Better fix for bug #3349 (almost all the needed code was there :-)
[claws.git] / src / codeconv.c
1 /*
2  * Sylpheed -- a GTK+ based, lightweight, and fast e-mail client
3  * Copyright (C) 1999-2012 Hiroyuki Yamamoto and the Claws Mail team
4  *
5  * This program is free software; you can redistribute it and/or modify
6  * it under the terms of the GNU General Public License as published by
7  * the Free Software Foundation; either version 3 of the License, or
8  * (at your option) any later version.
9  *
10  * This program is distributed in the hope that it will be useful,
11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
13  * GNU General Public License for more details.
14  *
15  * You should have received a copy of the GNU General Public License
16  * along with this program. If not, see <http://www.gnu.org/licenses/>.
17  * 
18  */
19
20 #ifdef HAVE_CONFIG_H
21 #  include "config.h"
22 #include "claws-features.h"
23 #endif
24
25 #include "defs.h"
26
27 #include <glib.h>
28 #include <glib/gi18n.h>
29 #include <string.h>
30 #include <ctype.h>
31 #include <stdlib.h>
32 #include <errno.h>
33
34 #if HAVE_LOCALE_H
35 #  include <locale.h>
36 #endif
37
38 #include "codeconv.h"
39 #include "unmime.h"
40 #include "quoted-printable.h"
41 #include "utils.h"
42 #include "prefs_common.h"
43
44 /* For unknown reasons the inconv.m4 macro undefs that macro if no
45    const is needed.  This would break the code below so we define it. */
46 #ifndef ICONV_CONST
47 #define ICONV_CONST
48 #endif
49
50 typedef enum
51 {
52         JIS_ASCII,
53         JIS_KANJI,
54         JIS_HWKANA,
55         JIS_AUXKANJI
56 } JISState;
57
58 #define SUBST_CHAR      0x5f;
59 #define ESC             '\033'
60
61 #define iseuckanji(c) \
62         (((c) & 0xff) >= 0xa1 && ((c) & 0xff) <= 0xfe)
63 #define iseuchwkana1(c) \
64         (((c) & 0xff) == 0x8e)
65 #define iseuchwkana2(c) \
66         (((c) & 0xff) >= 0xa1 && ((c) & 0xff) <= 0xdf)
67 #define iseucaux(c) \
68         (((c) & 0xff) == 0x8f)
69 #define issjiskanji1(c) \
70         ((((c) & 0xff) >= 0x81 && ((c) & 0xff) <= 0x9f) || \
71          (((c) & 0xff) >= 0xe0 && ((c) & 0xff) <= 0xfc))
72 #define issjiskanji2(c) \
73         ((((c) & 0xff) >= 0x40 && ((c) & 0xff) <= 0x7e) || \
74          (((c) & 0xff) >= 0x80 && ((c) & 0xff) <= 0xfc))
75 #define issjishwkana(c) \
76         (((c) & 0xff) >= 0xa1 && ((c) & 0xff) <= 0xdf)
77
78 #define K_IN()                          \
79         if (state != JIS_KANJI) {       \
80                 *out++ = ESC;           \
81                 *out++ = '$';           \
82                 *out++ = 'B';           \
83                 state = JIS_KANJI;      \
84         }
85
86 #define K_OUT()                         \
87         if (state != JIS_ASCII) {       \
88                 *out++ = ESC;           \
89                 *out++ = '(';           \
90                 *out++ = 'B';           \
91                 state = JIS_ASCII;      \
92         }
93
94 #define HW_IN()                         \
95         if (state != JIS_HWKANA) {      \
96                 *out++ = ESC;           \
97                 *out++ = '(';           \
98                 *out++ = 'I';           \
99                 state = JIS_HWKANA;     \
100         }
101
102 #define AUX_IN()                        \
103         if (state != JIS_AUXKANJI) {    \
104                 *out++ = ESC;           \
105                 *out++ = '$';           \
106                 *out++ = '(';           \
107                 *out++ = 'D';           \
108                 state = JIS_AUXKANJI;   \
109         }
110
111 static CodeConvFunc conv_get_code_conv_func     (const gchar    *src_charset_str,
112                                          const gchar    *dest_charset_str);
113
114 static gchar *conv_iconv_strdup_with_cd (const gchar    *inbuf,
115                                          iconv_t         cd);
116
117 static gchar *conv_iconv_strdup         (const gchar    *inbuf,
118                                          const gchar    *src_code,
119                                          const gchar    *dest_code);
120
121 static CharSet conv_get_locale_charset                  (void);
122 static CharSet conv_get_outgoing_charset                (void);
123 static CharSet conv_guess_ja_encoding(const gchar *str);
124 static gboolean conv_is_ja_locale                       (void);
125
126 static gint conv_jistoeuc(gchar *outbuf, gint outlen, const gchar *inbuf);
127 static gint conv_euctojis(gchar *outbuf, gint outlen, const gchar *inbuf);
128 static gint conv_sjistoeuc(gchar *outbuf, gint outlen, const gchar *inbuf);
129
130 static gint conv_jistoutf8(gchar *outbuf, gint outlen, const gchar *inbuf);
131 static gint conv_sjistoutf8(gchar *outbuf, gint outlen, const gchar *inbuf);
132 static gint conv_euctoutf8(gchar *outbuf, gint outlen, const gchar *inbuf);
133 static gint conv_anytoutf8(gchar *outbuf, gint outlen, const gchar *inbuf);
134
135 static gint conv_utf8toeuc(gchar *outbuf, gint outlen, const gchar *inbuf);
136 static gint conv_utf8tojis(gchar *outbuf, gint outlen, const gchar *inbuf);
137
138 static void conv_unreadable_8bit(gchar *str);
139
140 static gint conv_jistodisp(gchar *outbuf, gint outlen, const gchar *inbuf);
141 static gint conv_sjistodisp(gchar *outbuf, gint outlen, const gchar *inbuf);
142 static gint conv_euctodisp(gchar *outbuf, gint outlen, const gchar *inbuf);
143
144 static gint conv_anytodisp(gchar *outbuf, gint outlen, const gchar *inbuf);
145 static gint conv_ustodisp(gchar *outbuf, gint outlen, const gchar *inbuf);
146 static gint conv_noconv(gchar *outbuf, gint outlen, const gchar *inbuf);
147
148 static gboolean strict_mode = FALSE;
149
150 void codeconv_set_strict(gboolean mode)
151 {
152         strict_mode = mode;
153 }
154
155 static gint conv_jistoeuc(gchar *outbuf, gint outlen, const gchar *inbuf)
156 {
157         const guchar *in = inbuf;
158         guchar *out = outbuf;
159         JISState state = JIS_ASCII;
160
161         while (*in != '\0') {
162                 if (*in == ESC) {
163                         in++;
164                         if (*in == '$') {
165                                 if (*(in + 1) == '@' || *(in + 1) == 'B') {
