2007-01-23 [colin] 2.7.1cvs52
[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 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 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 CharSet conv_get_locale_charset                  (void);
118 static CharSet conv_get_outgoing_charset                (void);
119 static CharSet conv_guess_ja_encoding(const gchar *str);
120 static gboolean conv_is_ja_locale                       (void);
121
122 static void conv_jistoeuc(gchar *outbuf, gint outlen, const gchar *inbuf);
123 static void conv_euctojis(gchar *outbuf, gint outlen, const gchar *inbuf);
124 static void conv_sjistoeuc(gchar *outbuf, gint outlen, const gchar *inbuf);
125
126 static void conv_jistoutf8(gchar *outbuf, gint outlen, const gchar *inbuf);
127 static void conv_sjistoutf8(gchar *outbuf, gint outlen, const gchar *inbuf);
128 static void conv_euctoutf8(gchar *outbuf, gint outlen, const gchar *inbuf);
129 static void conv_anytoutf8(gchar *outbuf, gint outlen, const gchar *inbuf);
130
131 static void conv_utf8toeuc(gchar *outbuf, gint outlen, const gchar *inbuf);
132 static void conv_utf8tojis(gchar *outbuf, gint outlen, const gchar *inbuf);
133
134 static void conv_unreadable_8bit(gchar *str);
135
136 static void conv_jistodisp(gchar *outbuf, gint outlen, const gchar *inbuf);
137 static void conv_sjistodisp(gchar *outbuf, gint outlen, const gchar *inbuf);
138 static void conv_euctodisp(gchar *outbuf, gint outlen, const gchar *inbuf);
139
140 static void conv_anytodisp(gchar *outbuf, gint outlen, const gchar *inbuf);
141 static void conv_ustodisp(gchar *outbuf, gint outlen, const gchar *inbuf);
142 static void conv_noconv(gchar *outbuf, gint outlen, const gchar *inbuf);
143
144 static gboolean strict_mode = FALSE;
145
146 void codeconv_set_strict(gboolean mode)
147 {
148         strict_mode = mode;
149 }
150
151 static void conv_jistoeuc(gchar *outbuf, gint outlen, const gchar *inbuf)
152 {
153         const guchar *in = inbuf;
154         guchar *out = outbuf;
155         JISState state = JIS_ASCII;
156
157         while (*in != '\0') {
158                 if (*in == ESC) {
159                         in++;
160                         if (*in == '$') {
161                                 if (*(in + 1) == '@' || *(in + 1) == 'B') {
162                                         state = JIS_KANJI;
163                                         in += 2;
164                                 } else if (*(in + 1) == '(' &&
165                                            *(in + 2) == 'D') {
166                                         state = JIS_AUXKANJI;
167                                         in += 3;
168                                 } else {
169                                         /* unknown escape sequence */
170                                         state = JIS_ASCII;
171                                 }
172                         } else if (*in == '(') {
173                                 if (*(in + 1) == 'B' || *(in + 1) == 'J') {
174                                         state = JIS_ASCII;
175                                         in += 2;
176                                 } else if (*(in + 1) == 'I') {
177                                         state = JIS_HWKANA;
178                                         in += 2;
179                                 } else {
180                                         /* unknown escape sequence */
181                                         state = JIS_ASCII;
182                                 }
183                         } else {
184                                 /* unknown escape sequence */
185                                 state = JIS_ASCII;
186                         }
187                 } else if (*in == 0x0e) {
188                         state = JIS_HWKANA;
189                         in++;
190                 } else if (*in == 0x0f) {
191                         state = JIS_ASCII;
192                         in++;
193                 } else {
194                         switch (state) {
195                         case JIS_ASCII:
196                                 *out++ = *in++;
197                                 break;
198                         case JIS_KANJI:
199                                 *out++ = *in++ | 0x80;
200                                 if (*in == '\0') break;
201                                 *out++ = *in++ | 0x80;
202                                 break;
203                         case JIS_HWKANA:
204                                 *out++ = 0x8e;
205                                 *out++ = *in++ | 0x80;
206                                 break;
207                         case JIS_AUXKANJI:
208                                 *out++ = 0x8f;
209                                 *out++ = *in++ | 0x80;
210                                 if (*in == '\0') break;
211                                 *out++ = *in++ | 0x80;
212                                 break;
213                         }
214                 }
215         }
216
217         *out = '\0';
218 }
219
220 #define JIS_HWDAKUTEN           0x5e
221 #define JIS_HWHANDAKUTEN        0x5f
222
223 static gint conv_jis_hantozen(guchar *outbuf, guchar jis_code, guchar sound_sym)
224 {
225         static guint16 h2z_tbl[] = {
226                 /* 0x20 - 0x2f */
227                 0x0000, 0x2123, 0x2156, 0x2157, 0x2122, 0x2126, 0x2572, 0x2521,
228                 0x2523, 0x2525, 0x2527, 0x2529, 0x2563, 0x2565, 0x2567, 0x2543,
229                 /* 0x30 - 0x3f */
230                 0x213c, 0x2522, 0x2524, 0x2526, 0x2528, 0x252a, 0x252b, 0x252d,
231                 0x252f, 0x2531, 0x2533, 0x2535, 0x2537, 0x2539, 0x253b, 0x253d,
232                 /* 0x40 - 0x4f */
233                 0x253f, 0x2541, 0x2544, 0x2546, 0x2548, 0x254a, 0x254b, 0x254c,
234                 0x254d, 0x254e, 0x254f, 0x2552, 0x2555, 0x2558, 0x255b, 0x255e,
235                 /* 0x50 - 0x5f */
236                 0x255f, 0x2560, 0x2561, 0x2562, 0x2564, 0x2566, 0x2568, 0x2569,
237                 0x256a, 0x256b, 0x256c, 0x256d, 0x256f, 0x2573, 0x212b, 0x212c
238         };
239
240         static guint16 dakuten_tbl[] = {
241                 /* 0x30 - 0x3f */
242                 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x252c, 0x252e,
243                 0x2530, 0x2532, 0x2534, 0x2536, 0x2538, 0x253a, 0x253c, 0x253e,
244                 /* 0x40 - 0x4f */
245                 0x2540, 0x2542, 0x2545, 0x2547, 0x2549, 0x0000, 0x0000, 0x0000,
246                 0x0000, 0x0000, 0x2550, 0x2553, 0x2556, 0x2559, 0x255c, 0x0000
247         };
248
249         static guint16 handakuten_tbl[] = {
250                 /* 0x4a - 0x4e */
251                 0x2551, 0x2554, 0x2557, 0x255a, 0x255d
252         };
253
254         guint16 out_code;
255
256         jis_code &= 0x7f;
257         sound_sym &= 0x7f;
258
259         if (jis_code < 0x21 || jis_code > 0x5f)
260                 return 0;
261
262         if (sound_sym == JIS_HWDAKUTEN &&
263             jis_code >= 0x36 && jis_code <= 0x4e) {
264                 out_code = dakuten_tbl[jis_code - 0x30];
265                 if (out_code != 0) {
266                         *outbuf = out_code >> 8;
267                         *(outbuf + 1) = out_code & 0xff;
268                         return 2;
269                 }
270         }
271
272         if (sound_sym == JIS_HWHANDAKUTEN &&
273             jis_code >= 0x4a && jis_code <= 0x4e) {
274                 out_code = handakuten_tbl[jis_code - 0x4a];
275                 *outbuf = out_code >> 8;
276                 *(outbuf + 1) = out_code & 0xff;
277                 return 2;
278         }
279
280         out_code = h2z_tbl[jis_code - 0x20];
281         *outbuf = out_code >> 8;
282         *(outbuf + 1) = out_code & 0xff;
283         return 1;
284 }
285
286 static void conv_euctojis(gchar *outbuf, gint outlen, const gchar *inbuf)
287 {
288         const guchar *in = inbuf;
289         guchar *out = outbuf;
290         JISState state = JIS_ASCII;
291
292         while (*in != '\0') {
