(CS_EUC_JP_MS, CS_UTF_8);
493                 if (cd == (iconv_t)-1) {
494                         cd = iconv_open(CS_EUC_JP, CS_UTF_8);
495                         if (cd == (iconv_t)-1) {
496                                 g_warning("conv_utf8toeuc(): %s\n",
497                                           g_strerror(errno));
498                                 iconv_ok = FALSE;
499                                 strncpy2(outbuf, inbuf, outlen);
500                                 return;
501                         }
502                 }
503         }
504
505         tmpstr = conv_iconv_strdup_with_cd(inbuf, cd);
506         if (tmpstr) {
507                 strncpy2(outbuf, tmpstr, outlen);
508                 g_free(tmpstr);
509         } else
510                 strncpy2(outbuf, inbuf, outlen);
511 }
512
513 static void conv_utf8tojis(gchar *outbuf, gint outlen, const gchar *inbuf)
514 {
515         gchar *eucstr;
516
517         Xalloca(eucstr, outlen, return);
518
519         conv_utf8toeuc(eucstr, outlen, inbuf);
520         conv_euctojis(outbuf, outlen, eucstr);
521 }
522
523 static void conv_unreadable_8bit(gchar *str)
524 {
525         register guchar *p = str;
526
527         while (*p != '\0') {
528                 /* convert CR+LF -> LF */
529                 if (*p == '\r' && *(p + 1) == '\n')
530                         memmove(p, p + 1, strlen(p));
531                 else if (!IS_ASCII(*p)) *p = SUBST_CHAR;
532                 p++;
533         }
534 }
535
536 CharSet conv_guess_ja_encoding(const gchar *str)
537 {
538         const guchar *p = str;
539         CharSet guessed = C_US_ASCII;
540
541         while (*p != '\0') {
542                 if (*p == ESC && (*(p + 1) == '$' || *(p + 1) == '(')) {
543                         if (guessed == C_US_ASCII)
544                                 return C_ISO_2022_JP;
545                         p += 2;
546                 } else if (IS_ASCII(*p)) {
547                         p++;
548                 } else if (iseuckanji(*p) && iseuckanji(*(p + 1))) {
549                         if (*p >= 0xfd && *p <= 0xfe)
550                                 return C_EUC_JP;
551                         else if (guessed == C_SHIFT_JIS) {
552                                 if ((issjiskanji1(*p) &&
553                                      issjiskanji2(*(p + 1))) ||
554                                     issjishwkana(*p))
555                                         guessed = C_SHIFT_JIS;
556                                 else
557                                         guessed = C_EUC_JP;
558                         } else
559                                 guessed = C_EUC_JP;
560                         p += 2;
561                 } else if (issjiskanji1(*p) && issjiskanji2(*(p + 1))) {
562                         if (iseuchwkana1(*p) && iseuchwkana2(*(p + 1)))
563                                 guessed = C_SHIFT_JIS;
564                         else
565                                 return C_SHIFT_JIS;
566                         p += 2;
567                 } else if (issjishwkana(*p)) {
568                         guessed = C_SHIFT_JIS;
569                         p++;
570                 } else {
571                         p++;
572                 }
573         }
574
575         return guessed;
576 }
577
578 static void conv_jistodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
579 {
580         conv_jistoutf8(outbuf, outlen, inbuf);
581 }
582
583 static void conv_sjistodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
584 {
585         conv_sjistoutf8(outbuf, outlen, inbuf);
586 }
587
588 static void conv_euctodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
589 {
590         conv_euctoutf8(outbuf, outlen, inbuf);
591 }
592
593 void conv_utf8todisp(gchar *outbuf, gint outlen, const gchar *inbuf)
594 {
595         if (g_utf8_validate(inbuf, -1, NULL) == TRUE)
596                 strncpy2(outbuf, inbuf, outlen);
597         else
598                 conv_ustodisp(outbuf, outlen, inbuf);
599 }
600
601 static void conv_anytodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
602 {
603         conv_anytoutf8(outbuf, outlen, inbuf);
604         if (g_utf8_validate(outbuf, -1, NULL) != TRUE)
605                 conv_unreadable_8bit(outbuf);
606 }
607
608 static void conv_ustodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
609 {
610         strncpy2(outbuf, inbuf, outlen);
611         conv_unreadable_8bit(outbuf);
612 }
613
614 void conv_localetodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
615 {
616         gchar *tmpstr;
617
618         codeconv_set_strict(TRUE);
619         tmpstr = conv_iconv_strdup(inbuf, conv_get_locale_charset_str(),
620                                    CS_INTERNAL);
621         codeconv_set_strict(FALSE);
622         if (tmpstr && g_utf8_validate(tmpstr, -1, NULL)) {
623                 strncpy2(outbuf, tmpstr, outlen);
624                 g_free(tmpstr);
625                 return;
626         } else if (tmpstr && !g_utf8_validate(tmpstr, -1, NULL)) {
627                 g_free(tmpstr);
628                 codeconv_set_strict(TRUE);
629                 tmpstr = conv_iconv_strdup(inbuf, 
630                                 conv_get_locale_charset_str_no_utf8(),
631                                 CS_INTERNAL);
632                 codeconv_set_strict(FALSE);
633         }
634         if (tmpstr && g_utf8_validate(tmpstr, -1, NULL)) {
635                 strncpy2(outbuf, tmpstr, outlen);
636                 g_free(tmpstr);
637                 return;
638         } else {
639                 g_free(tmpstr);
640                 conv_utf8todisp(outbuf, outlen, inbuf);
641         }
642 }
643
644 static void conv_noconv(gchar *outbuf, gint outlen, const gchar *inbuf)
645 {
646         strncpy2(outbuf, inbuf, outlen);
647 }
648
649 static const gchar *
650 conv_get_fallback_for_private_encoding(const gchar *encoding)
651 {
652         if (encoding && (encoding[0] == 'X' || encoding[0] == 'x') &&
653             encoding[1] == '-') {
654                 if (!g_ascii_strcasecmp(encoding, CS_X_GBK))
655                         return CS_GBK;
656         }
657
658         return encoding;
659 }
660
661 CodeConverter *conv_code_converter_new(const gchar *src_charset)
662 {
663         CodeConverter *conv;
664
665         src_charset = conv_get_fallback_for_private_encoding(src_charset);