166                                         state = JIS_KANJI;
167                                         in += 2;
168                                 } else if (*(in + 1) == '(' &&
169                                            *(in + 2) == 'D') {
170                                         state = JIS_AUXKANJI;
171                                         in += 3;
172                                 } else {
173                                         /* unknown escape sequence */
174                                         state = JIS_ASCII;
175                                 }
176                         } else if (*in == '(') {
177                                 if (*(in + 1) == 'B' || *(in + 1) == 'J') {
178                                         state = JIS_ASCII;
179                                         in += 2;
180                                 } else if (*(in + 1) == 'I') {
181                                         state = JIS_HWKANA;
182                                         in += 2;
183                                 } else {
184                                         /* unknown escape sequence */
185                                         state = JIS_ASCII;
186                                 }
187                         } else {
188                                 /* unknown escape sequence */
189                                 state = JIS_ASCII;
190                         }
191                 } else if (*in == 0x0e) {
192                         state = JIS_HWKANA;
193                         in++;
194                 } else if (*in == 0x0f) {
195                         state = JIS_ASCII;
196                         in++;
197                 } else {
198                         switch (state) {
199                         case JIS_ASCII:
200                                 *out++ = *in++;
201                                 break;
202                         case JIS_KANJI:
203                                 *out++ = *in++ | 0x80;
204                                 if (*in == '\0') break;
205                                 *out++ = *in++ | 0x80;
206                                 break;
207                         case JIS_HWKANA:
208                                 *out++ = 0x8e;
209                                 *out++ = *in++ | 0x80;
210                                 break;
211                         case JIS_AUXKANJI:
212                                 *out++ = 0x8f;
213                                 *out++ = *in++ | 0x80;
214                                 if (*in == '\0') break;
215                                 *out++ = *in++ | 0x80;
216                                 break;
217                         }
218                 }
219         }
220
221         *out = '\0';
222         return 0;
223 }
224
225 #define JIS_HWDAKUTEN           0x5e
226 #define JIS_HWHANDAKUTEN        0x5f
227
228 static gint conv_jis_hantozen(guchar *outbuf, guchar jis_code, guchar sound_sym)
229 {
230         static guint16 h2z_tbl[] = {
231                 /* 0x20 - 0x2f */
232                 0x0000, 0x2123, 0x2156, 0x2157, 0x2122, 0x2126, 0x2572, 0x2521,
233                 0x2523, 0x2525, 0x2527, 0x2529, 0x2563, 0x2565, 0x2567, 0x2543,
234                 /* 0x30 - 0x3f */
235                 0x213c, 0x2522, 0x2524, 0x2526, 0x2528, 0x252a, 0x252b, 0x252d,
236                 0x252f, 0x2531, 0x2533, 0x2535, 0x2537, 0x2539, 0x253b, 0x253d,
237                 /* 0x40 - 0x4f */
238                 0x253f, 0x2541, 0x2544, 0x2546, 0x2548, 0x254a, 0x254b, 0x254c,
239                 0x254d, 0x254e, 0x254f, 0x2552, 0x2555, 0x2558, 0x255b, 0x255e,
240                 /* 0x50 - 0x5f */
241                 0x255f, 0x2560, 0x2561, 0x2562, 0x2564, 0x2566, 0x2568, 0x2569,
242                 0x256a, 0x256b, 0x256c, 0x256d, 0x256f, 0x2573, 0x212b, 0x212c
243         };
244
245         static guint16 dakuten_tbl[] = {
246                 /* 0x30 - 0x3f */
247                 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x252c, 0x252e,
248                 0x2530, 0x2532, 0x2534, 0x2536, 0x2538, 0x253a, 0x253c, 0x253e,
249                 /* 0x40 - 0x4f */
250                 0x2540, 0x2542, 0x2545, 0x2547, 0x2549, 0x0000, 0x0000, 0x0000,
251                 0x0000, 0x0000, 0x2550, 0x2553, 0x2556, 0x2559, 0x255c, 0x0000
252         };
253
254         static guint16 handakuten_tbl[] = {
255                 /* 0x4a - 0x4e */
256                 0x2551, 0x2554, 0x2557, 0x255a, 0x255d
257         };
258
259         guint16 out_code;
260
261         jis_code &= 0x7f;
262         sound_sym &= 0x7f;
263
264         if (jis_code < 0x21 || jis_code > 0x5f)
265                 return 0;
266
267         if (sound_sym == JIS_HWDAKUTEN &&
268             jis_code >= 0x36 && jis_code <= 0x4e) {
269                 out_code = dakuten_tbl[jis_code - 0x30];
270                 if (out_code != 0) {
271                         *outbuf = out_code >> 8;
272                         *(outbuf + 1) = out_code & 0xff;
273                         return 2;
274                 }
275         }
276
277         if (sound_sym == JIS_HWHANDAKUTEN &&
278             jis_code >= 0x4a && jis_code <= 0x4e) {
279                 out_code = handakuten_tbl[jis_code - 0x4a];
280                 *outbuf = out_code >> 8;
281                 *(outbuf + 1) = out_code & 0xff;
282                 return 2;
283         }
284
285         out_code = h2z_tbl[jis_code - 0x20];
286         *outbuf = out_code >> 8;
287         *(outbuf + 1) = out_code & 0xff;
288         return 1;
289 }
290
291 static gint conv_euctojis(gchar *outbuf, gint outlen, const gchar *inbuf)
292 {
293         const guchar *in = inbuf;
294         guchar *out = outbuf;
295         JISState state = JIS_ASCII;
296
297         while (*in != '\0') {
298                 if (IS_ASCII(*in)) {
299                         K_OUT();
300                         *out++ = *in++;
301                 } else if (iseuckanji(*in)) {
302                         if (iseuckanji(*(in + 1))) {
303                                 K_IN();
304                                 *out++ = *in++ & 0x7f;
305                                 *out++ = *in++ & 0x7f;
306                         } else {
307                                 K_OUT();