293                 if (IS_ASCII(*in)) {
294                         K_OUT();
295                         *out++ = *in++;
296                 } else if (iseuckanji(*in)) {
297                         if (iseuckanji(*(in + 1))) {
298                                 K_IN();
299                                 *out++ = *in++ & 0x7f;
300                                 *out++ = *in++ & 0x7f;
301                         } else {
302                                 K_OUT();
303                                 *out++ = SUBST_CHAR;
304                                 in++;
305                                 if (*in != '\0' && !IS_ASCII(*in)) {
306                                         *out++ = SUBST_CHAR;
307                                         in++;
308                                 }
309                         }
310                 } else if (iseuchwkana1(*in)) {
311                         if (iseuchwkana2(*(in + 1))) {
312                                 if (prefs_common.allow_jisx0201_kana) {
313                                         HW_IN();
314                                         in++;
315                                         *out++ = *in++ & 0x7f;
316                                 } else {
317                                         guchar jis_ch[2];
318                                         gint len;
319
320                                         if (iseuchwkana1(*(in + 2)) &&
321                                             iseuchwkana2(*(in + 3)))
322                                                 len = conv_jis_hantozen
323                                                         (jis_ch,
324                                                          *(in + 1), *(in + 3));
325                                         else
326                                                 len = conv_jis_hantozen
327                                                         (jis_ch,
328                                                          *(in + 1), '\0');
329                                         if (len == 0)
330                                                 in += 2;
331                                         else {
332                                                 K_IN();
333                                                 in += len * 2;
334                                                 *out++ = jis_ch[0];
335                                                 *out++ = jis_ch[1];
336                                         }
337                                 }
338                         } else {
339                                 K_OUT();
340                                 in++;
341                                 if (*in != '\0' && !IS_ASCII(*in)) {
342                                         *out++ = SUBST_CHAR;
343                                         in++;
344                                 }
345                         }
346                 } else if (iseucaux(*in)) {
347                         in++;
348                         if (iseuckanji(*in) && iseuckanji(*(in + 1))) {
349                                 AUX_IN();
350                                 *out++ = *in++ & 0x7f;
351                                 *out++ = *in++ & 0x7f;
352                         } else {
353                                 K_OUT();
354                                 if (*in != '\0' && !IS_ASCII(*in)) {
355                                         *out++ = SUBST_CHAR;
356                                         in++;
357                                         if (*in != '\0' && !IS_ASCII(*in)) {
358                                                 *out++ = SUBST_CHAR;
359                                                 in++;
360                                         }
361                                 }
362                         }
363                 } else {
364                         K_OUT();
365                         *out++ = SUBST_CHAR;
366                         in++;
367                 }
368         }
369
370         K_OUT();
371         *out = '\0';
372 }
373
374 static void conv_sjistoeuc(gchar *outbuf, gint outlen, const gchar *inbuf)
375 {
376         const guchar *in = inbuf;
377         guchar *out = outbuf;
378
379         while (*in != '\0') {
380                 if (IS_ASCII(*in)) {
381                         *out++ = *in++;
382                 } else if (issjiskanji1(*in)) {
383                         if (issjiskanji2(*(in + 1))) {
384                                 guchar out1 = *in;
385                                 guchar out2 = *(in + 1);
386                                 guchar row;
387
388                                 row = out1 < 0xa0 ? 0x70 : 0xb0;
389                                 if (out2 < 0x9f) {
390                                         out1 = (out1 - row) * 2 - 1;
391                                         out2 -= out2 > 0x7f ? 0x20 : 0x1f;
392                                 } else {
393                                         out1 = (out1 - row) * 2;
394                                         out2 -= 0x7e;
395                                 }
396
397                                 *out++ = out1 | 0x80;
398                                 *out++ = out2 | 0x80;
399                                 in += 2;
400                         } else {
401                                 *out++ = SUBST_CHAR;
402                                 in++;
403                                 if (*in != '\0' && !IS_ASCII(*in)) {
404                                         *out++ = SUBST_CHAR;
405                                         in++;
406                                 }
407                         }
408                 } else if (issjishwkana(*in)) {
409                         *out++ = 0x8e;
410                         *out++ = *in++;
411                 } else {
412                         *out++ = SUBST_CHAR;
413                         in++;
414                 }
415         }
416
417         *out = '\0';
418 }
419
420 static void conv_jistoutf8(gchar *outbuf, gint outlen, const gchar *inbuf)
421 {
422         gchar *eucstr;
423
424         Xalloca(eucstr, outlen, return);
425
426         conv_jistoeuc(eucstr, outlen, inbuf);
427         conv_euctoutf8(outbuf, outlen, eucstr);
428 }
429
430 static void conv_sjistoutf8(gchar *outbuf, gint outlen, const gchar *inbuf)
431 {
432         gchar *tmpstr;
433
434         tmpstr = conv_iconv_strdup(inbuf, CS_SHIFT_JIS, CS_UTF_8);
435         if (tmpstr) {
436                 strncpy2(outbuf, tmpstr, outlen);
437                 g_free(tmpstr);
438         } else
439                 strncpy2(outbuf, inbuf, outlen);
440 }
441
442 static void conv_euctoutf8(gchar *outbuf, gint outlen, const gchar *inbuf)
443 {
444         static iconv_t cd = (iconv_t)-1;
445         static gboolean iconv_ok = TRUE;
446         gchar *tmpstr;
447
448         if (cd == (iconv_t)-1) {
449                 if (!iconv_ok) {
450                         strncpy2(outbuf, inbuf, outlen);
451                         return;
452                 }
453                 cd = iconv_open(CS_UTF_8, CS_EUC_JP_MS);
454                 if (cd == (iconv_t)-1) {
455                         cd = iconv_open(CS_UTF_8, CS_EUC_JP);
456                         if (cd == (iconv_t)-1) {
457                                 g_warning("conv_euctoutf8(): %s\n",
458                                           g_strerror(errno));
459                                 iconv_ok = FALSE;
460                                 strncpy2(outbuf, inbuf, outlen);
461                                 return;
462                         }
463                 }
464         }
465
466         tmpstr = conv_iconv_strdup_with_cd(inbuf, cd);
467         if (tmpstr) {
468                 strncpy2(outbuf, tmpstr, outlen);
469                 g_free(tmpstr);
470         } else
471                 strncpy2(outbuf, inbuf, outlen);
472 }
473
474 static void conv_anytoutf8(gchar *outbuf, gint outlen, const gchar *inbuf)
475 {
476         switch (conv_guess_ja_encoding(inbuf)) {
477         case C_ISO_2022_JP:
478                 conv_jistoutf8(outbuf, outlen, inbuf);
479                 break;
480         case C_SHIFT_JIS:
481                 conv_sjistoutf8(outbuf, outlen, inbuf);
482                 break;
483         case C_EUC_JP:
484                 conv_euctoutf8(outbuf, outlen, inbuf);
485                 break;
486         default:
487                 strncpy2(outbuf, inbuf, outlen);
488                 break;
489         }
490 }
491
492 static void conv_utf8toeuc(gchar *outbuf, gint outlen, const gchar *inbuf)
493 {
494         static iconv_t cd = (iconv_t)-1;
495         static gboolean iconv_ok = TRUE;
496         gchar *tmpstr;
497
498         if (cd == (iconv_t)-1) {
499                 if (!iconv_ok) {
500                         strncpy2(outbuf, inbuf, outlen);
501                         return;
502                 }
503                 cd = iconv_open(CS_EUC_JP_MS, CS_UTF_8);
504                 if (cd == (iconv_t)-1) {
505                         cd = iconv_open(CS_EUC_JP, CS_UTF_8);
506                         if (cd == (iconv_t)-1) {
507                                 g_warning("conv_utf8toeuc(): %s\n",
508                                           g_strerror(errno));
509                                 iconv_ok = FALSE;
510                                 strncpy2(outbuf, inbuf, outlen);
511                                 return;
512                         }
513                 }
514         }
515
516         tmpstr = conv_iconv_strdup_with_cd(inbuf, cd);
517         if (tmpstr) {
518                 strncpy2(outbuf, tmpstr, outlen);
519                 g_free(tmpstr);
520         } else
521                 strncpy2(outbuf, inbuf, outlen);
522 }
523
524 static void conv_utf8tojis(gchar *outbuf, gint outlen, const gchar *inbuf)
525 {
526         gchar *eucstr;
527
528         Xalloca(eucstr, outlen, return);
529
530         conv_utf8toeuc(eucstr, outlen, inbuf);
531         conv_euctojis(outbuf, outlen, eucstr);
532 }
533
534 static void conv_unreadable_8bit(gchar *str)
535 {
536         register guchar *p = str;
537
538         while (*p != '\0') {
539                 /* convert CR+LF -> LF */
540                 if (*p == '\r' && *(p + 1) == '\n')
541                         memmove(p, p + 1, strlen(p));
542                 else if (!IS_ASCII(*p)) *p = SUBST_CHAR;
543                 p++;
544         }
545 }
546
547 static CharSet conv_guess_ja_encoding(const gchar *str)
548 {
549         const guchar *p = str;
550         CharSet guessed = C_US_ASCII;
551
552         while (*p != '\0') {
553                 if (*p == ESC && (*(p + 1) == '$' || *(p + 1) == '(')) {
554                         if (guessed == C_US_ASCII)
555                                 return C_ISO_2022_JP;
556                         p += 2;
557                 } else if (IS_ASCII(*p)) {
558                         p++;
559                 } else if (iseuckanji(*p) && iseuckanji(*(p + 1))) {
560                         if (*p >= 0xfd && *p <= 0xfe)
561                                 return C_EUC_JP;
562                         else if (guessed == C_SHIFT_JIS) {
563                                 if ((issjiskanji1(*p) &&
564                                      issjiskanji2(*(p + 1))) ||
565                                     issjishwkana(*p))
566                                         guessed = C_SHIFT_JIS;
567                                 else
568                                         guessed = C_EUC_JP;
569                         } else
570                                 guessed = C_EUC_JP;
571                         p += 2;
572                 } else if (issjiskanji1(*p) && issjiskanji2(*(p + 1))) {
573                         if (iseuchwkana1(*p) && iseuchwkana2(*(p + 1)))
574                                 guessed = C_SHIFT_JIS;
575                         else
576                                 return C_SHIFT_JIS;
577                         p += 2;
578                 } else if (issjishwkana(*p)) {
579                         guessed = C_SHIFT_JIS;
580                         p++;
581                 } else {
582                         p++;
583                 }
584         }
585
586         return guessed;
587 }
588
589 static void conv_jistodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
590 {
591         conv_jistoutf8(outbuf, outlen, inbuf);
592 }
593
594 static void conv_sjistodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
595 {
596         conv_sjistoutf8(outbuf, outlen, inbuf);
597 }
598
599 static void conv_euctodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
600 {
601         conv_euctoutf8(outbuf, outlen, inbuf);
602 }
603
604 void conv_utf8todisp(gchar *outbuf, gint outlen, const gchar *inbuf)
605 {
606         if (g_utf8_validate(inbuf, -1, NULL) == TRUE)
607                 strncpy2(outbuf, inbuf, outlen);
608         else
609                 conv_ustodisp(outbuf, outlen, inbuf);
610 }
611
612 static void conv_anytodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
613 {
614         conv_anytoutf8(outbuf, outlen, inbuf);
615         if (g_utf8_validate(outbuf, -1, NULL) != TRUE)
616                 conv_unreadable_8bit(outbuf);
617 }
618
619 static void conv_ustodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
620 {
621         strncpy2(outbuf, inbuf, outlen);
622         conv_unreadable_8bit(outbuf);
623 }
624
625 void conv_localetodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
626 {
627         gchar *tmpstr;
628
629         codeconv_set_strict(TRUE);
630         tmpstr = conv_iconv_strdup(inbuf, conv_get_locale_charset_str(),
631                                    CS_INTERNAL);
632         codeconv_set_strict(FALSE);
633         if (tmpstr && g_utf8_validate(tmpstr, -1, NULL)) {
634                 strncpy2(outbuf, tmpstr, outlen);
635                 g_free(tmpstr);
636                 return;
637         } else if (tmpstr && !g_utf8_validate(tmpstr, -1, NULL)) {
638                 g_free(tmpstr);
639                 codeconv_set_strict(TRUE);
640                 tmpstr = conv_iconv_strdup(inbuf, 
641                                 conv_get_locale_charset_str_no_utf8(),
642                                 CS_INTERNAL);
643                 codeconv_set_strict(FALSE);
644         }
645         if (tmpstr && g_utf8_validate(tmpstr, -1, NULL)) {
646                 strncpy2(outbuf, tmpstr, outlen);
647                 g_free(tmpstr);
648                 return;
649         } else {
650                 g_free(tmpstr);
651                 conv_utf8todisp(outbuf, outlen, inbuf);
652         }
653 }
654
655 static void conv_noconv(gchar *outbuf, gint outlen, const gchar *inbuf)
656 {
657         strncpy2(outbuf, inbuf, outlen);
658 }
659
660 static const gchar *
661 conv_get_fallback_for_private_encoding(const gchar *encoding)
662 {
663         if (encoding && (encoding[0] == 'X' || encoding[0] == 'x') &&
664             encoding[1] == '-') {
665                 if (!g_ascii_strcasecmp(encoding, CS_X_GBK))
666                         return CS_GBK;
667         }
668
669         return encoding;
670 }
671
672 CodeConverter *conv_code_converter_new(const gchar *src_charset)
673 {
674         CodeConverter *conv;
675
676         src_charset = conv_get_fallback_for_private_encoding(src_charset);
677
678         conv = g_new0(CodeConverter, 1);
679         conv->code_conv_func = conv_get_code_conv_func(src_charset, NULL);
680         conv->charset_str = g_strdup(src_charset);
681         conv->charset = conv_get_charset_from_str(src_charset);
682
683         return conv;
684 }
685
686 void conv_code_converter_destroy(CodeConverter *conv)
687 {
688         g_free(conv->charset_str);
689         g_free(conv);
690 }
691
692 gint conv_convert(CodeConverter *conv, gchar *outbuf, gint outlen,
693                   const gchar *inbuf)
694 {
695         if (conv->code_conv_func != conv_noconv)
696                 conv->code_conv_func(outbuf, outlen, inbuf);
697         else {
698                 gchar *str;
699
700                 str = conv_iconv_strdup(inbuf, conv->charset_str, NULL);