666
667         conv = g_new0(CodeConverter, 1);
668         conv->code_conv_func = conv_get_code_conv_func(src_charset, NULL);
669         conv->charset_str = g_strdup(src_charset);
670         conv->charset = conv_get_charset_from_str(src_charset);
671
672         return conv;
673 }
674
675 void conv_code_converter_destroy(CodeConverter *conv)
676 {
677         g_free(conv->charset_str);
678         g_free(conv);
679 }
680
681 gint conv_convert(CodeConverter *conv, gchar *outbuf, gint outlen,
682                   const gchar *inbuf)
683 {
684         if (conv->code_conv_func != conv_noconv)
685                 conv->code_conv_func(outbuf, outlen, inbuf);
686         else {
687                 gchar *str;
688
689                 str = conv_iconv_strdup(inbuf, conv->charset_str, NULL);
690                 if (!str)
691                         return -1;
692                 else {
693                         strncpy2(outbuf, str, outlen);
694                         g_free(str);
695                 }
696         }
697
698         return 0;
699 }
700
701 gchar *conv_codeset_strdup(const gchar *inbuf,
702                            const gchar *src_code, const gchar *dest_code)
703 {
704         gchar *buf;
705         size_t len;
706         CodeConvFunc conv_func;
707
708         src_code = conv_get_fallback_for_private_encoding(src_code);
709         conv_func = conv_get_code_conv_func(src_code, dest_code);
710         if (conv_func != conv_noconv) {
711                 len = (strlen(inbuf) + 1) * 3;
712                 buf = g_malloc(len);
713                 if (!buf) return NULL;
714
715                 conv_func(buf, len, inbuf);
716                 return g_realloc(buf, strlen(buf) + 1);
717         }
718
719         return conv_iconv_strdup(inbuf, src_code, dest_code);
720 }
721
722 CodeConvFunc conv_get_code_conv_func(const gchar *src_charset_str,
723                                      const gchar *dest_charset_str)
724 {
725         CodeConvFunc code_conv = conv_noconv;
726         CharSet src_charset;
727         CharSet dest_charset;
728
729         if (!src_charset_str)
730                 src_charset = conv_get_locale_charset();
731         else
732                 src_charset = conv_get_charset_from_str(src_charset_str);
733
734         /* auto detection mode */
735         if (!src_charset_str && !dest_charset_str) {
736                 if (conv_is_ja_locale())
737                         return conv_anytodisp;
738                 else
739                         return conv_noconv;
740         }
741
742         dest_charset = conv_get_charset_from_str(dest_charset_str);
743
744         if (dest_charset == C_US_ASCII)
745                 return conv_ustodisp;
746
747         switch (src_charset) {
748         case C_US_ASCII:
749         case C_ISO_8859_1:
750         case C_ISO_8859_2:
751         case C_ISO_8859_3:
752         case C_ISO_8859_4:
753         case C_ISO_8859_5:
754         case C_ISO_8859_6:
755         case C_ISO_8859_7:
756         case C_ISO_8859_8:
757         case C_ISO_8859_9:
758         case C_ISO_8859_10:
759         case C_ISO_8859_11:
760         case C_ISO_8859_13:
761         case C_ISO_8859_14:
762         case C_ISO_8859_15:
763                 break;
764         case C_ISO_2022_JP:
765         case C_ISO_2022_JP_2:
766         case C_ISO_2022_JP_3:
767                 if (dest_charset == C_AUTO)
768                         code_conv = conv_jistodisp;
769                 else if (dest_charset == C_EUC_JP)
770                         code_conv = conv_jistoeuc;
771                 else if (dest_charset == C_UTF_8)
772                         code_conv = conv_jistoutf8;
773                 break;
774         case C_SHIFT_JIS:
775                 if (dest_charset == C_AUTO)
776                         code_conv = conv_sjistodisp;
777                 else if (dest_charset == C_EUC_JP)
778                         code_conv = conv_sjistoeuc;
779                 else if (dest_charset == C_UTF_8)
780                         code_conv = conv_sjistoutf8;
781                 break;
782         case C_EUC_JP:
783                 if (dest_charset == C_AUTO)
784                         code_conv = conv_euctodisp;
785                 else if (dest_charset == C_ISO_2022_JP   ||
786                          dest_charset == C_ISO_2022_JP_2 ||
787                          dest_charset == C_ISO_2022_JP_3)
788                         code_conv = conv_euctojis;
789                 else if (dest_charset == C_UTF_8)
790                         code_conv = conv_euctoutf8;
791                 break;
792         case C_UTF_8:
793                 if (dest_charset == C_EUC_JP)
794                         code_conv = conv_utf8toeuc;
795                 else if (dest_charset == C_ISO_2022_JP   ||
796                          dest_charset == C_ISO_2022_JP_2 ||
797                          dest_charset == C_ISO_2022_JP_3)
798                         code_conv = conv_utf8tojis;
799                 break;
800         default:
801                 break;
802         }
803
804         return code_conv;
805 }
806
807 gchar *conv_iconv_strdup(const gchar *inbuf,
808                          const gchar *src_code, const gchar *dest_code)
809 {
810         iconv_t cd;
811         gchar *outbuf;
812
813         if (!src_code)
814                 src_code = conv_get_outgoing_charset_str();
815         if (!dest_code)
816                 dest_code = CS_INTERNAL;
817
818         /* don't convert if src and dest codeset are identical */
819         if (!strcasecmp(src_code, dest_code))
820                 return g_strdup(inbuf);
821
822         /* don't convert if current codeset is US-ASCII */
823         if (!strcasecmp(dest_code, CS_US_ASCII))
824                 return g_strdup(inbuf);
825
826         cd = iconv_open(dest_code, src_code);
827         if (cd == (iconv_t)-1)
828                 return NULL;
829
830         outbuf = conv_iconv_strdup_with_cd(inbuf, cd);
831
832         iconv_close(cd);
833
834         return outbuf;
835 }
836
837 gchar *conv_iconv_strdup_with_cd(const gchar *inbuf, iconv_t cd)
838 {
839         const gchar *inbuf_p;
840         gchar *outbuf;
841         gchar *outbuf_p;
842         size_t in_size;
843         size_t in_left;
844         size_t out_size;
845         size_t out_left;