308                                 *out++ = SUBST_CHAR;
309                                 in++;
310                                 if (*in != '\0' && !IS_ASCII(*in)) {
311                                         *out++ = SUBST_CHAR;
312                                         in++;
313                                 }
314                         }
315                 } else if (iseuchwkana1(*in)) {
316                         if (iseuchwkana2(*(in + 1))) {
317                                 if (prefs_common.allow_jisx0201_kana) {
318                                         HW_IN();
319                                         in++;
320                                         *out++ = *in++ & 0x7f;
321                                 } else {
322                                         guchar jis_ch[2];
323                                         gint len;
324
325                                         if (iseuchwkana1(*(in + 2)) &&
326                                             iseuchwkana2(*(in + 3)))
327                                                 len = conv_jis_hantozen
328                                                         (jis_ch,
329                                                          *(in + 1), *(in + 3));
330                                         else
331                                                 len = conv_jis_hantozen
332                                                         (jis_ch,
333                                                          *(in + 1), '\0');
334                                         if (len == 0)
335                                                 in += 2;
336                                         else {
337                                                 K_IN();
338                                                 in += len * 2;
339                                                 *out++ = jis_ch[0];
340                                                 *out++ = jis_ch[1];
341                                         }
342                                 }
343                         } else {
344                                 K_OUT();
345                                 in++;
346                                 if (*in != '\0' && !IS_ASCII(*in)) {
347                                         *out++ = SUBST_CHAR;
348                                         in++;
349                                 }
350                         }
351                 } else if (iseucaux(*in)) {
352                         in++;
353                         if (iseuckanji(*in) && iseuckanji(*(in + 1))) {
354                                 AUX_IN();
355                                 *out++ = *in++ & 0x7f;
356                                 *out++ = *in++ & 0x7f;
357                         } else {
358                                 K_OUT();
359                                 if (*in != '\0' && !IS_ASCII(*in)) {
360                                         *out++ = SUBST_CHAR;
361                                         in++;
362                                         if (*in != '\0' && !IS_ASCII(*in)) {
363                                                 *out++ = SUBST_CHAR;
364                                                 in++;
365                                         }
366                                 }
367                         }
368                 } else {
369                         K_OUT();
370                         *out++ = SUBST_CHAR;
371                         in++;
372                 }
373         }
374
375         K_OUT();
376         *out = '\0';
377         return 0;
378 }
379
380 static gint conv_sjistoeuc(gchar *outbuf, gint outlen, const gchar *inbuf)
381 {
382         const guchar *in = inbuf;
383         guchar *out = outbuf;
384
385         while (*in != '\0') {
386                 if (IS_ASCII(*in)) {
387                         *out++ = *in++;
388                 } else if (issjiskanji1(*in)) {
389                         if (issjiskanji2(*(in + 1))) {
390                                 guchar out1 = *in;
391                                 guchar out2 = *(in + 1);
392                                 guchar row;
393
394                                 row = out1 < 0xa0 ? 0x70 : 0xb0;
395                                 if (out2 < 0x9f) {
396                                         out1 = (out1 - row) * 2 - 1;
397                                         out2 -= out2 > 0x7f ? 0x20 : 0x1f;
398                                 } else {
399                                         out1 = (out1 - row) * 2;
400                                         out2 -= 0x7e;
401                                 }
402
403                                 *out++ = out1 | 0x80;
404                                 *out++ = out2 | 0x80;
405                                 in += 2;
406                         } else {
407                                 *out++ = SUBST_CHAR;
408                                 in++;
409                                 if (*in != '\0' && !IS_ASCII(*in)) {
410                                         *out++ = SUBST_CHAR;
411                                         in++;
412                                 }
413                         }
414                 } else if (issjishwkana(*in)) {
415                         *out++ = 0x8e;
416                         *out++ = *in++;
417                 } else {
418                         *out++ = SUBST_CHAR;
419                         in++;
420                 }
421         }
422
423         *out = '\0';
424         return 0;
425 }
426
427 static gint conv_jistoutf8(gchar *outbuf, gint outlen, const gchar *inbuf)
428 {
429         gchar *eucstr;
430
431         Xalloca(eucstr, outlen, return -1);
432
433         if (conv_jistoeuc(eucstr, outlen, inbuf) <0)
434                 return -1;
435         if (conv_euctoutf8(outbuf, outlen, eucstr) < 0)
436                 return -1;
437         return 0;
438 }
439
440 static gint conv_sjistoutf8(gchar *outbuf, gint outlen, const gchar *inbuf)
441 {
442         gchar *tmpstr;
443
444         tmpstr = conv_iconv_strdup(inbuf, CS_SHIFT_JIS, CS_UTF_8);
445         if (tmpstr) {
446                 strncpy2(outbuf, tmpstr, outlen);
447                 g_free(tmpstr);
448                 return 0;