701                 if (!str)
702                         return -1;
703                 else {
704                         strncpy2(outbuf, str, outlen);
705                         g_free(str);
706                 }
707         }
708
709         return 0;
710 }
711
712 gchar *conv_codeset_strdup(const gchar *inbuf,
713                            const gchar *src_code, const gchar *dest_code)
714 {
715         gchar *buf;
716         size_t len;
717         CodeConvFunc conv_func;
718
719         src_code = conv_get_fallback_for_private_encoding(src_code);
720         conv_func = conv_get_code_conv_func(src_code, dest_code);
721         if (conv_func != conv_noconv) {
722                 len = (strlen(inbuf) + 1) * 3;
723                 buf = g_malloc(len);
724                 if (!buf) return NULL;
725
726                 conv_func(buf, len, inbuf);
727                 return g_realloc(buf, strlen(buf) + 1);
728         }
729
730         return conv_iconv_strdup(inbuf, src_code, dest_code);
731 }
732
733 static CodeConvFunc conv_get_code_conv_func(const gchar *src_charset_str,
734                                      const gchar *dest_charset_str)
735 {
736         CodeConvFunc code_conv = conv_noconv;
737         CharSet src_charset;
738         CharSet dest_charset;
739
740         if (!src_charset_str)
741                 src_charset = conv_get_locale_charset();
742         else
743                 src_charset = conv_get_charset_from_str(src_charset_str);
744
745         /* auto detection mode */
746         if (!src_charset_str && !dest_charset_str) {
747                 if (conv_is_ja_locale())
748                         return conv_anytodisp;
749                 else
750                         return conv_noconv;
751         }
752
753         dest_charset = conv_get_charset_from_str(dest_charset_str);
754
755         if (dest_charset == C_US_ASCII)
756                 return conv_ustodisp;
757
758         switch (src_charset) {
759         case C_US_ASCII:
760         case C_ISO_8859_1:
761         case C_ISO_8859_2:
762         case C_ISO_8859_3:
763         case C_ISO_8859_4:
764         case C_ISO_8859_5:
765         case C_ISO_8859_6:
766         case C_ISO_8859_7:
767         case C_ISO_8859_8:
768         case C_ISO_8859_9:
769         case C_ISO_8859_10:
770         case C_ISO_8859_11:
771         case C_ISO_8859_13:
772         case C_ISO_8859_14:
773         case C_ISO_8859_15:
774                 break;
775         case C_ISO_2022_JP:
776         case C_ISO_2022_JP_2:
777         case C_ISO_2022_JP_3:
778                 if (dest_charset == C_AUTO)
779                         code_conv = conv_jistodisp;
780                 else if (dest_charset == C_EUC_JP)
781                         code_conv = conv_jistoeuc;
782                 else if (dest_charset == C_UTF_8)
783                         code_conv = conv_jistoutf8;
784                 break;
785         case C_SHIFT_JIS:
786                 if (dest_charset == C_AUTO)
787                         code_conv = conv_sjistodisp;
788                 else if (dest_charset == C_EUC_JP)
789                         code_conv = conv_sjistoeuc;
790                 else if (dest_charset == C_UTF_8)
791                         code_conv = conv_sjistoutf8;
792                 break;
793         case C_EUC_JP:
794                 if (dest_charset == C_AUTO)
795                         code_conv = conv_euctodisp;
796                 else if (dest_charset == C_ISO_2022_JP   ||
797                          dest_charset == C_ISO_2022_JP_2 ||
798                          dest_charset == C_ISO_2022_JP_3)
799                         code_conv = conv_euctojis;
800                 else if (dest_charset == C_UTF_8)
801                         code_conv = conv_euctoutf8;
802                 break;
803         case C_UTF_8:
804                 if (dest_charset == C_EUC_JP)
805                         code_conv = conv_utf8toeuc;
806                 else if (dest_charset == C_ISO_2022_JP   ||
807                          dest_charset == C_ISO_2022_JP_2 ||
808                          dest_charset == C_ISO_2022_JP_3)
809                         code_conv = conv_utf8tojis;
810                 break;
811         default:
812                 break;
813         }
814
815         return code_conv;
816 }
817
818 gchar *conv_iconv_strdup(const gchar *inbuf,
819                          const gchar *src_code, const gchar *dest_code)
820 {
821         iconv_t cd;
822         gchar *outbuf;
823
824         if (!src_code && !dest_code && 
825             g_utf8_validate(inbuf, -1, NULL))
826                 return g_strdup(inbuf);
827
828         if (!src_code)
829                 src_code = conv_get_outgoing_charset_str();
830         if (!dest_code)
831                 dest_code = CS_INTERNAL;
832
833         /* don't convert if src and dest codeset are identical */
834         if (!strcasecmp(src_code, dest_code))
835                 return g_strdup(inbuf);
836
837         /* don't convert if dest codeset is US-ASCII */
838         if (!strcasecmp(src_code, CS_US_ASCII))
839                 return g_strdup(inbuf);
840
841         /* don't convert if dest codeset is US-ASCII */
842         if (!strcasecmp(dest_code, CS_US_ASCII))
843                 return g_strdup(inbuf);
844
845         cd = iconv_open(dest_code, src_code);
846         if (cd == (iconv_t)-1)
847                 return NULL;
848
849         outbuf = conv_iconv_strdup_with_cd(inbuf, cd);
850
851         iconv_close(cd);
852
853         return outbuf;
854 }
855
856 gchar *conv_iconv_strdup_with_cd(const gchar *inbuf, iconv_t cd)
857 {
858         const gchar *inbuf_p;
859         gchar *outbuf;
860         gchar *outbuf_p;
861         size_t in_size;
862         size_t in_left;
863         size_t out_size;
864         size_t out_left;
865         size_t n_conv;
866         size_t len;
867
868         inbuf_p = inbuf;
869         in_size = strlen(inbuf);
870         in_left = in_size;
871         out_size = (in_size + 1) * 2;
872         outbuf = g_malloc(out_size);
873         outbuf_p = outbuf;
874         out_left = out_size;
875
876 #define EXPAND_BUF()                            \
877 {                                               \
878         len = outbuf_p - outbuf;                \
879         out_size *= 2;                          \
880         outbuf = g_realloc(outbuf, out_size);   \
881         outbuf_p = outbuf + len;                \
882         out_left = out_size - len;              \
883 }
884
885         while ((n_conv = iconv(cd, (ICONV_CONST gchar **)&inbuf_p, &in_left,
886                                &outbuf_p, &out_left)) == (size_t)-1) {
887                 if (EILSEQ == errno) {
888                         if (strict_mode) {
889                                 g_free(outbuf);
890                                 return NULL;
891                         }
892                         //g_print("iconv(): at %d: %s\n", in_size - in_left, g_strerror(errno));
893                         inbuf_p++;
894                         in_left--;
895                         if (out_left == 0) {
896                                 EXPAND_BUF();
897                         }
898                         *outbuf_p++ = SUBST_CHAR;
899                         out_left--;
900                 } else if (EINVAL == errno) {
901                         break;
902                 } else if (E2BIG == errno) {
903                         EXPAND_BUF();
904                 } else {
905                         g_warning("conv_iconv_strdup(): %s\n",
906                                   g_strerror(errno));
907                         break;
908                 }
909         }
910
911         while ((n_conv = iconv(cd, NULL, NULL, &outbuf_p, &out_left)) ==
912                (size_t)-1) {
913                 if (E2BIG == errno) {
914                         EXPAND_BUF();
915                 } else {
916                         g_warning("conv_iconv_strdup(): %s\n",
917                                   g_strerror(errno));
918                         break;