846         size_t n_conv;
847         size_t len;
848
849         inbuf_p = inbuf;
850         in_size = strlen(inbuf);
851         in_left = in_size;
852         out_size = (in_size + 1) * 2;
853         outbuf = g_malloc(out_size);
854         outbuf_p = outbuf;
855         out_left = out_size;
856
857 #define EXPAND_BUF()                            \
858 {                                               \
859         len = outbuf_p - outbuf;                \
860         out_size *= 2;                          \
861         outbuf = g_realloc(outbuf, out_size);   \
862         outbuf_p = outbuf + len;                \
863         out_left = out_size - len;              \
864 }
865
866         while ((n_conv = iconv(cd, (ICONV_CONST gchar **)&inbuf_p, &in_left,
867                                &outbuf_p, &out_left)) == (size_t)-1) {
868                 if (EILSEQ == errno) {
869                         if (strict_mode) {
870                                 g_free(outbuf);
871                                 return NULL;
872                         }
873                         //g_print("iconv(): at %d: %s\n", in_size - in_left, g_strerror(errno));
874                         inbuf_p++;
875                         in_left--;
876                         if (out_left == 0) {
877                                 EXPAND_BUF();
878                         }
879                         *outbuf_p++ = SUBST_CHAR;
880                         out_left--;
881                 } else if (EINVAL == errno) {
882                         break;
883                 } else if (E2BIG == errno) {
884                         EXPAND_BUF();
885                 } else {
886                         g_warning("conv_iconv_strdup(): %s\n",
887                                   g_strerror(errno));
888                         break;
889                 }
890         }
891
892         while ((n_conv = iconv(cd, NULL, NULL, &outbuf_p, &out_left)) ==
893                (size_t)-1) {
894                 if (E2BIG == errno) {
895                         EXPAND_BUF();
896                 } else {
897                         g_warning("conv_iconv_strdup(): %s\n",
898                                   g_strerror(errno));
899                         break;
900                 }
901         }
902
903 #undef EXPAND_BUF
904
905         len = outbuf_p - outbuf;
906         outbuf = g_realloc(outbuf, len + 1);
907         outbuf[len] = '\0';
908
909         return outbuf;
910 }
911
912 static const struct {
913         CharSet charset;
914         gchar *const name;
915 } charsets[] = {
916         {C_US_ASCII,            CS_US_ASCII},
917         {C_US_ASCII,            CS_ANSI_X3_4_1968},
918         {C_UTF_8,               CS_UTF_8},
919         {C_UTF_7,               CS_UTF_7},
920         {C_ISO_8859_1,          CS_ISO_8859_1},
921         {C_ISO_8859_2,          CS_ISO_8859_2},
922         {C_ISO_8859_3,          CS_ISO_8859_3},
923         {C_ISO_8859_4,          CS_ISO_8859_4},
924         {C_ISO_8859_5,          CS_ISO_8859_5},
925         {C_ISO_8859_6,          CS_ISO_8859_6},
926         {C_ISO_8859_7,          CS_ISO_8859_7},
927         {C_ISO_8859_8,          CS_ISO_8859_8},
928         {C_ISO_8859_9,          CS_ISO_8859_9},
929         {C_ISO_8859_10,         CS_ISO_8859_10},
930         {C_ISO_8859_11,         CS_ISO_8859_11},
931         {C_ISO_8859_13,         CS_ISO_8859_13},
932         {C_ISO_8859_14,         CS_ISO_8859_14},
933         {C_ISO_8859_15,         CS_ISO_8859_15},
934         {C_BALTIC,              CS_BALTIC},
935         {C_CP1250,              CS_CP1250},
936         {C_CP1251,              CS_CP1251},
937         {C_CP1252,              CS_CP1252},
938         {C_CP1253,              CS_CP1253},
939         {C_CP1254,              CS_CP1254},
940         {C_CP1255,              CS_CP1255},
941         {C_CP1256,              CS_CP1256},
942         {C_CP1257,              CS_CP1257},
943         {C_CP1258,              CS_CP1258},
944         {C_WINDOWS_1250,        CS_WINDOWS_1250},
945         {C_WINDOWS_1251,        CS_WINDOWS_1251},
946         {C_WINDOWS_1252,        CS_WINDOWS_1252},
947         {C_WINDOWS_1253,        CS_WINDOWS_1253},
948         {C_WINDOWS_1254,        CS_WINDOWS_1254},
949         {C_WINDOWS_1255,        CS_WINDOWS_1255},
950         {C_WINDOWS_1256,        CS_WINDOWS_1256},
951         {C_WINDOWS_1257,        CS_WINDOWS_1257},
952         {C_WINDOWS_1258,        CS_WINDOWS_1258},
953         {C_KOI8_R,              CS_KOI8_R},
954         {C_KOI8_T,              CS_KOI8_T},
955         {C_KOI8_U,              CS_KOI8_U},
956         {C_ISO_2022_JP,         CS_ISO_2022_JP},
957         {C_ISO_2022_JP_2,       CS_ISO_2022_JP_2},
958         {C_ISO_2022_JP_3,       CS_ISO_2022_JP_3},
959         {C_EUC_JP,              CS_EUC_JP},
960         {C_EUC_JP,              CS_EUCJP},
961         {C_EUC_JP_MS,           CS_EUC_JP_MS},
962         {C_SHIFT_JIS,           CS_SHIFT_JIS},
963         {C_SHIFT_JIS,           CS_SHIFT__JIS},
964         {C_SHIFT_JIS,           CS_SJIS},
965         {C_ISO_2022_KR,         CS_ISO_2022_KR},
966         {C_EUC_KR,              CS_EUC_KR},
967         {C_ISO_2022_CN,         CS_ISO_2022_CN},
968         {C_EUC_CN,              CS_EUC_CN},
969         {C_GB2312,              CS_GB2312},
970         {C_GBK,                 CS_GBK},
971         {C_EUC_TW,              CS_EUC_TW},
972         {C_BIG5,                CS_BIG5},
973         {C_BIG5_HKSCS,          CS_BIG5_HKSCS},
974         {C_TIS_620,             CS_TIS_620},
975         {C_WINDOWS_874,         CS_WINDOWS_874},
976         {C_GEORGIAN_PS,         CS_GEORGIAN_PS},
977         {C_TCVN5712_1,          CS_TCVN5712_1},
978 };
979
980 static const struct {
981         gchar *const locale;
982         CharSet charset;
983         CharSet out_charset;
984 } locale_table[] = {
985         {"ja_JP.eucJP"  , C_EUC_JP      , C_ISO_2022_JP},
986         {"ja_JP.EUC-JP" , C_EUC_JP      , C_ISO_2022_JP},
987         {"ja_JP.EUC"    , C_EUC_JP      , C_ISO_2022_JP},
988         {"ja_JP.ujis"   , C_EUC_JP      , C_ISO_2022_JP},
989         {"ja_JP.SJIS"   , C_SHIFT_JIS   , C_ISO_2022_JP},
990         {"ja_JP.JIS"    , C_ISO_2022_JP , C_ISO_2022_JP},
991 #ifdef G_OS_WIN32
992         {"ja_JP"        , C_SHIFT_JIS   , C_ISO_2022_JP},
993 #else