449         } else {
450                 strncpy2(outbuf, inbuf, outlen);
451                 return -1;
452         }
453 }
454
455 static gint conv_euctoutf8(gchar *outbuf, gint outlen, const gchar *inbuf)
456 {
457         static iconv_t cd = (iconv_t)-1;
458         static gboolean iconv_ok = TRUE;
459         gchar *tmpstr;
460
461         if (cd == (iconv_t)-1) {
462                 if (!iconv_ok) {
463                         strncpy2(outbuf, inbuf, outlen);
464                         return -1;
465                 }
466                 cd = iconv_open(CS_UTF_8, CS_EUC_JP_MS);
467                 if (cd == (iconv_t)-1) {
468                         cd = iconv_open(CS_UTF_8, CS_EUC_JP);
469                         if (cd == (iconv_t)-1) {
470                                 g_warning("conv_euctoutf8(): %s\n",
471                                           g_strerror(errno));
472                                 iconv_ok = FALSE;
473                                 strncpy2(outbuf, inbuf, outlen);
474                                 return -1;
475                         }
476                 }
477         }
478
479         tmpstr = conv_iconv_strdup_with_cd(inbuf, cd);
480         if (tmpstr) {
481                 strncpy2(outbuf, tmpstr, outlen);
482                 g_free(tmpstr);
483                 return 0;
484         } else {
485                 strncpy2(outbuf, inbuf, outlen);
486                 return -1;
487         }
488 }
489
490 static gint conv_anytoutf8(gchar *outbuf, gint outlen, const gchar *inbuf)
491 {
492         gint r = -1;
493         switch (conv_guess_ja_encoding(inbuf)) {
494         case C_ISO_2022_JP:
495                 r = conv_jistoutf8(outbuf, outlen, inbuf);
496                 break;
497         case C_SHIFT_JIS:
498                 r = conv_sjistoutf8(outbuf, outlen, inbuf);
499                 break;
500         case C_EUC_JP:
501                 r = conv_euctoutf8(outbuf, outlen, inbuf);
502                 break;
503         default:
504                 r = 0;
505                 strncpy2(outbuf, inbuf, outlen);
506                 break;
507         }
508         
509         return r;
510 }
511
512 static gint conv_utf8toeuc(gchar *outbuf, gint outlen, const gchar *inbuf)
513 {
514         static iconv_t cd = (iconv_t)-1;
515         static gboolean iconv_ok = TRUE;
516         gchar *tmpstr;
517
518         if (cd == (iconv_t)-1) {
519                 if (!iconv_ok) {
520                         strncpy2(outbuf, inbuf, outlen);
521                         return -1;
522                 }
523                 cd = iconv_open(CS_EUC_JP_MS, CS_UTF_8);
524                 if (cd == (iconv_t)-1) {
525                         cd = iconv_open(CS_EUC_JP, CS_UTF_8);
526                         if (cd == (iconv_t)-1) {
527                                 g_warning("conv_utf8toeuc(): %s\n",
528                                           g_strerror(errno));
529                                 iconv_ok = FALSE;
530                                 strncpy2(outbuf, inbuf, outlen);
531                                 return -1;
532                         }
533                 }
534         }
535
536         tmpstr = conv_iconv_strdup_with_cd(inbuf, cd);
537         if (tmpstr) {
538                 strncpy2(outbuf, tmpstr, outlen);
539                 g_free(tmpstr);
540                 return 0;
541         } else {
542                 strncpy2(outbuf, inbuf, outlen);
543                 return -1;
544         }
545 }
546
547 static gint conv_utf8tojis(gchar *outbuf, gint outlen, const gchar *inbuf)
548 {
549         gchar *eucstr;
550
551         Xalloca(eucstr, outlen, return -1);
552
553         if (conv_utf8toeuc(eucstr, outlen, inbuf) < 0)
554                 return -1;
555         if (conv_euctojis(outbuf, outlen, eucstr) < 0)
556                 return -1;
557                 
558         return 0;
559 }
560
561 static void conv_unreadable_8bit(gchar *str)
562 {
563         register guchar *p = str;
564
565         while (*p != '\0') {
566                 /* convert CR+LF -> LF */
567                 if (*p == '\r' && *(p + 1) == '\n')
568                         memmove(p, p + 1, strlen(p));
569                 else if (!IS_ASCII(*p)) *p = SUBST_CHAR;
570                 p++;
571         }
572 }
573
574 static CharSet conv_guess_ja_encoding(const gchar *str)
575 {
576         const guchar *p = str;
577         CharSet guessed = C_US_ASCII;
578
579         while (*p != '\0') {
580                 if (*p == ESC && (*(p + 1) == '$' || *(p + 1) == '(')) {
581                         if (guessed == C_US_ASCII)
582                                 return C_ISO_2022_JP;
583                         p += 2;
584                 } else if (IS_ASCII(*p)) {
585                         p++;
586                 } else if (iseuckanji(*p) && iseuckanji(*(p + 1))) {
587                         if (*p >= 0xfd && *p <= 0xfe)
588                                 return C_EUC_JP;
589                         else if (guessed == C_SHIFT_JIS) {
590                                 if ((issjiskanji1(*p) &&
591                                      issjiskanji2(*(p + 1))) ||
592                                     issjishwkana(*p))
593                                         guessed = C_SHIFT_JIS;
594                                 else
595                                         guessed = C_EUC_JP;
596                         } else
597                                 guessed = C_EUC_JP;
598                         p += 2;
599                 } else if (issjiskanji1(*p) && issjiskanji2(*(p + 1))) {
600                         if (iseuchwkana1(*p) && iseuchwkana2(*(p + 1)))
601                                 guessed = C_SHIFT_JIS;
602                         else
603                                 return C_SHIFT_JIS;
604                         p += 2;
605                 } else if (issjishwkana(*p)) {
606                         guessed = C_SHIFT_JIS;
607                         p++;
608                 } else {
609                         p++;
610                 }
611         }
612
613         return guessed;
614 }
615
616 static gint conv_jistodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
617 {