919                 }
920         }
921
922 #undef EXPAND_BUF
923
924         len = outbuf_p - outbuf;
925         outbuf = g_realloc(outbuf, len + 1);
926         outbuf[len] = '\0';
927
928         return outbuf;
929 }
930
931 static const struct {
932         CharSet charset;
933         gchar *const name;
934 } charsets[] = {
935         {C_US_ASCII,            CS_US_ASCII},
936         {C_US_ASCII,            CS_ANSI_X3_4_1968},
937         {C_UTF_8,               CS_UTF_8},
938         {C_UTF_7,               CS_UTF_7},
939         {C_ISO_8859_1,          CS_ISO_8859_1},
940         {C_ISO_8859_2,          CS_ISO_8859_2},
941         {C_ISO_8859_3,          CS_ISO_8859_3},
942         {C_ISO_8859_4,          CS_ISO_8859_4},
943         {C_ISO_8859_5,          CS_ISO_8859_5},
944         {C_ISO_8859_6,          CS_ISO_8859_6},
945         {C_ISO_8859_7,          CS_ISO_8859_7},
946         {C_ISO_8859_8,          CS_ISO_8859_8},
947         {C_ISO_8859_9,          CS_ISO_8859_9},
948         {C_ISO_8859_10,         CS_ISO_8859_10},
949         {C_ISO_8859_11,         CS_ISO_8859_11},
950         {C_ISO_8859_13,         CS_ISO_8859_13},
951         {C_ISO_8859_14,         CS_ISO_8859_14},
952         {C_ISO_8859_15,         CS_ISO_8859_15},
953         {C_BALTIC,              CS_BALTIC},
954         {C_CP1250,              CS_CP1250},
955         {C_CP1251,              CS_CP1251},
956         {C_CP1252,              CS_CP1252},
957         {C_CP1253,              CS_CP1253},
958         {C_CP1254,              CS_CP1254},
959         {C_CP1255,              CS_CP1255},
960         {C_CP1256,              CS_CP1256},
961         {C_CP1257,              CS_CP1257},
962         {C_CP1258,              CS_CP1258},
963         {C_WINDOWS_1250,        CS_WINDOWS_1250},
964         {C_WINDOWS_1251,        CS_WINDOWS_1251},
965         {C_WINDOWS_1252,        CS_WINDOWS_1252},
966         {C_WINDOWS_1253,        CS_WINDOWS_1253},
967         {C_WINDOWS_1254,        CS_WINDOWS_1254},
968         {C_WINDOWS_1255,        CS_WINDOWS_1255},
969         {C_WINDOWS_1256,        CS_WINDOWS_1256},
970         {C_WINDOWS_1257,        CS_WINDOWS_1257},
971         {C_WINDOWS_1258,        CS_WINDOWS_1258},
972         {C_KOI8_R,              CS_KOI8_R},
973         {C_KOI8_T,              CS_KOI8_T},
974         {C_KOI8_U,              CS_KOI8_U},
975         {C_ISO_2022_JP,         CS_ISO_2022_JP},
976         {C_ISO_2022_JP_2,       CS_ISO_2022_JP_2},
977         {C_ISO_2022_JP_3,       CS_ISO_2022_JP_3},
978         {C_EUC_JP,              CS_EUC_JP},
979         {C_EUC_JP,              CS_EUCJP},
980         {C_EUC_JP_MS,           CS_EUC_JP_MS},
981         {C_SHIFT_JIS,           CS_SHIFT_JIS},
982         {C_SHIFT_JIS,           CS_SHIFT__JIS},
983         {C_SHIFT_JIS,           CS_SJIS},
984         {C_ISO_2022_KR,         CS_ISO_2022_KR},
985         {C_EUC_KR,              CS_EUC_KR},
986         {C_ISO_2022_CN,         CS_ISO_2022_CN},
987         {C_EUC_CN,              CS_EUC_CN},
988         {C_GB2312,              CS_GB2312},
989         {C_GBK,                 CS_GBK},
990         {C_EUC_TW,              CS_EUC_TW},
991         {C_BIG5,                CS_BIG5},
992         {C_BIG5_HKSCS,          CS_BIG5_HKSCS},
993         {C_TIS_620,             CS_TIS_620},
994         {C_WINDOWS_874,         CS_WINDOWS_874},
995         {C_GEORGIAN_PS,         CS_GEORGIAN_PS},
996         {C_TCVN5712_1,          CS_TCVN5712_1},
997 };
998
999 static const struct {
1000         gchar *const locale;
1001         CharSet charset;
1002         CharSet out_charset;
1003 } locale_table[] = {
1004         {"ja_JP.eucJP"  , C_EUC_JP      , C_ISO_2022_JP},
1005         {"ja_JP.EUC-JP" , C_EUC_JP      , C_ISO_2022_JP},
1006         {"ja_JP.EUC"    , C_EUC_JP      , C_ISO_2022_JP},
1007         {"ja_JP.ujis"   , C_EUC_JP      , C_ISO_2022_JP},
1008         {"ja_JP.SJIS"   , C_SHIFT_JIS   , C_ISO_2022_JP},
1009         {"ja_JP.JIS"    , C_ISO_2022_JP , C_ISO_2022_JP},
1010 #ifdef G_OS_WIN32
1011         {"ja_JP"        , C_SHIFT_JIS   , C_ISO_2022_JP},
1012 #else
1013         {"ja_JP"        , C_EUC_JP      , C_ISO_2022_JP},
1014 #endif
1015         {"ko_KR.EUC-KR" , C_EUC_KR      , C_EUC_KR},
1016         {"ko_KR"        , C_EUC_KR      , C_EUC_KR},
1017         {"zh_CN.GB2312" , C_GB2312      , C_GB2312},
1018         {"zh_CN.GBK"    , C_GBK         , C_GBK},
1019         {"zh_CN"        , C_GB2312      , C_GB2312},
1020         {"zh_HK"        , C_BIG5_HKSCS  , C_BIG5_HKSCS},
1021         {"zh_TW.eucTW"  , C_EUC_TW      , C_BIG5},
1022         {"zh_TW.EUC-TW" , C_EUC_TW      , C_BIG5},
1023         {"zh_TW.Big5"   , C_BIG5        , C_BIG5},
1024         {"zh_TW"        , C_BIG5        , C_BIG5},
1025
1026         {"ru_RU.KOI8-R" , C_KOI8_R      , C_KOI8_R},
1027         {"ru_RU.KOI8R"  , C_KOI8_R      , C_KOI8_R},
1028         {"ru_RU.CP1251" , C_WINDOWS_1251, C_KOI8_R},
1029         {"ru_RU"        , C_ISO_8859_5  , C_KOI8_R},
1030         {"tg_TJ"        , C_KOI8_T      , C_KOI8_T},
1031         {"ru_UA"        , C_KOI8_U      , C_KOI8_U},
1032         {"uk_UA.CP1251" , C_WINDOWS_1251, C_KOI8_U},
1033         {"uk_UA"        , C_KOI8_U      , C_KOI8_U},
1034
1035         {"be_BY"        , C_WINDOWS_1251, C_WINDOWS_1251},
1036         {"bg_BG"        , C_WINDOWS_1251, C_WINDOWS_1251},
1037
1038         {"yi_US"        , C_WINDOWS_1255, C_WINDOWS_1255},
1039
1040         {"af_ZA"        , C_ISO_8859_1  , C_ISO_8859_1},
1041         {"br_FR"        , C_ISO_8859_1  , C_ISO_8859_1},
1042         {"ca_ES"        , C_ISO_8859_1  , C_ISO_8859_1},
1043         {"da_DK"        , C_ISO_8859_1  , C_ISO_8859_1},
1044         {"de_AT"        , C_ISO_8859_1  , C_ISO_8859_1},
1045         {"de_BE"        , C_ISO_8859_1  , C_ISO_8859_1},
1046         {"de_CH"        , C_ISO_8859_1  , C_ISO_8859_1},
1047         {"de_DE"        , C_ISO_8859_1  , C_ISO_8859_1},
1048         {"de_LU"        , C_ISO_8859_1  , C_ISO_8859_1},
1049         {"en_AU"        , C_ISO_8859_1  , C_ISO_8859_1},
1050         {"en_BW"        , C_ISO_8859_1  , C_ISO_8859_1},
1051         {"en_CA"        , C_ISO_8859_1  , C_ISO_8859_1},
1052         {"en_DK"        , C_ISO_8859_1  , C_ISO_8859_1},
1053         {"en_GB"        , C_ISO_8859_1  , C_ISO_8859_1},
1054         {"en_HK"        , C_ISO_8859_1  , C_ISO_8859_1},
1055         {"en_IE"        , C_ISO_8859_1  , C_ISO_8859_1},
1056         {"en_NZ"        , C_ISO_8859_1  , C_ISO_8859_1},
1057         {"en_PH"        , C_ISO_8859_1  , C_ISO_8859_1},
1058         {"en_SG"        , C_ISO_8859_1  , C_ISO_8859_1},
1059         {"en_US"        , C_ISO_8859_1  , C_ISO_8859_1},
1060         {"en_ZA"        , C_ISO_8859_1  , C_ISO_8859_1},
1061         {"en_ZW"        , C_ISO_8859_1  , C_ISO_8859_1},
1062         {"es_AR"        , C_ISO_8859_1  , C_ISO_8859_1},
1063         {"es_BO"        , C_ISO_8859_1  , C_ISO_8859_1},
1064         {"es_CL"        , C_ISO_8859_1  , C_ISO_8859_1},
1065         {"es_CO"        , C_ISO_8859_1  , C_ISO_8859_1},
1066         {"es_CR"        , C_ISO_8859_1  , C_ISO_8859_1},
1067         {"es_DO"        , C_ISO_8859_1  , C_ISO_8859_1},
1068         {"es_EC"        , C_ISO_8859_1  , C_ISO_8859_1},
1069         {"es_ES"        , C_ISO_8859_1  , C_ISO_8859_1},
1070         {"es_GT"        , C_ISO_8859_1  , C_ISO_8859_1},
1071         {"es_HN"        , C_ISO_8859_1  , C_ISO_8859_1},
1072         {"es_MX"        , C_ISO_8859_1  , C_ISO_8859_1},
1073         {"es_NI"        , C_ISO_8859_1  , C_ISO_8859_1},
1074         {"es_PA"        , C_ISO_8859_1  , C_ISO_8859_1},
1075         {"es_PE"        , C_ISO_8859_1  , C_ISO_8859_1},
1076         {"es_PR"        , C_ISO_8859_1  , C_ISO_8859_1},
1077         {"es_PY"        , C_ISO_8859_1  , C_ISO_8859_1},
1078         {"es_SV"        , C_ISO_8859_1  , C_ISO_8859_1},
1079         {"es_US"        , C_ISO_8859_1  , C_ISO_8859_1},
1080         {"es_UY"        , C_ISO_8859_1  , C_ISO_8859_1},
1081         {"es_VE"        , C_ISO_8859_1  , C_ISO_8859_1},