994         {"ja_JP"        , C_EUC_JP      , C_ISO_2022_JP},
995 #endif
996         {"ko_KR.EUC-KR" , C_EUC_KR      , C_EUC_KR},
997         {"ko_KR"        , C_EUC_KR      , C_EUC_KR},
998         {"zh_CN.GB2312" , C_GB2312      , C_GB2312},
999         {"zh_CN.GBK"    , C_GBK         , C_GBK},
1000         {"zh_CN"        , C_GB2312      , C_GB2312},
1001         {"zh_HK"        , C_BIG5_HKSCS  , C_BIG5_HKSCS},
1002         {"zh_TW.eucTW"  , C_EUC_TW      , C_BIG5},
1003         {"zh_TW.EUC-TW" , C_EUC_TW      , C_BIG5},
1004         {"zh_TW.Big5"   , C_BIG5        , C_BIG5},
1005         {"zh_TW"        , C_BIG5        , C_BIG5},
1006
1007         {"ru_RU.KOI8-R" , C_KOI8_R      , C_KOI8_R},
1008         {"ru_RU.KOI8R"  , C_KOI8_R      , C_KOI8_R},
1009         {"ru_RU.CP1251" , C_WINDOWS_1251, C_KOI8_R},
1010         {"ru_RU"        , C_ISO_8859_5  , C_KOI8_R},
1011         {"tg_TJ"        , C_KOI8_T      , C_KOI8_T},
1012         {"ru_UA"        , C_KOI8_U      , C_KOI8_U},
1013         {"uk_UA.CP1251" , C_WINDOWS_1251, C_KOI8_U},
1014         {"uk_UA"        , C_KOI8_U      , C_KOI8_U},
1015
1016         {"be_BY"        , C_WINDOWS_1251, C_WINDOWS_1251},
1017         {"bg_BG"        , C_WINDOWS_1251, C_WINDOWS_1251},
1018
1019         {"yi_US"        , C_WINDOWS_1255, C_WINDOWS_1255},
1020
1021         {"af_ZA"        , C_ISO_8859_1  , C_ISO_8859_1},
1022         {"br_FR"        , C_ISO_8859_1  , C_ISO_8859_1},
1023         {"ca_ES"        , C_ISO_8859_1  , C_ISO_8859_1},
1024         {"da_DK"        , C_ISO_8859_1  , C_ISO_8859_1},
1025         {"de_AT"        , C_ISO_8859_1  , C_ISO_8859_1},
1026         {"de_BE"        , C_ISO_8859_1  , C_ISO_8859_1},
1027         {"de_CH"        , C_ISO_8859_1  , C_ISO_8859_1},
1028         {"de_DE"        , C_ISO_8859_1  , C_ISO_8859_1},
1029         {"de_LU"        , C_ISO_8859_1  , C_ISO_8859_1},
1030         {"en_AU"        , C_ISO_8859_1  , C_ISO_8859_1},
1031         {"en_BW"        , C_ISO_8859_1  , C_ISO_8859_1},
1032         {"en_CA"        , C_ISO_8859_1  , C_ISO_8859_1},
1033         {"en_DK"        , C_ISO_8859_1  , C_ISO_8859_1},
1034         {"en_GB"        , C_ISO_8859_1  , C_ISO_8859_1},
1035         {"en_HK"        , C_ISO_8859_1  , C_ISO_8859_1},
1036         {"en_IE"        , C_ISO_8859_1  , C_ISO_8859_1},
1037         {"en_NZ"        , C_ISO_8859_1  , C_ISO_8859_1},
1038         {"en_PH"        , C_ISO_8859_1  , C_ISO_8859_1},
1039         {"en_SG"        , C_ISO_8859_1  , C_ISO_8859_1},
1040         {"en_US"        , C_ISO_8859_1  , C_ISO_8859_1},
1041         {"en_ZA"        , C_ISO_8859_1  , C_ISO_8859_1},
1042         {"en_ZW"        , C_ISO_8859_1  , C_ISO_8859_1},
1043         {"es_AR"        , C_ISO_8859_1  , C_ISO_8859_1},
1044         {"es_BO"        , C_ISO_8859_1  , C_ISO_8859_1},
1045         {"es_CL"        , C_ISO_8859_1  , C_ISO_8859_1},
1046         {"es_CO"        , C_ISO_8859_1  , C_ISO_8859_1},
1047         {"es_CR"        , C_ISO_8859_1  , C_ISO_8859_1},
1048         {"es_DO"        , C_ISO_8859_1  , C_ISO_8859_1},
1049         {"es_EC"        , C_ISO_8859_1  , C_ISO_8859_1},
1050         {"es_ES"        , C_ISO_8859_1  , C_ISO_8859_1},
1051         {"es_GT"        , C_ISO_8859_1  , C_ISO_8859_1},
1052         {"es_HN"        , C_ISO_8859_1  , C_ISO_8859_1},
1053         {"es_MX"        , C_ISO_8859_1  , C_ISO_8859_1},
1054         {"es_NI"        , C_ISO_8859_1  , C_ISO_8859_1},
1055         {"es_PA"        , C_ISO_8859_1  , C_ISO_8859_1},
1056         {"es_PE"        , C_ISO_8859_1  , C_ISO_8859_1},
1057         {"es_PR"        , C_ISO_8859_1  , C_ISO_8859_1},
1058         {"es_PY"        , C_ISO_8859_1  , C_ISO_8859_1},
1059         {"es_SV"        , C_ISO_8859_1  , C_ISO_8859_1},
1060         {"es_US"        , C_ISO_8859_1  , C_ISO_8859_1},
1061         {"es_UY"        , C_ISO_8859_1  , C_ISO_8859_1},
1062         {"es_VE"        , C_ISO_8859_1  , C_ISO_8859_1},
1063         {"et_EE"        , C_ISO_8859_1  , C_ISO_8859_1},
1064         {"eu_ES"        , C_ISO_8859_1  , C_ISO_8859_1},
1065         {"fi_FI"        , C_ISO_8859_1  , C_ISO_8859_1},
1066         {"fo_FO"        , C_ISO_8859_1  , C_ISO_8859_1},
1067         {"fr_BE"        , C_ISO_8859_1  , C_ISO_8859_1},
1068         {"fr_CA"        , C_ISO_8859_1  , C_ISO_8859_1},
1069         {"fr_CH"        , C_ISO_8859_1  , C_ISO_8859_1},
1070         {"fr_FR"        , C_ISO_8859_1  , C_ISO_8859_1},
1071         {"fr_LU"        , C_ISO_8859_1  , C_ISO_8859_1},
1072         {"ga_IE"        , C_ISO_8859_1  , C_ISO_8859_1},
1073         {"gl_ES"        , C_ISO_8859_1  , C_ISO_8859_1},
1074         {"gv_GB"        , C_ISO_8859_1  , C_ISO_8859_1},
1075         {"id_ID"        , C_ISO_8859_1  , C_ISO_8859_1},
1076         {"is_IS"        , C_ISO_8859_1  , C_ISO_8859_1},
1077         {"it_CH"        , C_ISO_8859_1  , C_ISO_8859_1},
1078         {"it_IT"        , C_ISO_8859_1  , C_ISO_8859_1},
1079         {"kl_GL"        , C_ISO_8859_1  , C_ISO_8859_1},
1080         {"kw_GB"        , C_ISO_8859_1  , C_ISO_8859_1},
1081         {"ms_MY"        , C_ISO_8859_1  , C_ISO_8859_1},
1082         {"nl_BE"        , C_ISO_8859_1  , C_ISO_8859_1},
1083         {"nl_NL"        , C_ISO_8859_1  , C_ISO_8859_1},
1084         {"nn_NO"        , C_ISO_8859_1  , C_ISO_8859_1},
1085         {"no_NO"        , C_ISO_8859_1  , C_ISO_8859_1},
1086         {"oc_FR"        , C_ISO_8859_1  , C_ISO_8859_1},
1087         {"pt_BR"        , C_ISO_8859_1  , C_ISO_8859_1},
1088         {"pt_PT"        , C_ISO_8859_1  , C_ISO_8859_1},
1089         {"sq_AL"        , C_ISO_8859_1  , C_ISO_8859_1},
1090         {"sv_FI"        , C_ISO_8859_1  , C_ISO_8859_1},
1091         {"sv_SE"        , C_ISO_8859_1  , C_ISO_8859_1},
1092         {"tl_PH"        , C_ISO_8859_1  , C_ISO_8859_1},
1093         {"uz_UZ"        , C_ISO_8859_1  , C_ISO_8859_1},
1094         {"wa_BE"        , C_ISO_8859_1  , C_ISO_8859_1},
1095
1096         {"bs_BA"        , C_ISO_8859_2  , C_ISO_8859_2},
1097         {"cs_CZ"        , C_ISO_8859_2  , C_ISO_8859_2},
1098         {"hr_HR"        , C_ISO_8859_2  , C_ISO_8859_2},
1099         {"hu_HU"        , C_ISO_8859_2  , C_ISO_8859_2},
1100         {"pl_PL"        , C_ISO_8859_2  , C_ISO_8859_2},
1101         {"ro_RO"        , C_ISO_8859_2  , C_ISO_8859_2},
1102         {"sk_SK"        , C_ISO_8859_2  , C_ISO_8859_2},
1103         {"sl_SI"        , C_ISO_8859_2  , C_ISO_8859_2},