618         return conv_jistoutf8(outbuf, outlen, inbuf);
619 }
620
621 static gint conv_sjistodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
622 {
623         return conv_sjistoutf8(outbuf, outlen, inbuf);
624 }
625
626 static gint conv_euctodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
627 {
628         return conv_euctoutf8(outbuf, outlen, inbuf);
629 }
630
631 void conv_utf8todisp(gchar *outbuf, gint outlen, const gchar *inbuf)
632 {
633         if (g_utf8_validate(inbuf, -1, NULL) == TRUE)
634                 strncpy2(outbuf, inbuf, outlen);
635         else
636                 conv_ustodisp(outbuf, outlen, inbuf);
637 }
638
639 static gint conv_anytodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
640 {
641         gint r = 0;
642         if (conv_anytoutf8(outbuf, outlen, inbuf) < 0)
643                 r = -1;
644         if (g_utf8_validate(outbuf, -1, NULL) != TRUE)
645                 conv_unreadable_8bit(outbuf);
646         return r;
647 }
648
649 static gint conv_ustodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
650 {
651         strncpy2(outbuf, inbuf, outlen);
652         conv_unreadable_8bit(outbuf);
653         
654         return 0;
655 }
656
657 void conv_localetodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
658 {
659         gchar *tmpstr;
660
661         codeconv_set_strict(TRUE);
662         tmpstr = conv_iconv_strdup(inbuf, conv_get_locale_charset_str(),
663                                    CS_INTERNAL);
664         codeconv_set_strict(FALSE);
665         if (tmpstr && g_utf8_validate(tmpstr, -1, NULL)) {
666                 strncpy2(outbuf, tmpstr, outlen);
667                 g_free(tmpstr);
668                 return;
669         } else if (tmpstr && !g_utf8_validate(tmpstr, -1, NULL)) {
670                 g_free(tmpstr);
671                 codeconv_set_strict(TRUE);
672                 tmpstr = conv_iconv_strdup(inbuf, 
673                                 conv_get_locale_charset_str_no_utf8(),
674                                 CS_INTERNAL);
675                 codeconv_set_strict(FALSE);
676         }
677         if (tmpstr && g_utf8_validate(tmpstr, -1, NULL)) {
678                 strncpy2(outbuf, tmpstr, outlen);
679                 g_free(tmpstr);
680                 return;
681         } else {
682                 g_free(tmpstr);
683                 conv_utf8todisp(outbuf, outlen, inbuf);
684         }
685 }
686
687 static gint conv_noconv(gchar *outbuf, gint outlen, const gchar *inbuf)
688 {
689         strncpy2(outbuf, inbuf, outlen);
690         return 0;
691 }
692
693 static const gchar *
694 conv_get_fallback_for_private_encoding(const gchar *encoding)
695 {
696         if (encoding && (encoding[0] == 'X' || encoding[0] == 'x') &&
697             encoding[1] == '-') {
698                 if (!g_ascii_strcasecmp(encoding, CS_X_GBK))
699                         return CS_GBK;
700         }
701
702         return encoding;
703 }
704
705 CodeConverter *conv_code_converter_new(const gchar *src_charset)
706 {
707         CodeConverter *conv;
708
709         src_charset = conv_get_fallback_for_private_encoding(src_charset);
710
711         conv = g_new0(CodeConverter, 1);
712         conv->code_conv_func = conv_get_code_conv_func(src_charset, NULL);
713         conv->charset_str = g_strdup(src_charset);
714         conv->charset = conv_get_charset_from_str(src_charset);
715
716         return conv;
717 }
718
719 void conv_code_converter_destroy(CodeConverter *conv)
720 {
721         g_free(conv->charset_str);
722         g_free(conv);
723 }
724
725 gint conv_convert(CodeConverter *conv, gchar *outbuf, gint outlen,
726                   const gchar *inbuf)
727 {
728         if (conv->code_conv_func != conv_noconv)
729                 return conv->code_conv_func(outbuf, outlen, inbuf);
730         else {
731                 gchar *str;
732
733                 str = conv_iconv_strdup(inbuf, conv->charset_str, NULL);
734                 if (!str)
735                         return -1;
736                 else {
737                         strncpy2(outbuf, str, outlen);
738                         g_free(str);
739                 }
740         }
741
742         return 0;
743 }
744
745 gchar *conv_codeset_strdup(const gchar *inbuf,
746                            const gchar *src_code, const gchar *dest_code)
747 {
748         gchar *buf;
749         size_t len;
750         CodeConvFunc conv_func;
751
752         if (!strcmp2(src_code, dest_code)) {
753                 CharSet dest_charset = conv_get_charset_from_str(dest_code);
754                 if (strict_mode && dest_charset == C_UTF_8) {
755                         /* ensure valid UTF-8 if target is UTF-8 */
756                         if (!g_utf8_validate(inbuf, -1, NULL)) {
757                                 return NULL;
758                         }
759                 }
760                 /* otherwise, try for a lucky day */
761                 return g_strdup(inbuf);
762         }
763
764         src_code = conv_get_fallback_for_private_encoding(src_code);
765         conv_func = conv_get_code_conv_func(src_code, dest_code);
766         if (conv_func == conv_ustodisp && strict_mode && !is_ascii_str(inbuf))
767                 return NULL;
768
769         if (conv_func != conv_noconv) {
770                 len = (strlen(inbuf) + 1) * 3;
771                 buf = g_malloc(len);
772                 if (!buf) return NULL;
773
774                 if (conv_func(buf, len, inbuf) == 0 || !strict_mode)
775                         return g_realloc(buf, strlen(buf) + 1);
776                 else {
777                         g_free(buf);
778                         return NULL;
779                 }
780         }
781
782         return conv_iconv_strdup(inbuf, src_code, dest_code);
783 }
784
785 static CodeConvFunc conv_get_code_conv_func(const gchar *src_charset_str,
786                                      const gchar *dest_charset_str)
787 {
788         CodeConvFunc code_conv = conv_noconv;
789         CharSet src_charset;
790         CharSet dest_charset;
791
792         if (!src_charset_str)
793                 src_charset = conv_get_locale_charset();
794         else
795                 src_charset = conv_get_charset_from_str(src_charset_str);