1082         {"et_EE"        , C_ISO_8859_1  , C_ISO_8859_1},
1083         {"eu_ES"        , C_ISO_8859_1  , C_ISO_8859_1},
1084         {"fi_FI"        , C_ISO_8859_1  , C_ISO_8859_1},
1085         {"fo_FO"        , C_ISO_8859_1  , C_ISO_8859_1},
1086         {"fr_BE"        , C_ISO_8859_1  , C_ISO_8859_1},
1087         {"fr_CA"        , C_ISO_8859_1  , C_ISO_8859_1},
1088         {"fr_CH"        , C_ISO_8859_1  , C_ISO_8859_1},
1089         {"fr_FR"        , C_ISO_8859_1  , C_ISO_8859_1},
1090         {"fr_LU"        , C_ISO_8859_1  , C_ISO_8859_1},
1091         {"ga_IE"        , C_ISO_8859_1  , C_ISO_8859_1},
1092         {"gl_ES"        , C_ISO_8859_1  , C_ISO_8859_1},
1093         {"gv_GB"        , C_ISO_8859_1  , C_ISO_8859_1},
1094         {"id_ID"        , C_ISO_8859_1  , C_ISO_8859_1},
1095         {"is_IS"        , C_ISO_8859_1  , C_ISO_8859_1},
1096         {"it_CH"        , C_ISO_8859_1  , C_ISO_8859_1},
1097         {"it_IT"        , C_ISO_8859_1  , C_ISO_8859_1},
1098         {"kl_GL"        , C_ISO_8859_1  , C_ISO_8859_1},
1099         {"kw_GB"        , C_ISO_8859_1  , C_ISO_8859_1},
1100         {"ms_MY"        , C_ISO_8859_1  , C_ISO_8859_1},
1101         {"nl_BE"        , C_ISO_8859_1  , C_ISO_8859_1},
1102         {"nl_NL"        , C_ISO_8859_1  , C_ISO_8859_1},
1103         {"nn_NO"        , C_ISO_8859_1  , C_ISO_8859_1},
1104         {"no_NO"        , C_ISO_8859_1  , C_ISO_8859_1},
1105         {"oc_FR"        , C_ISO_8859_1  , C_ISO_8859_1},
1106         {"pt_BR"        , C_ISO_8859_1  , C_ISO_8859_1},
1107         {"pt_PT"        , C_ISO_8859_1  , C_ISO_8859_1},
1108         {"sq_AL"        , C_ISO_8859_1  , C_ISO_8859_1},
1109         {"sv_FI"        , C_ISO_8859_1  , C_ISO_8859_1},
1110         {"sv_SE"        , C_ISO_8859_1  , C_ISO_8859_1},
1111         {"tl_PH"        , C_ISO_8859_1  , C_ISO_8859_1},
1112         {"uz_UZ"        , C_ISO_8859_1  , C_ISO_8859_1},
1113         {"wa_BE"        , C_ISO_8859_1  , C_ISO_8859_1},
1114
1115         {"bs_BA"        , C_ISO_8859_2  , C_ISO_8859_2},
1116         {"cs_CZ"        , C_ISO_8859_2  , C_ISO_8859_2},
1117         {"hr_HR"        , C_ISO_8859_2  , C_ISO_8859_2},
1118         {"hu_HU"        , C_ISO_8859_2  , C_ISO_8859_2},
1119         {"pl_PL"        , C_ISO_8859_2  , C_ISO_8859_2},
1120         {"ro_RO"        , C_ISO_8859_2  , C_ISO_8859_2},
1121         {"sk_SK"        , C_ISO_8859_2  , C_ISO_8859_2},
1122         {"sl_SI"        , C_ISO_8859_2  , C_ISO_8859_2},
1123
1124         {"sr_YU@cyrillic"       , C_ISO_8859_5  , C_ISO_8859_5},
1125         {"sr_YU"                , C_ISO_8859_2  , C_ISO_8859_2},
1126
1127         {"mt_MT"                , C_ISO_8859_3  , C_ISO_8859_3},
1128
1129         {"lt_LT.iso88594"       , C_ISO_8859_4  , C_ISO_8859_4},
1130         {"lt_LT.ISO8859-4"      , C_ISO_8859_4  , C_ISO_8859_4},
1131         {"lt_LT.ISO_8859-4"     , C_ISO_8859_4  , C_ISO_8859_4},
1132         {"lt_LT"                , C_ISO_8859_13 , C_ISO_8859_13},
1133
1134         {"mk_MK"        , C_ISO_8859_5  , C_ISO_8859_5},
1135
1136         {"ar_AE"        , C_ISO_8859_6  , C_ISO_8859_6},
1137         {"ar_BH"        , C_ISO_8859_6  , C_ISO_8859_6},
1138         {"ar_DZ"        , C_ISO_8859_6  , C_ISO_8859_6},
1139         {"ar_EG"        , C_ISO_8859_6  , C_ISO_8859_6},
1140         {"ar_IQ"        , C_ISO_8859_6  , C_ISO_8859_6},
1141         {"ar_JO"        , C_ISO_8859_6  , C_ISO_8859_6},
1142         {"ar_KW"        , C_ISO_8859_6  , C_ISO_8859_6},
1143         {"ar_LB"        , C_ISO_8859_6  , C_ISO_8859_6},
1144         {"ar_LY"        , C_ISO_8859_6  , C_ISO_8859_6},
1145         {"ar_MA"        , C_ISO_8859_6  , C_ISO_8859_6},
1146         {"ar_OM"        , C_ISO_8859_6  , C_ISO_8859_6},
1147         {"ar_QA"        , C_ISO_8859_6  , C_ISO_8859_6},
1148         {"ar_SA"        , C_ISO_8859_6  , C_ISO_8859_6},
1149         {"ar_SD"        , C_ISO_8859_6  , C_ISO_8859_6},
1150         {"ar_SY"        , C_ISO_8859_6  , C_ISO_8859_6},
1151         {"ar_TN"        , C_ISO_8859_6  , C_ISO_8859_6},
1152         {"ar_YE"        , C_ISO_8859_6  , C_ISO_8859_6},
1153
1154         {"el_GR"        , C_ISO_8859_7  , C_ISO_8859_7},
1155         {"he_IL"        , C_ISO_8859_8  , C_ISO_8859_8},
1156         {"iw_IL"        , C_ISO_8859_8  , C_ISO_8859_8},
1157         {"tr_TR"        , C_ISO_8859_9  , C_ISO_8859_9},
1158
1159         {"lv_LV"        , C_ISO_8859_13 , C_ISO_8859_13},
1160         {"mi_NZ"        , C_ISO_8859_13 , C_ISO_8859_13},
1161
1162         {"cy_GB"        , C_ISO_8859_14 , C_ISO_8859_14},
1163
1164         {"ar_IN"        , C_UTF_8       , C_UTF_8},
1165         {"en_IN"        , C_UTF_8       , C_UTF_8},
1166         {"se_NO"        , C_UTF_8       , C_UTF_8},
1167         {"ta_IN"        , C_UTF_8       , C_UTF_8},
1168         {"te_IN"        , C_UTF_8       , C_UTF_8},
1169         {"ur_PK"        , C_UTF_8       , C_UTF_8},
1170
1171         {"th_TH"        , C_TIS_620     , C_TIS_620},
1172         /* {"th_TH"     , C_WINDOWS_874}, */
1173         /* {"th_TH"     , C_ISO_8859_11}, */
1174
1175         {"ka_GE"        , C_GEORGIAN_PS , C_GEORGIAN_PS},
1176         {"vi_VN.TCVN"   , C_TCVN5712_1  , C_TCVN5712_1},
1177
1178         {"C"                    , C_US_ASCII    , C_US_ASCII},
1179         {"POSIX"                , C_US_ASCII    , C_US_ASCII},
1180         {"ANSI_X3.4-1968"       , C_US_ASCII    , C_US_ASCII},
1181 };
1182
1183 static GHashTable *conv_get_charset_to_str_table(void)
1184 {
1185         static GHashTable *table;
1186         gint i;
1187
1188         if (table)
1189                 return table;
1190
1191         table = g_hash_table_new(NULL, g_direct_equal);
1192
1193         for (i = 0; i < sizeof(charsets) / sizeof(charsets[0]); i++) {
1194                 if (g_hash_table_lookup(table, GUINT_TO_POINTER(charsets[i].charset))
1195                     == NULL) {
1196                         g_hash_table_insert
1197                                 (table, GUINT_TO_POINTER(charsets[i].charset),
1198                                  charsets[i].name);
1199                 }
1200         }
1201
1202         return table;
1203 }
1204
1205 static GHashTable *conv_get_charset_from_str_table(void)
1206 {
1207         static GHashTable *table;
1208         gint i;
1209
1210         if (table)
1211                 return table;
1212
1213         table = g_hash_table_new(str_case_hash, str_case_equal);
1214
1215         for (i = 0; i < sizeof(charsets) / sizeof(charsets[0]); i++) {
1216                 g_hash_table_insert(table, charsets[i].name,
1217                                     GUINT_TO_POINTER(charsets[i].charset));
1218         }
1219
1220         return table;
1221 }
1222
1223 const gchar *conv_get_charset_str(CharSet charset)
1224 {
1225         GHashTable *table;
1226
1227         table = conv_get_charset_to_str_table();
1228         return g_hash_table_lookup(table, GUINT_TO_POINTER(charset));
1229 }
1230
1231 CharSet conv_get_charset_from_str(const gchar *charset)
1232 {
1233         GHashTable *table;
1234
1235         if (!charset) return C_AUTO;
1236
1237         table = conv_get_charset_from_str_table();
1238         return GPOINTER_TO_UINT(g_hash_table_lookup(table, charset));
1239 }
1240
1241 static CharSet conv_get_locale_charset(void)
1242 {
1243         static CharSet cur_charset = -1;
1244         const gchar *cur_locale;
1245         const gchar *p;
1246         gint i;
1247
1248         if (cur_charset != -1)
1249                 return cur_charset;
1250
1251         cur_locale = conv_get_current_locale();
1252         if (!cur_locale) {
1253                 cur_charset = C_US_ASCII;
1254                 return cur_charset;
1255         }
1256
1257         if (strcasestr(cur_locale, "UTF-8")) {
1258                 cur_charset = C_UTF_8;
1259                 return cur_charset;
1260         }
1261
1262         if ((p = strcasestr(cur_locale, "@euro")) && p[5] == '\0') {
1263                 cur_charset = C_ISO_8859_15;
1264                 return cur_charset;
1265         }
1266