1104
1105         {"sr_YU@cyrillic"       , C_ISO_8859_5  , C_ISO_8859_5},
1106         {"sr_YU"                , C_ISO_8859_2  , C_ISO_8859_2},
1107
1108         {"mt_MT"                , C_ISO_8859_3  , C_ISO_8859_3},
1109
1110         {"lt_LT.iso88594"       , C_ISO_8859_4  , C_ISO_8859_4},
1111         {"lt_LT.ISO8859-4"      , C_ISO_8859_4  , C_ISO_8859_4},
1112         {"lt_LT.ISO_8859-4"     , C_ISO_8859_4  , C_ISO_8859_4},
1113         {"lt_LT"                , C_ISO_8859_13 , C_ISO_8859_13},
1114
1115         {"mk_MK"        , C_ISO_8859_5  , C_ISO_8859_5},
1116
1117         {"ar_AE"        , C_ISO_8859_6  , C_ISO_8859_6},
1118         {"ar_BH"        , C_ISO_8859_6  , C_ISO_8859_6},
1119         {"ar_DZ"        , C_ISO_8859_6  , C_ISO_8859_6},
1120         {"ar_EG"        , C_ISO_8859_6  , C_ISO_8859_6},
1121         {"ar_IQ"        , C_ISO_8859_6  , C_ISO_8859_6},
1122         {"ar_JO"        , C_ISO_8859_6  , C_ISO_8859_6},
1123         {"ar_KW"        , C_ISO_8859_6  , C_ISO_8859_6},
1124         {"ar_LB"        , C_ISO_8859_6  , C_ISO_8859_6},
1125         {"ar_LY"        , C_ISO_8859_6  , C_ISO_8859_6},
1126         {"ar_MA"        , C_ISO_8859_6  , C_ISO_8859_6},
1127         {"ar_OM"        , C_ISO_8859_6  , C_ISO_8859_6},
1128         {"ar_QA"        , C_ISO_8859_6  , C_ISO_8859_6},
1129         {"ar_SA"        , C_ISO_8859_6  , C_ISO_8859_6},
1130         {"ar_SD"        , C_ISO_8859_6  , C_ISO_8859_6},
1131         {"ar_SY"        , C_ISO_8859_6  , C_ISO_8859_6},
1132         {"ar_TN"        , C_ISO_8859_6  , C_ISO_8859_6},
1133         {"ar_YE"        , C_ISO_8859_6  , C_ISO_8859_6},
1134
1135         {"el_GR"        , C_ISO_8859_7  , C_ISO_8859_7},
1136         {"he_IL"        , C_ISO_8859_8  , C_ISO_8859_8},
1137         {"iw_IL"        , C_ISO_8859_8  , C_ISO_8859_8},
1138         {"tr_TR"        , C_ISO_8859_9  , C_ISO_8859_9},
1139
1140         {"lv_LV"        , C_ISO_8859_13 , C_ISO_8859_13},
1141         {"mi_NZ"        , C_ISO_8859_13 , C_ISO_8859_13},
1142
1143         {"cy_GB"        , C_ISO_8859_14 , C_ISO_8859_14},
1144
1145         {"ar_IN"        , C_UTF_8       , C_UTF_8},
1146         {"en_IN"        , C_UTF_8       , C_UTF_8},
1147         {"se_NO"        , C_UTF_8       , C_UTF_8},
1148         {"ta_IN"        , C_UTF_8       , C_UTF_8},
1149         {"te_IN"        , C_UTF_8       , C_UTF_8},
1150         {"ur_PK"        , C_UTF_8       , C_UTF_8},
1151
1152         {"th_TH"        , C_TIS_620     , C_TIS_620},
1153         /* {"th_TH"     , C_WINDOWS_874}, */
1154         /* {"th_TH"     , C_ISO_8859_11}, */
1155
1156         {"ka_GE"        , C_GEORGIAN_PS , C_GEORGIAN_PS},
1157         {"vi_VN.TCVN"   , C_TCVN5712_1  , C_TCVN5712_1},
1158
1159         {"C"                    , C_US_ASCII    , C_US_ASCII},
1160         {"POSIX"                , C_US_ASCII    , C_US_ASCII},
1161         {"ANSI_X3.4-1968"       , C_US_ASCII    , C_US_ASCII},
1162 };
1163
1164 static GHashTable *conv_get_charset_to_str_table(void)
1165 {
1166         static GHashTable *table;
1167         gint i;
1168
1169         if (table)
1170                 return table;
1171
1172         table = g_hash_table_new(NULL, g_direct_equal);
1173
1174         for (i = 0; i < sizeof(charsets) / sizeof(charsets[0]); i++) {
1175                 if (g_hash_table_lookup(table, GUINT_TO_POINTER(charsets[i].charset))
1176                     == NULL) {
1177                         g_hash_table_insert
1178                                 (table, GUINT_TO_POINTER(charsets[i].charset),
1179                                  charsets[i].name);
1180                 }
1181         }
1182
1183         return table;
1184 }
1185
1186 static GHashTable *conv_get_charset_from_str_table(void)
1187 {
1188         static GHashTable *table;
1189         gint i;
1190
1191         if (table)
1192                 return table;
1193
1194         table = g_hash_table_new(str_case_hash, str_case_equal);
1195
1196         for (i = 0; i < sizeof(charsets) / sizeof(charsets[0]); i++) {
1197                 g_hash_table_insert(table, charsets[i].name,
1198                                     GUINT_TO_POINTER(charsets[i].charset));
1199         }
1200
1201         return table;
1202 }
1203
1204 const gchar *conv_get_charset_str(CharSet charset)
1205 {
1206         GHashTable *table;
1207
1208         table = conv_get_charset_to_str_table();
1209         return g_hash_table_lookup(table, GUINT_TO_POINTER(charset));
1210 }
1211
1212 CharSet conv_get_charset_from_str(const gchar *charset)
1213 {
1214         GHashTable *table;
1215
1216         if (!charset) return C_AUTO;
1217
1218         table = conv_get_charset_from_str_table();
1219         return GPOINTER_TO_UINT(g_hash_table_lookup(table, charset));
1220 }
1221
1222 CharSet conv_get_locale_charset(void)
1223 {
1224         static CharSet cur_charset = -1;
1225         const gchar *cur_locale;
1226         const gchar *p;
1227         gint i;
1228
1229         if (cur_charset != -1)
1230                 return cur_charset;
1231
1232         cur_locale = conv_get_current_locale();
1233         if (!cur_locale) {
1234                 cur_charset = C_US_ASCII;
1235                 return cur_charset;
1236         }
1237
1238         if (strcasestr(cur_locale, "UTF-8")) {
1239                 cur_charset = C_UTF_8;
1240                 return cur_charset;
1241         }
1242
1243         if ((p = strcasestr(cur_locale, "@euro")) && p[5] == '\0') {
1244                 cur_charset = C_ISO_8859_15;
1245                 return cur_charset;
1246         }
1247
1248         for (i = 0; i < sizeof(locale_table) / sizeof(locale_table[0]); i++) {
1249                 const gchar *p;
1250
1251                 /* "ja_JP.EUC" matches with "ja_JP.eucJP", "ja_JP.EUC" and
1252                    "ja_JP". "ja_JP" matches with "ja_JP.xxxx" and "ja" */
1253                 if (!g_ascii_strncasecmp(cur_locale, locale_table[i].locale,
1254                                  strlen(locale_table[i].locale))) {
1255                         cur_charset = locale_table[i].charset;
1256                         return cur_charset;
1257                 } else if ((p = strchr(locale_table[i].locale, '_')) &&
1258                          !strchr(p + 1, '.')) {
1259                         if (strlen(cur_locale) == 2 &&