796
797         /* auto detection mode */
798         if (!src_charset_str && !dest_charset_str) {
799                 if (conv_is_ja_locale())
800                         return conv_anytodisp;
801                 else
802                         return conv_noconv;
803         }
804
805         dest_charset = conv_get_charset_from_str(dest_charset_str);
806
807         if (dest_charset == C_US_ASCII)
808                 return conv_ustodisp;
809
810         switch (src_charset) {
811         case C_US_ASCII:
812         case C_ISO_8859_1:
813         case C_ISO_8859_2:
814         case C_ISO_8859_3:
815         case C_ISO_8859_4:
816         case C_ISO_8859_5:
817         case C_ISO_8859_6:
818         case C_ISO_8859_7:
819         case C_ISO_8859_8:
820         case C_ISO_8859_9:
821         case C_ISO_8859_10:
822         case C_ISO_8859_11:
823         case C_ISO_8859_13:
824         case C_ISO_8859_14:
825         case C_ISO_8859_15:
826                 break;
827         case C_ISO_2022_JP:
828         case C_ISO_2022_JP_2:
829         case C_ISO_2022_JP_3:
830                 if (dest_charset == C_AUTO)
831                         code_conv = conv_jistodisp;
832                 else if (dest_charset == C_EUC_JP)
833                         code_conv = conv_jistoeuc;
834                 else if (dest_charset == C_UTF_8)
835                         code_conv = conv_jistoutf8;
836                 break;
837         case C_SHIFT_JIS:
838                 if (dest_charset == C_AUTO)
839                         code_conv = conv_sjistodisp;
840                 else if (dest_charset == C_EUC_JP)
841                         code_conv = conv_sjistoeuc;
842                 else if (dest_charset == C_UTF_8)
843                         code_conv = conv_sjistoutf8;
844                 break;
845         case C_EUC_JP:
846                 if (dest_charset == C_AUTO)
847                         code_conv = conv_euctodisp;
848                 else if (dest_charset == C_ISO_2022_JP   ||
849                          dest_charset == C_ISO_2022_JP_2 ||
850                          dest_charset == C_ISO_2022_JP_3)
851                         code_conv = conv_euctojis;
852                 else if (dest_charset == C_UTF_8)
853                         code_conv = conv_euctoutf8;
854                 break;
855         case C_UTF_8:
856                 if (dest_charset == C_EUC_JP)
857                         code_conv = conv_utf8toeuc;
858                 else if (dest_charset == C_ISO_2022_JP   ||
859                          dest_charset == C_ISO_2022_JP_2 ||
860                          dest_charset == C_ISO_2022_JP_3)
861                         code_conv = conv_utf8tojis;
862                 break;
863         default:
864                 break;
865         }
866
867         return code_conv;
868 }
869
870 static gchar *conv_iconv_strdup(const gchar *inbuf,
871                          const gchar *src_code, const gchar *dest_code)
872 {
873         iconv_t cd;
874         gchar *outbuf;
875
876         if (!src_code && !dest_code && 
877             g_utf8_validate(inbuf, -1, NULL))
878                 return g_strdup(inbuf);
879
880         if (!src_code)
881                 src_code = conv_get_outgoing_charset_str();
882         if (!dest_code)
883                 dest_code = CS_INTERNAL;
884
885         /* don't convert if src and dest codeset are identical */
886         if (!strcasecmp(src_code, dest_code))
887                 return g_strdup(inbuf);
888
889         /* don't convert if dest codeset is US-ASCII */
890         if (!strcasecmp(src_code, CS_US_ASCII))
891                 return g_strdup(inbuf);
892
893         /* don't convert if dest codeset is US-ASCII */
894         if (!strcasecmp(dest_code, CS_US_ASCII))
895                 return g_strdup(inbuf);
896
897         cd = iconv_open(dest_code, src_code);
898         if (cd == (iconv_t)-1)
899                 return NULL;
900
901         outbuf = conv_iconv_strdup_with_cd(inbuf, cd);
902
903         iconv_close(cd);
904
905         return outbuf;
906 }
907
908 gchar *conv_iconv_strdup_with_cd(const gchar *inbuf, iconv_t cd)
909 {
910         const gchar *inbuf_p;
911         gchar *outbuf;
912         gchar *outbuf_p;
913         size_t in_size;
914         size_t in_left;
915         size_t out_size;
916         size_t out_left;
917         size_t n_conv;
918         size_t len;
919
920         inbuf_p = inbuf;
921         in_size = strlen(inbuf);
922         in_left = in_size;
923         out_size = (in_size + 1) * 2;
924         outbuf = g_malloc(out_size);
925         outbuf_p = outbuf;
926         out_left = out_size;
927
928 #define EXPAND_BUF()                            \
929 {                                               \
930         len = outbuf_p - outbuf;                \
931         out_size *= 2;                          \
932         outbuf = g_realloc(outbuf, out_size);   \
933         outbuf_p = outbuf + len;                \
934         out_left = out_size - len;              \
935 }
936
937         while ((n_conv = iconv(cd, (ICONV_CONST gchar **)&inbuf_p, &in_left,
938                                &outbuf_p, &out_left)) == (size_t)-1) {
939                 if (EILSEQ == errno) {
940                         if (strict_mode) {
941                                 g_free(outbuf);
942                                 return NULL;
943                         }
944                         //g_print("iconv(): at %d: %s\n", in_size - in_left, g_strerror(errno));
945                         inbuf_p++;
946                         in_left--;
947                         if (out_left == 0) {
948                                 EXPAND_BUF();
949                         }
950                         *outbuf_p++ = SUBST_CHAR;
951                         out_left--;
952                 } else if (EINVAL == errno) {
953                         break;
954                 } else if (E2BIG == errno) {
955                         EXPAND_BUF();
956                 } else {
957                         g_warning("conv_iconv_strdup(): %s\n",
958                                   g_strerror(errno));
959                         break;
960                 }
961         }
962