1267         for (i = 0; i < sizeof(locale_table) / sizeof(locale_table[0]); i++) {
1268                 const gchar *p;
1269
1270                 /* "ja_JP.EUC" matches with "ja_JP.eucJP", "ja_JP.EUC" and
1271                    "ja_JP". "ja_JP" matches with "ja_JP.xxxx" and "ja" */
1272                 if (!g_ascii_strncasecmp(cur_locale, locale_table[i].locale,
1273                                  strlen(locale_table[i].locale))) {
1274                         cur_charset = locale_table[i].charset;
1275                         return cur_charset;
1276                 } else if ((p = strchr(locale_table[i].locale, '_')) &&
1277                          !strchr(p + 1, '.')) {
1278                         if (strlen(cur_locale) == 2 &&
1279                             !g_ascii_strncasecmp(cur_locale, locale_table[i].locale, 2)) {
1280                                 cur_charset = locale_table[i].charset;
1281                                 return cur_charset;
1282                         }
1283                 }
1284         }
1285
1286         cur_charset = C_AUTO;
1287         return cur_charset;
1288 }
1289
1290 static CharSet conv_get_locale_charset_no_utf8(void)
1291 {
1292         static CharSet cur_charset = -1;
1293         const gchar *cur_locale;
1294         const gchar *p;
1295         gchar *tmp;
1296         gint i;
1297
1298         if (prefs_common.broken_are_utf8)
1299                 return conv_get_locale_charset();
1300
1301         if (cur_charset != -1)
1302                 return cur_charset;
1303
1304         cur_locale = conv_get_current_locale();
1305         if (!cur_locale) {
1306                 cur_charset = C_US_ASCII;
1307                 return cur_charset;
1308         }
1309
1310         if (strcasestr(cur_locale, "UTF-8")) {
1311                 tmp = g_strdup(cur_locale);
1312                 *(strcasestr(tmp, ".UTF-8")) = '\0';
1313                 cur_locale = tmp;
1314         }
1315
1316         if ((p = strcasestr(cur_locale, "@euro")) && p[5] == '\0') {
1317                 cur_charset = C_ISO_8859_15;
1318                 return cur_charset;
1319         }
1320
1321         for (i = 0; i < sizeof(locale_table) / sizeof(locale_table[0]); i++) {
1322                 const gchar *p;
1323
1324                 /* "ja_JP.EUC" matches with "ja_JP.eucJP", "ja_JP.EUC" and
1325                    "ja_JP". "ja_JP" matches with "ja_JP.xxxx" and "ja" */
1326                 if (!g_ascii_strncasecmp(cur_locale, locale_table[i].locale,
1327                                  strlen(locale_table[i].locale))) {
1328                         cur_charset = locale_table[i].charset;
1329                         return cur_charset;
1330                 } else if ((p = strchr(locale_table[i].locale, '_')) &&
1331                          !strchr(p + 1, '.')) {
1332                         if (strlen(cur_locale) == 2 &&
1333                             !g_ascii_strncasecmp(cur_locale, locale_table[i].locale, 2)) {
1334                                 cur_charset = locale_table[i].charset;
1335                                 return cur_charset;
1336                         }
1337                 }
1338         }
1339
1340         cur_charset = C_AUTO;
1341         return cur_charset;
1342 }
1343
1344 const gchar *conv_get_locale_charset_str(void)
1345 {
1346         static const gchar *codeset = NULL;
1347
1348         if (!codeset)
1349                 codeset = conv_get_charset_str(conv_get_locale_charset());
1350
1351         return codeset ? codeset : CS_INTERNAL;
1352 }
1353
1354 const gchar *conv_get_locale_charset_str_no_utf8(void)
1355 {
1356         static const gchar *codeset = NULL;
1357
1358         if (!codeset)
1359                 codeset = conv_get_charset_str(conv_get_locale_charset_no_utf8());
1360
1361         return codeset ? codeset : CS_INTERNAL;
1362 }
1363
1364 static CharSet conv_get_outgoing_charset(void)
1365 {
1366         static CharSet out_charset = -1;
1367         const gchar *cur_locale;
1368         const gchar *p;
1369         gint i;
1370
1371         if (out_charset != -1)
1372                 return out_charset;
1373
1374         cur_locale = conv_get_current_locale();
1375         if (!cur_locale) {
1376                 out_charset = C_AUTO;
1377                 return out_charset;
1378         }
1379
1380         if (strcasestr(cur_locale, "UTF-8")) {
1381                 out_charset = C_UTF_8;
1382                 return out_charset;
1383         }
1384
1385         if ((p = strcasestr(cur_locale, "@euro")) && p[5] == '\0') {
1386                 out_charset = C_ISO_8859_15;
1387                 return out_charset;
1388         }
1389
1390         for (i = 0; i < sizeof(locale_table) / sizeof(locale_table[0]); i++) {
1391                 const gchar *p;
1392
1393                 if (!g_ascii_strncasecmp(cur_locale, locale_table[i].locale,
1394                                  strlen(locale_table[i].locale))) {
1395                         out_charset = locale_table[i].out_charset;
1396                         break;
1397                 } else if ((p = strchr(locale_table[i].locale, '_')) &&
1398                          !strchr(p + 1, '.')) {
1399                         if (strlen(cur_locale) == 2 &&
1400                             !g_ascii_strncasecmp(cur_locale, locale_table[i].locale, 2)) {
1401                                 out_charset = locale_table[i].out_charset;
1402                                 break;
1403                         }
1404                 }
1405         }
1406
1407         return out_charset;
1408 }
1409
1410 const gchar *conv_get_outgoing_charset_str(void)
1411 {
1412         CharSet out_charset;
1413         const gchar *str;
1414
1415         out_charset = conv_get_outgoing_charset();
1416         str = conv_get_charset_str(out_charset);
1417
1418         return str ? str : CS_UTF_8;
1419 }
1420
1421 const gchar *conv_get_current_locale(void)
1422 {
1423         const gchar *cur_locale;
1424
1425 #ifdef G_OS_WIN32
1426         cur_locale = g_win32_getlocale();
1427 #else
1428         cur_locale = g_getenv("LC_ALL");
1429         if (!cur_locale) cur_locale = g_getenv("LC_CTYPE");
1430         if (!cur_locale) cur_locale = g_getenv("LANG");
1431         if (!cur_locale) cur_locale = setlocale(LC_CTYPE, NULL);
1432 #endif /* G_OS_WIN32 */
1433
1434         debug_print("current locale: %s\n",
1435                     cur_locale ? cur_locale : "(none)");
1436
1437         return cur_locale;
1438 }
1439
1440 static gboolean conv_is_ja_locale(void)
1441 {
1442         static gint is_ja_locale = -1;
1443         const gchar *cur_locale;
1444
1445         if (is_ja_locale != -1)
1446                 return is_ja_locale != 0;
1447
1448         is_ja_locale = 0;
1449         cur_locale = conv_get_current_locale();
1450         if (cur_locale) {
1451                 if (g_ascii_strncasecmp(cur_locale, "ja", 2) == 0)
1452                         is_ja_locale = 1;
1453         }
1454
1455         return is_ja_locale != 0;
1456 }
1457
1458 gchar *conv_unmime_header(const gchar *str, const gchar *default_encoding)
1459 {
1460         gchar buf[BUFFSIZE];
1461
1462         if (is_ascii_str(str))
1463                 return unmime_header(str);
1464
1465         if (default_encoding) {
1466                 gchar *utf8_buf;
1467
1468                 utf8_buf = conv_codeset_strdup
1469                         (str, default_encoding, CS_INTERNAL);
1470                 if (utf8_buf) {
1471                         gchar *decoded_str;
1472
1473                         decoded_str = unmime_header(utf8_buf);
1474                         g_free(utf8_buf);
1475                         return decoded_str;
1476                 }
1477         }
1478
1479         if (conv_is_ja_locale())
1480                 conv_anytodisp(buf, sizeof(buf), str);
1481         else
1482                 conv_localetodisp(buf, sizeof(buf), str);
1483
1484         return unmime_header(buf);
1485 }
1486
1487 #define MAX_LINELEN             76
1488 #define MAX_HARD_LINELEN        996
1489 #define MIMESEP_BEGIN           "=?"