1260                             !g_ascii_strncasecmp(cur_locale, locale_table[i].locale, 2)) {
1261                                 cur_charset = locale_table[i].charset;
1262                                 return cur_charset;
1263                         }
1264                 }
1265         }
1266
1267         cur_charset = C_AUTO;
1268         return cur_charset;
1269 }
1270
1271 static CharSet conv_get_locale_charset_no_utf8(void)
1272 {
1273         static CharSet cur_charset = -1;
1274         const gchar *cur_locale;
1275         const gchar *p;
1276         gchar *tmp;
1277         gint i;
1278
1279         if (prefs_common.broken_are_utf8)
1280                 return conv_get_locale_charset();
1281
1282         if (cur_charset != -1)
1283                 return cur_charset;
1284
1285         cur_locale = conv_get_current_locale();
1286         if (!cur_locale) {
1287                 cur_charset = C_US_ASCII;
1288                 return cur_charset;
1289         }
1290
1291         if (strcasestr(cur_locale, "UTF-8")) {
1292                 tmp = g_strdup(cur_locale);
1293                 *(strcasestr(tmp, ".UTF-8")) = '\0';
1294                 cur_locale = tmp;
1295         }
1296
1297         if ((p = strcasestr(cur_locale, "@euro")) && p[5] == '\0') {
1298                 cur_charset = C_ISO_8859_15;
1299                 return cur_charset;
1300         }
1301
1302         for (i = 0; i < sizeof(locale_table) / sizeof(locale_table[0]); i++) {
1303                 const gchar *p;
1304
1305                 /* "ja_JP.EUC" matches with "ja_JP.eucJP", "ja_JP.EUC" and
1306                    "ja_JP". "ja_JP" matches with "ja_JP.xxxx" and "ja" */
1307                 if (!g_ascii_strncasecmp(cur_locale, locale_table[i].locale,
1308                                  strlen(locale_table[i].locale))) {
1309                         cur_charset = locale_table[i].charset;
1310                         return cur_charset;
1311                 } else if ((p = strchr(locale_table[i].locale, '_')) &&
1312                          !strchr(p + 1, '.')) {
1313                         if (strlen(cur_locale) == 2 &&
1314                             !g_ascii_strncasecmp(cur_locale, locale_table[i].locale, 2)) {
1315                                 cur_charset = locale_table[i].charset;
1316                                 return cur_charset;
1317                         }
1318                 }
1319         }
1320
1321         cur_charset = C_AUTO;
1322         return cur_charset;
1323 }
1324
1325 const gchar *conv_get_locale_charset_str(void)
1326 {
1327         static const gchar *codeset = NULL;
1328
1329         if (!codeset)
1330                 codeset = conv_get_charset_str(conv_get_locale_charset());
1331
1332         return codeset ? codeset : CS_INTERNAL;
1333 }
1334
1335 const gchar *conv_get_locale_charset_str_no_utf8(void)
1336 {
1337         static const gchar *codeset = NULL;
1338
1339         if (!codeset)
1340                 codeset = conv_get_charset_str(conv_get_locale_charset_no_utf8());
1341
1342         return codeset ? codeset : CS_INTERNAL;
1343 }
1344
1345 CharSet conv_get_internal_charset(void)
1346 {
1347         return C_INTERNAL;
1348 }
1349
1350 const gchar *conv_get_internal_charset_str(void)
1351 {
1352         return CS_INTERNAL;
1353 }
1354
1355 CharSet conv_get_outgoing_charset(void)
1356 {
1357         static CharSet out_charset = -1;
1358         const gchar *cur_locale;
1359         const gchar *p;
1360         gint i;
1361
1362         if (out_charset != -1)
1363                 return out_charset;
1364
1365         cur_locale = conv_get_current_locale();
1366         if (!cur_locale) {
1367                 out_charset = C_AUTO;
1368                 return out_charset;
1369         }
1370
1371         if (strcasestr(cur_locale, "UTF-8")) {
1372                 out_charset = C_UTF_8;
1373                 return out_charset;
1374         }
1375
1376         if ((p = strcasestr(cur_locale, "@euro")) && p[5] == '\0') {
1377                 out_charset = C_ISO_8859_15;
1378                 return out_charset;
1379         }
1380
1381         for (i = 0; i < sizeof(locale_table) / sizeof(locale_table[0]); i++) {
1382                 const gchar *p;
1383
1384                 if (!g_ascii_strncasecmp(cur_locale, locale_table[i].locale,
1385                                  strlen(locale_table[i].locale))) {
1386                         out_charset = locale_table[i].out_charset;
1387                         break;
1388                 } else if ((p = strchr(locale_table[i].locale, '_')) &&
1389                          !strchr(p + 1, '.')) {
1390                         if (strlen(cur_locale) == 2 &&
1391                             !g_ascii_strncasecmp(cur_locale, locale_table[i].locale, 2)) {
1392                                 out_charset = locale_table[i].out_charset;
1393                                 break;
1394                         }
1395                 }
1396         }
1397
1398         return out_charset;
1399 }
1400
1401 const gchar *conv_get_outgoing_charset_str(void)
1402 {
1403         CharSet out_charset;
1404         const gchar *str;
1405
1406         out_charset = conv_get_outgoing_charset();
1407         str = conv_get_charset_str(out_charset);
1408
1409         return str ? str : CS_UTF_8;
1410 }
1411
1412 gboolean conv_is_multibyte_encoding(CharSet encoding)
1413 {
1414         switch (encoding) {
1415         case C_EUC_JP:
1416         case C_EUC_JP_MS:
1417         case C_EUC_KR:
1418         case C_EUC_TW:
1419         case C_EUC_CN:
1420         case C_ISO_2022_JP:
1421         case C_ISO_2022_JP_2:
1422         case C_ISO_2022_JP_3:
1423         case C_ISO_2022_KR:
1424         case C_ISO_2022_CN:
1425         case C_SHIFT_JIS:
1426         case C_GB2312:
1427         case C_GBK:
1428         case C_BIG5:
1429         case C_UTF_8:
1430         case C_UTF_7:
1431                 return TRUE;
1432         default:
1433                 return FALSE;
1434         }
1435 }
1436
1437 const gchar *conv_get_current_locale(void)
1438 {
1439         const gchar *cur_locale;
1440
1441 #ifdef G_OS_WIN32
1442         cur_locale = g_win32_getlocale();
1443 #else
1444         cur_locale = g_getenv("LC_ALL");
1445         if (!cur_locale) cur_locale = g_getenv("LC_CTYPE");
1446         if (!cur_locale) cur_locale = g_getenv("LANG");