963         while ((n_conv = iconv(cd, NULL, NULL, &outbuf_p, &out_left)) ==
964                (size_t)-1) {
965                 if (E2BIG == errno) {
966                         EXPAND_BUF();
967                 } else {
968                         g_warning("conv_iconv_strdup(): %s\n",
969                                   g_strerror(errno));
970                         break;
971                 }
972         }
973
974 #undef EXPAND_BUF
975
976         len = outbuf_p - outbuf;
977         outbuf = g_realloc(outbuf, len + 1);
978         outbuf[len] = '\0';
979
980         return outbuf;
981 }
982
983 static const struct {
984         CharSet charset;
985         gchar *const name;
986 } charsets[] = {
987         {C_US_ASCII,            CS_US_ASCII},
988         {C_US_ASCII,            CS_ANSI_X3_4_1968},
989         {C_UTF_8,               CS_UTF_8},
990         {C_UTF_7,               CS_UTF_7},
991         {C_ISO_8859_1,          CS_ISO_8859_1},
992         {C_ISO_8859_2,          CS_ISO_8859_2},
993         {C_ISO_8859_3,          CS_ISO_8859_3},
994         {C_ISO_8859_4,          CS_ISO_8859_4},
995         {C_ISO_8859_5,          CS_ISO_8859_5},
996         {C_ISO_8859_6,          CS_ISO_8859_6},
997         {C_ISO_8859_7,          CS_ISO_8859_7},
998         {C_ISO_8859_8,          CS_ISO_8859_8},
999         {C_ISO_8859_9,          CS_ISO_8859_9},
1000         {C_ISO_8859_10,         CS_ISO_8859_10},
1001         {C_ISO_8859_11,         CS_ISO_8859_11},
1002         {C_ISO_8859_13,         CS_ISO_8859_13},
1003         {C_ISO_8859_14,         CS_ISO_8859_14},
1004         {C_ISO_8859_15,         CS_ISO_8859_15},
1005         {C_BALTIC,              CS_BALTIC},
1006         {C_CP1250,              CS_CP1250},
1007         {C_CP1251,              CS_CP1251},
1008         {C_CP1252,              CS_CP1252},
1009         {C_CP1253,              CS_CP1253},
1010         {C_CP1254,              CS_CP1254},
1011         {C_CP1255,              CS_CP1255},
1012         {C_CP1256,              CS_CP1256},
1013         {C_CP1257,              CS_CP1257},
1014         {C_CP1258,              CS_CP1258},
1015         {C_WINDOWS_1250,        CS_WINDOWS_1250},
1016         {C_WINDOWS_1251,        CS_WINDOWS_1251},
1017         {C_WINDOWS_1252,        CS_WINDOWS_1252},
1018         {C_WINDOWS_1253,        CS_WINDOWS_1253},
1019         {C_WINDOWS_1254,        CS_WINDOWS_1254},
1020         {C_WINDOWS_1255,        CS_WINDOWS_1255},
1021         {C_WINDOWS_1256,        CS_WINDOWS_1256},
1022         {C_WINDOWS_1257,        CS_WINDOWS_1257},
1023         {C_WINDOWS_1258,        CS_WINDOWS_1258},
1024         {C_KOI8_R,              CS_KOI8_R},
1025         {C_KOI8_T,              CS_KOI8_T},
1026         {C_KOI8_U,              CS_KOI8_U},
1027         {C_ISO_2022_JP,         CS_ISO_2022_JP},
1028         {C_ISO_2022_JP_2,       CS_ISO_2022_JP_2},
1029         {C_ISO_2022_JP_3,       CS_ISO_2022_JP_3},
1030         {C_EUC_JP,              CS_EUC_JP},
1031         {C_EUC_JP,              CS_EUCJP},
1032         {C_EUC_JP_MS,           CS_EUC_JP_MS},
1033         {C_SHIFT_JIS,           CS_SHIFT_JIS},
1034         {C_SHIFT_JIS,           CS_SHIFT__JIS},
1035         {C_SHIFT_JIS,           CS_SJIS},
1036         {C_ISO_2022_KR,         CS_ISO_2022_KR},
1037         {C_EUC_KR,              CS_EUC_KR},
1038         {C_ISO_2022_CN,         CS_ISO_2022_CN},
1039         {C_EUC_CN,              CS_EUC_CN},
1040         {C_GB18030,             CS_GB18030},
1041         {C_GB2312,              CS_GB2312},
1042         {C_GBK,                 CS_GBK},
1043         {C_EUC_TW,              CS_EUC_TW},
1044         {C_BIG5,                CS_BIG5},
1045         {C_BIG5_HKSCS,          CS_BIG5_HKSCS},
1046         {C_TIS_620,             CS_TIS_620},
1047         {C_WINDOWS_874,         CS_WINDOWS_874},
1048         {C_GEORGIAN_PS,         CS_GEORGIAN_PS},
1049         {C_TCVN5712_1,          CS_TCVN5712_1},
1050 };
1051
1052 static const struct {
1053         gchar *const locale;
1054         CharSet charset;
1055         CharSet out_charset;
1056 } locale_table[] = {
1057         {"ja_JP.eucJP"  , C_EUC_JP      , C_ISO_2022_JP},
1058         {"ja_JP.EUC-JP" , C_EUC_JP      , C_ISO_2022_JP},
1059         {"ja_JP.EUC"    , C_EUC_JP      , C_ISO_2022_JP},
1060         {"ja_JP.ujis"   , C_EUC_JP      , C_ISO_2022_JP},
1061         {"ja_JP.SJIS"   , C_SHIFT_JIS   , C_ISO_2022_JP},
1062         {"ja_JP.JIS"    , C_ISO_2022_JP , C_ISO_2022_JP},
1063 #ifdef G_OS_WIN32
1064         {"ja_JP"        , C_SHIFT_JIS   , C_ISO_2022_JP},
1065 #else
1066         {"ja_JP"        , C_EUC_JP      , C_ISO_2022_JP},
1067 #endif
1068         {"ko_KR.EUC-KR" , C_EUC_KR      , C_EUC_KR},
1069         {"ko_KR"        , C_EUC_KR      , C_EUC_KR},
1070         {"zh_CN.GB18030"        , C_GB18030     , C_GB18030},
1071         {"zh_CN.GB2312" , C_GB2312      , C_GB2312},
1072         {"zh_CN.GBK"    , C_GBK         , C_GBK},
1073         {"zh_CN"        , C_GB18030     , C_GB18030},
1074         {"zh_HK"        , C_BIG5_HKSCS  , C_BIG5_HKSCS},
1075         {"zh_TW.eucTW"  , C_EUC_TW      , C_BIG5},
1076         {"zh_TW.EUC-TW" , C_EUC_TW      , C_BIG5},
1077         {"zh_TW.Big5"   , C_BIG5        , C_BIG5},
1078         {"zh_TW"        , C_BIG5        , C_BIG5},
1079
1080         {"ru_RU.KOI8-R" , C_KOI8_R      , C_KOI8_R},
1081         {"ru_RU.KOI8R"  , C_KOI8_R      , C_KOI8_R},
1082         {"ru_RU.CP1251" , C_WINDOWS_1251, C_KOI8_R},
1083 #ifdef G_OS_WIN32
1084         {"ru_RU"        , C_WINDOWS_1251, C_KOI8_R},
1085 #else
1086         {"ru_RU"        , C_ISO_8859_5  , C_KOI8_R},
1087 #endif
1088         {"tg_TJ"        , C_KOI8_T      , C_KOI8_T},
1089         {"ru_UA"        , C_KOI8_U      , C_KOI8_U},
1090         {"uk_UA.CP1251" , C_WINDOWS_1251, C_KOI8_U},
1091         {"uk_UA"        , C_KOI8_U      , C_KOI8_U},
1092
1093         {"be_BY"        , C_WINDOWS_1251, C_WINDOWS_1251},
1094         {"bg_BG"        , C_WINDOWS_1251, C_WINDOWS_1251},
1095
1096         {"yi_US"        , C_WINDOWS_1255, C_WINDOWS_1255},
1097
1098         {"af_ZA"        , C_ISO_8859_1  , C_ISO_8859_1},
1099         {"br_FR"        , C_ISO_8859_1  , C_ISO_8859_1},
1100         {"ca_ES"        , C_ISO_8859_1  , C_ISO_8859_1},