1490 #define MIMESEP_END             "?="
1491
1492 #define LBREAK_IF_REQUIRED(cond, is_plain_text)                         \
1493 {                                                                       \
1494         if (len - (destp - (guchar *)dest) < MAX_LINELEN + 2) {         \
1495                 *destp = '\0';                                          \
1496                 return;                                                 \
1497         }                                                               \
1498                                                                         \
1499         if ((cond) && *srcp) {                                          \
1500                 if (destp > (guchar *)dest && left < MAX_LINELEN - 1) { \
1501                         if (isspace(*(destp - 1)))                      \
1502                                 destp--;                                \
1503                         else if (is_plain_text && isspace(*srcp))       \
1504                                 srcp++;                                 \
1505                         if (*srcp) {                                    \
1506                                 *destp++ = '\n';                        \
1507                                 *destp++ = ' ';                         \
1508                                 left = MAX_LINELEN - 1;                 \
1509                         }                                               \
1510                 }                                                       \
1511         }                                                               \
1512 }
1513
1514 void conv_encode_header_full(gchar *dest, gint len, const gchar *src,
1515                         gint header_len, gboolean addr_field,
1516                         const gchar *out_encoding_)
1517 {
1518         const gchar *cur_encoding;
1519         const gchar *out_encoding;
1520         gint mimestr_len;
1521         gchar *mimesep_enc;
1522         gint left;
1523         const guchar *srcp = src;
1524         guchar *destp = dest;
1525         gboolean use_base64;
1526
1527         g_return_if_fail(g_utf8_validate(src, -1, NULL) == TRUE);
1528
1529         if (MB_CUR_MAX > 1) {
1530                 use_base64 = TRUE;
1531                 mimesep_enc = "?B?";
1532         } else {
1533                 use_base64 = FALSE;
1534                 mimesep_enc = "?Q?";
1535         }
1536
1537         cur_encoding = CS_INTERNAL;
1538
1539         if (out_encoding_)
1540                 out_encoding = out_encoding_;
1541         else
1542                 out_encoding = conv_get_outgoing_charset_str();
1543
1544         if (!strcmp(out_encoding, CS_US_ASCII))
1545                 out_encoding = CS_ISO_8859_1;
1546
1547         mimestr_len = strlen(MIMESEP_BEGIN) + strlen(out_encoding) +
1548                 strlen(mimesep_enc) + strlen(MIMESEP_END);
1549
1550         left = MAX_LINELEN - header_len;
1551
1552         while (*srcp) {
1553                 LBREAK_IF_REQUIRED(left <= 0, TRUE);
1554
1555                 while (isspace(*srcp)) {
1556                         *destp++ = *srcp++;
1557                         left--;
1558                         LBREAK_IF_REQUIRED(left <= 0, TRUE);
1559                 }
1560
1561                 /* output as it is if the next word is ASCII string */
1562                 if (!is_next_nonascii(srcp)) {
1563                         gint word_len;
1564
1565                         word_len = get_next_word_len(srcp);
1566                         LBREAK_IF_REQUIRED(left < word_len, TRUE);
1567                         while (word_len > 0) {
1568                                 LBREAK_IF_REQUIRED(left + (MAX_HARD_LINELEN - MAX_LINELEN) <= 0, TRUE)
1569                                 *destp++ = *srcp++;
1570                                 left--;
1571                                 word_len--;
1572                         }
1573
1574                         continue;
1575                 }
1576
1577                 /* don't include parentheses in encoded strings */
1578                 if (addr_field && (*srcp == '(' || *srcp == ')')) {
1579                         LBREAK_IF_REQUIRED(left < 2, FALSE);
1580                         *destp++ = *srcp++;
1581                         left--;
1582                 }
1583
1584                 while (1) {
1585                         gint mb_len = 0;
1586                         gint cur_len = 0;
1587                         gchar *part_str;
1588                         gchar *out_str;
1589                         gchar *enc_str;
1590                         const guchar *p = srcp;
1591                         gint out_str_len;
1592                         gint out_enc_str_len;
1593                         gint mime_block_len;
1594                         gboolean cont = FALSE;
1595
1596                         while (*p != '\0') {
1597                                 if (isspace(*p) && !is_next_nonascii(p + 1))
1598                                         break;
1599                                 /* don't include parentheses in encoded
1600                                    strings */
1601                                 if (addr_field && (*p == '(' || *p == ')'))
1602                                         break;
1603
1604                                 mb_len = g_utf8_skip[*p];
1605
1606                                 Xstrndup_a(part_str, srcp, cur_len + mb_len, );
1607                                 out_str = conv_codeset_strdup
1608                                         (part_str, cur_encoding, out_encoding);
1609                                 if (!out_str) {
1610                                         if (strict_mode) {
1611                                                 *dest = '\0';
1612                                                 return;
1613                                         } else {
1614                                                 g_warning("conv_encode_header(): code conversion failed\n");
1615                                                 conv_unreadable_8bit(part_str);
1616                                                 out_str = g_strdup(part_str);
1617                                         }
1618                                 }
1619                                 out_str_len = strlen(out_str);
1620
1621                                 if (use_base64)
1622                                         out_enc_str_len = B64LEN(out_str_len);
1623                                 else
1624                                         out_enc_str_len =
1625                                                 qp_get_q_encoding_len(out_str);
1626
1627                                 g_free(out_str);
1628
1629                                 if (mimestr_len + out_enc_str_len <= left) {
1630                                         cur_len += mb_len;
1631                                         p += mb_len;
1632                                 } else if (cur_len == 0) {
1633                                         LBREAK_IF_REQUIRED(1, FALSE);
1634                                         continue;
1635                                 } else {
1636                                         cont = TRUE;
1637                                         break;
1638                                 }
1639                         }
1640
1641                         if (cur_len > 0) {
1642                                 Xstrndup_a(part_str, srcp, cur_len, );
1643                                 out_str = conv_codeset_strdup
1644                                         (part_str, cur_encoding, out_encoding);
1645                                 if (!out_str) {
1646                                         g_warning("conv_encode_header(): code conversion failed\n");
1647                                         conv_unreadable_8bit(part_str);
1648                                         out_str = g_strdup(part_str);
1649                                 }
1650                                 out_str_len = strlen(out_str);
1651
1652                                 if (use_base64)
1653                                         out_enc_str_len = B64LEN(out_str_len);
1654                                 else
1655                                         out_enc_str_len =
1656                                                 qp_get_q_encoding_len(out_str);
1657
1658                                 Xalloca(enc_str, out_enc_str_len + 1, );
1659                                 if (use_base64)
1660                                         base64_encode(enc_str, out_str, out_str_len);
1661                                 else
1662                                         qp_q_encode(enc_str, out_str);
1663
1664                                 g_free(out_str);
1665
1666                                 /* output MIME-encoded string block */
1667                                 mime_block_len = mimestr_len + strlen(enc_str);
1668                                 g_snprintf(destp, mime_block_len + 1,
1669                                            MIMESEP_BEGIN "%s%s%s" MIMESEP_END,
1670                                            out_encoding, mimesep_enc, enc_str);
1671                                 destp += mime_block_len;
1672                                 srcp += cur_len;
1673
1674                                 left -= mime_block_len;
1675                         }
1676
1677                         LBREAK_IF_REQUIRED(cont, FALSE);
1678
1679                         if (cur_len == 0)
1680                                 break;
1681                 }
1682         }
1683
1684         *destp = '\0';
1685 }
1686
1687 void conv_encode_header(gchar *dest, gint len, const gchar *src,
1688                         gint header_len, gboolean addr_field)
1689 {
1690         conv_encode_header_full(dest,len,src,header_len,addr_field,NULL);
1691 }
1692
1693 #undef LBREAK_IF_REQUIRED
1694 gchar *conv_filename_from_utf8(const gchar *utf8_file)
1695 {
1696         gchar *fs_file;
1697         GError *error = NULL;
1698
1699         fs_file = g_filename_from_utf8(utf8_file, -1, NULL, NULL, &error);
1700         if (error) {
1701                 g_warning("failed to convert encoding of file name: %s\n",
1702                           error->message);
1703                 g_error_free(error);
1704         }
1705         if (!fs_file)
1706                 fs_file = g_strdup(utf8_file);
1707
1708         return fs_file;
1709 }
1710
1711 gchar *conv_filename_to_utf8(const gchar *fs_file)
1712 {
1713         gchar *utf8_file = NULL;
1714         GError *error = NULL;
1715
1716         utf8_file = g_filename_to_utf8(fs_file, -1, NULL, NULL, &error);
1717         if (error) {
1718                 g_warning("failed to convert encoding of file name: %s\n",
1719                           error->message);
1720                 g_error_free(error);
1721         }
1722
1723         if (!utf8_file || !g_utf8_validate(utf8_file, -1, NULL)) {
1724                 g_free(utf8_file);
1725                 utf8_file = g_strdup(fs_file);
1726                 conv_unreadable_8bit(utf8_file);
1727         }
1728
1729         return utf8_file;
1730 }