1447         if (!cur_locale) cur_locale = setlocale(LC_CTYPE, NULL);
1448 #endif /* G_OS_WIN32 */
1449
1450         debug_print("current locale: %s\n",
1451                     cur_locale ? cur_locale : "(none)");
1452
1453         return cur_locale;
1454 }
1455
1456 gboolean conv_is_ja_locale(void)
1457 {
1458         static gint is_ja_locale = -1;
1459         const gchar *cur_locale;
1460
1461         if (is_ja_locale != -1)
1462                 return is_ja_locale != 0;
1463
1464         is_ja_locale = 0;
1465         cur_locale = conv_get_current_locale();
1466         if (cur_locale) {
1467                 if (g_ascii_strncasecmp(cur_locale, "ja", 2) == 0)
1468                         is_ja_locale = 1;
1469         }
1470
1471         return is_ja_locale != 0;
1472 }
1473
1474 gchar *conv_unmime_header(const gchar *str, const gchar *default_encoding)
1475 {
1476         gchar buf[BUFFSIZE];
1477
1478         if (is_ascii_str(str))
1479                 return unmime_header(str);
1480
1481         if (default_encoding) {
1482                 gchar *utf8_buf;
1483
1484                 utf8_buf = conv_codeset_strdup
1485                         (str, default_encoding, CS_INTERNAL);
1486                 if (utf8_buf) {
1487                         gchar *decoded_str;
1488
1489                         decoded_str = unmime_header(utf8_buf);
1490                         g_free(utf8_buf);
1491                         return decoded_str;
1492                 }
1493         }
1494
1495         if (conv_is_ja_locale())
1496                 conv_anytodisp(buf, sizeof(buf), str);
1497         else
1498                 conv_localetodisp(buf, sizeof(buf), str);
1499
1500         return unmime_header(buf);
1501 }
1502
1503 #define MAX_LINELEN             76
1504 #define MAX_HARD_LINELEN        996
1505 #define MIMESEP_BEGIN           "=?"
1506 #define MIMESEP_END             "?="
1507
1508 #define LBREAK_IF_REQUIRED(cond, is_plain_text)                         \
1509 {                                                                       \
1510         if (len - (destp - (guchar *)dest) < MAX_LINELEN + 2) {         \
1511                 *destp = '\0';                                          \
1512                 return;                                                 \
1513         }                                                               \
1514                                                                         \
1515         if ((cond) && *srcp) {                                          \
1516                 if (destp > (guchar *)dest && left < MAX_LINELEN - 1) { \
1517                         if (isspace(*(destp - 1)))                      \
1518                                 destp--;                                \
1519                         else if (is_plain_text && isspace(*srcp))       \
1520                                 srcp++;                                 \
1521                         if (*srcp) {                                    \
1522                                 *destp++ = '\n';                        \
1523                                 *destp++ = ' ';                         \
1524                                 left = MAX_LINELEN - 1;                 \
1525                         }                                               \
1526                 }                                                       \
1527         }                                                               \
1528 }
1529
1530 void conv_encode_header_full(gchar *dest, gint len, const gchar *src,
1531                         gint header_len, gboolean addr_field,
1532                         const gchar *out_encoding_)
1533 {
1534         const gchar *cur_encoding;
1535         const gchar *out_encoding;
1536         gint mimestr_len;
1537         gchar *mimesep_enc;
1538         gint left;
1539         const guchar *srcp = src;
1540         guchar *destp = dest;
1541         gboolean use_base64;
1542
1543         g_return_if_fail(g_utf8_validate(src, -1, NULL) == TRUE);
1544
1545         if (MB_CUR_MAX > 1) {
1546                 use_base64 = TRUE;
1547                 mimesep_enc = "?B?";
1548         } else {
1549                 use_base64 = FALSE;
1550                 mimesep_enc = "?Q?";
1551         }
1552
1553         cur_encoding = CS_INTERNAL;
1554
1555         if (out_encoding_)
1556                 out_encoding = out_encoding_;
1557         else
1558                 out_encoding = conv_get_outgoing_charset_str();
1559
1560         if (!strcmp(out_encoding, CS_US_ASCII))
1561                 out_encoding = CS_ISO_8859_1;
1562
1563         mimestr_len = strlen(MIMESEP_BEGIN) + strlen(out_encoding) +
1564                 strlen(mimesep_enc) + strlen(MIMESEP_END);
1565
1566         left = MAX_LINELEN - header_len;
1567
1568         while (*srcp) {
1569                 LBREAK_IF_REQUIRED(left <= 0, TRUE);
1570
1571                 while (isspace(*srcp)) {
1572                         *destp++ = *srcp++;
1573                         left--;
1574                         LBREAK_IF_REQUIRED(left <= 0, TRUE);
1575                 }
1576
1577                 /* output as it is if the next word is ASCII string */
1578                 if (!is_next_nonascii(srcp)) {
1579                         gint word_len;
1580
1581                         word_len = get_next_word_len(srcp);
1582                         LBREAK_IF_REQUIRED(left < word_len, TRUE);
1583                         while (word_len > 0) {
1584                                 LBREAK_IF_REQUIRED(left + (MAX_HARD_LINELEN - MAX_LINELEN) <= 0, TRUE)
1585                                 *destp++ = *srcp++;
1586                                 left--;
1587                                 word_len--;
1588                         }
1589
1590                         continue;
1591                 }
1592
1593                 /* don't include parentheses in encoded strings */
1594                 if (addr_field && (*srcp == '(' || *srcp == ')')) {
1595                         LBREAK_IF_REQUIRED(left < 2, FALSE);
1596                         *destp++ = *srcp++;
1597                         left--;
1598                 }
1599
1600                 while (1) {
1601                         gint mb_len = 0;
1602                         gint cur_len = 0;
1603                         gchar *part_str;
1604                         gchar *out_str;
1605                         gchar *enc_str;
1606                         const guchar *p = srcp;
1607                         gint out_str_len;
1608                         gint out_enc_str_len;
1609                         gint mime_block_len;
1610                         gboolean cont = FALSE;
1611
1612                         while (*p != '\0') {
1613                                 if (isspace(*p) && !is_next_nonascii(p + 1))
1614                                         break;
1615                                 /* don't include parentheses in encoded
1616                                    strings */
1617                                 if (addr_field && (*p == '(' || *p == ')'))
1618                                         break;
1619
1620                                 mb_len = g_utf8_skip[*p];
1621
1622                                 Xstrndup_a(part_str, srcp, cur_len + mb_len, );
1623                                 out_str = conv_codeset_strdup
1624                                         (part_str, cur_encoding, out_encoding);
1625                                 if (!out_str) {
1626                                         if (strict_mode) {
1627                                                 *dest = '\0';
1628                                                 return;
1629                                         } else {
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                                 }
1635                                 out_str_len = strlen(out_str);
1636
1637                                 if (use_base64)
1638                                         out_enc_str_len = B64LEN(out_str_len);
1639                                 else
1640                                         out_enc_str_len =
1641                                                 qp_get_q_encoding_len(out_str);
1642
1643                                 g_free(out_str);
1644
1645                                 if (mimestr_len + out_enc_str_len <= left) {
1646                                         cur_len += mb_len;
1647                                         p += mb_len;
1648                                 } else if (cur_len == 0) {
1649                                         LBREAK_IF_REQUIRED(1, FALSE);
1650                                         continue;
1651                                 } else {
1652                                         cont = TRUE;
1653                                         break;
1654                                 }
1655                         }
1656
1657                         if (cur_len > 0) {
1658                                 Xstrndup_a(part_str, srcp, cur_len, );
1659                                 out_str = conv_codeset_strdup
1660                                         (part_str, cur_encoding, out_encoding);
1661                                 if (!out_str) {
1662                                         g_warning("conv_encode_header(): code conversion failed\n");
1663                                         conv_unreadable_8bit(part_str);
1664                                         out_str = g_strdup(part_str);
1665                                 }
1666                                 out_str_len = strlen(out_str);
1667
1668                                 if (use_base64)
1669                                         out_enc_str_len = B64LEN(out_str_len);
1670                                 else
1671                                         out_enc_str_len =
1672                                                 qp_get_q_encoding_len(out_str);
1673
1674                                 Xalloca(enc_str, out_enc_str_len + 1, );
1675                                 if (use_base64)
1676                                         base64_encode(enc_str, out_str, out_str_len);
1677                                 else
1678                                         qp_q_encode(enc_str, out_str);
1679
1680                                 g_free(out_str);
1681
1682                                 /* output MIME-encoded string block */
1683                                 mime_block_len = mimestr_len + strlen(enc_str);
1684                                 g_snprintf(destp, mime_block_len + 1,
1685                                            MIMESEP_BEGIN "%s%s%s" MIMESEP_END,
1686                                            out_encoding, mimesep_enc, enc_str);
1687                                 destp += mime_block_len;
1688                                 srcp += cur_len;
1689
1690                                 left -= mime_block_len;
1691                         }
1692
1693                         LBREAK_IF_REQUIRED(cont, FALSE);
1694
1695                         if (cur_len == 0)
1696                                 break;
1697                 }
1698         }
1699
1700         *destp = '\0';
1701 }
1702
1703 void conv_encode_header(gchar *dest, gint len, const gchar *src,
1704                         gint header_len, gboolean addr_field)
1705 {
1706         conv_encode_header_full(dest,len,src,header_len,addr_field,NULL);
1707 }
1708
1709 #undef LBREAK_IF_REQUIRED
1710 gchar *conv_filename_from_utf8(const gchar *utf8_file)
1711 {
1712         gchar *fs_file;
1713         GError *error = NULL;
1714
1715         fs_file = g_filename_from_utf8(utf8_file, -1, NULL, NULL, &error);
1716         if (error) {
1717                 g_warning("failed to convert encoding of file name: %s\n",
1718                           error->message);
1719                 g_error_free(error);
1720         }
1721         if (!fs_file)
1722                 fs_file = g_strdup(utf8_file);
1723
1724         return fs_file;
1725 }
1726
1727 gchar *conv_filename_to_utf8(const gchar *fs_file)
1728 {
1729         gchar *utf8_file = NULL;
1730         GError *error = NULL;
1731
1732         utf8_file = g_filename_to_utf8(fs_file, -1, NULL, NULL, &error);
1733         if (error) {
1734                 g_warning("failed to convert encoding of file name: %s\n",
1735                           error->message);
1736                 g_error_free(error);
1737         }
1738
1739         if (!utf8_file || !g_utf8_validate(utf8_file, -1, NULL)) {
1740                 g_free(utf8_file);
1741                 utf8_file = g_strdup(fs_file);
1742                 conv_unreadable_8bit(utf8_file);
1743         }
1744
1745         return utf8_file;
1746 }