1101         {"da_DK"        , C_ISO_8859_1  , C_ISO_8859_1},
1102         {"de_AT"        , C_ISO_8859_1  , C_ISO_8859_1},
1103         {"de_BE"        , C_ISO_8859_1  , C_ISO_8859_1},
1104         {"de_CH"        , C_ISO_8859_1  , C_ISO_8859_1},
1105         {"de_DE"        , C_ISO_8859_1  , C_ISO_8859_1},
1106         {"de_LU"        , C_ISO_8859_1  , C_ISO_8859_1},
1107         {"en_AU"        , C_ISO_8859_1  , C_ISO_8859_1},
1108         {"en_BW"        , C_ISO_8859_1  , C_ISO_8859_1},
1109         {"en_CA"        , C_ISO_8859_1  , C_ISO_8859_1},
1110         {"en_DK"        , C_ISO_8859_1  , C_ISO_8859_1},
1111         {"en_GB"        , C_ISO_8859_1  , C_ISO_8859_1},
1112         {"en_HK"        , C_ISO_8859_1  , C_ISO_8859_1},
1113         {"en_IE"        , C_ISO_8859_1  , C_ISO_8859_1},
1114         {"en_NZ"        , C_ISO_8859_1  , C_ISO_8859_1},
1115         {"en_PH"        , C_ISO_8859_1  , C_ISO_8859_1},
1116         {"en_SG"        , C_ISO_8859_1  , C_ISO_8859_1},
1117         {"en_US"        , C_ISO_8859_1  , C_ISO_8859_1},
1118         {"en_ZA"        , C_ISO_8859_1  , C_ISO_8859_1},
1119         {"en_ZW"        , C_ISO_8859_1  , C_ISO_8859_1},
1120         {"es_AR"        , C_ISO_8859_1  , C_ISO_8859_1},
1121         {"es_BO"        , C_ISO_8859_1  , C_ISO_8859_1},
1122         {"es_CL"        , C_ISO_8859_1  , C_ISO_8859_1},
1123         {"es_CO"        , C_ISO_8859_1  , C_ISO_8859_1},
1124         {"es_CR"        , C_ISO_8859_1  , C_ISO_8859_1},
1125         {"es_DO"        , C_ISO_8859_1  , C_ISO_8859_1},
1126         {"es_EC"        , C_ISO_8859_1  , C_ISO_8859_1},
1127         {"es_ES"        , C_ISO_8859_1  , C_ISO_8859_1},
1128         {"es_GT"        , C_ISO_8859_1  , C_ISO_8859_1},
1129         {"es_HN"        , C_ISO_8859_1  , C_ISO_8859_1},
1130         {"es_MX"        , C_ISO_8859_1  , C_ISO_8859_1},
1131         {"es_NI"        , C_ISO_8859_1  , C_ISO_8859_1},
1132         {"es_PA"        , C_ISO_8859_1  , C_ISO_8859_1},
1133         {"es_PE"        , C_ISO_8859_1  , C_ISO_8859_1},
1134         {"es_PR"        , C_ISO_8859_1  , C_ISO_8859_1},
1135         {"es_PY"        , C_ISO_8859_1  , C_ISO_8859_1},
1136         {"es_SV"        , C_ISO_8859_1  , C_ISO_8859_1},
1137         {"es_US"        , C_ISO_8859_1  , C_ISO_8859_1},
1138         {"es_UY"        , C_ISO_8859_1  , C_ISO_8859_1},
1139         {"es_VE"        , C_ISO_8859_1  , C_ISO_8859_1},
1140         {"et_EE"        , C_ISO_8859_1  , C_ISO_8859_1},
1141         {"eu_ES"        , C_ISO_8859_1  , C_ISO_8859_1},
1142         {"fi_FI"        , C_ISO_8859_1  , C_ISO_8859_1},
1143         {"fo_FO"        , C_ISO_8859_1  , C_ISO_8859_1},
1144         {"fr_BE"        , C_ISO_8859_1  , C_ISO_8859_1},
1145         {"fr_CA"        , C_ISO_8859_1  , C_ISO_8859_1},
1146         {"fr_CH"        , C_ISO_8859_1  , C_ISO_8859_1},
1147         {"fr_FR"        , C_ISO_8859_1  , C_ISO_8859_1},
1148         {"fr_LU"        , C_ISO_8859_1  , C_ISO_8859_1},
1149         {"ga_IE"        , C_ISO_8859_1  , C_ISO_8859_1},
1150         {"gl_ES"        , C_ISO_8859_1  , C_ISO_8859_1},
1151         {"gv_GB"        , C_ISO_8859_1  , C_ISO_8859_1},
1152         {"id_ID"        , C_ISO_8859_1  , C_ISO_8859_1},
1153         {"is_IS"        , C_ISO_8859_1  , C_ISO_8859_1},
1154         {"it_CH"        , C_ISO_8859_1  , C_ISO_8859_1},
1155         {"it_IT"        , C_ISO_8859_1  , C_ISO_8859_1},
1156         {"kl_GL"        , C_ISO_8859_1  , C_ISO_8859_1},
1157         {"kw_GB"        , C_ISO_8859_1  , C_ISO_8859_1},
1158         {"ms_MY"        , C_ISO_8859_1  , C_ISO_8859_1},
1159         {"nl_BE"        , C_ISO_8859_1  , C_ISO_8859_1},
1160         {"nl_NL"        , C_ISO_8859_1  , C_ISO_8859_1},
1161         {"nb_NO"        , C_ISO_8859_1  , C_ISO_8859_1},
1162         {"nn_NO"        , C_ISO_8859_1  , C_ISO_8859_1},
1163         {"no_NO"        , C_ISO_8859_1  , C_ISO_8859_1},
1164         {"oc_FR"        , C_ISO_8859_1  , C_ISO_8859_1},
1165         {"pt_BR"        , C_ISO_8859_1  , C_ISO_8859_1},
1166         {"pt_PT"        , C_ISO_8859_1  , C_ISO_8859_1},
1167         {"sq_AL"        , C_ISO_8859_1  , C_ISO_8859_1},
1168         {"sv_FI"        , C_ISO_8859_1  , C_ISO_8859_1},
1169         {"sv_SE"        , C_ISO_8859_1  , C_ISO_8859_1},
1170         {"tl_PH"        , C_ISO_8859_1  , C_ISO_8859_1},
1171         {"uz_UZ"        , C_ISO_8859_1  , C_ISO_8859_1},
1172         {"wa_BE"        , C_ISO_8859_1  , C_ISO_8859_1},
1173
1174         {"bs_BA"        , C_ISO_8859_2  , C_ISO_8859_2},
1175         {"cs_CZ"        , C_ISO_8859_2  , C_ISO_8859_2},
1176         {"hr_HR"        , C_ISO_8859_2  , C_ISO_8859_2},
1177         {"hu_HU"        , C_ISO_8859_2  , C_ISO_8859_2},
1178         {"pl_PL"        , C_ISO_8859_2  , C_ISO_8859_2},
1179         {"ro_RO"        , C_ISO_8859_2  , C_ISO_8859_2},
1180         {"sk_SK"        , C_ISO_8859_2  , C_ISO_8859_2},
1181         {"sl_SI"        , C_ISO_8859_2  , C_ISO_8859_2},
1182
1183         {"sr_YU@cyrillic"       , C_ISO_8859_5  , C_ISO_8859_5},
1184         {"sr_YU"                , C_ISO_8859_2  , C_ISO_8859_2},
1185
1186         {"mt_MT"                , C_ISO_8859_3  , C_ISO_8859_3},
1187
1188         {"lt_LT.iso88594"       , C_ISO_8859_4  , C_ISO_8859_4},
1189         {"lt_LT.ISO8859-4"      , C_ISO_8859_4  , C_ISO_8859_4},
1190         {"lt_LT.ISO_8859-4"     , C_ISO_8859_4  , C_ISO_8859_4},
1191         {"lt_LT"                , C_ISO_8859_13 , C_ISO_8859_13},
1192
1193         {"mk_MK"        , C_ISO_8859_5  , C_ISO_8859_5},
1194
1195         {"ar_AE"        , C_ISO_8859_6  , C_ISO_8859_6},
1196         {"ar_BH"        , C_ISO_8859_6  , C_ISO_8859_6},
1197         {"ar_DZ"        , C_ISO_8859_6  , C_ISO_8859_6},
1198         {"ar_EG"        , C_ISO_8859_6  , C_ISO_8859_6},
1199         {"ar_IQ"        , C_ISO_8859_6  , C_ISO_8859_6},
1200         {"ar_JO"        , C_ISO_8859_6  , C_ISO_8859_6},
1201         {"ar_KW"        , C_ISO_8859_6  , C_ISO_8859_6},
1202         {"ar_LB"        , C_ISO_8859_6  , C_ISO_8859_6},
1203         {"ar_LY"        , C_ISO_8859_6  , C_ISO_8859_6},
1204         {"ar_MA"        , C_ISO_8859_6  , C_ISO_8859_6},
1205         {"ar_OM"        , C_ISO_8859_6  , C_ISO_8859_6},
1206         {"ar_QA"        , C_ISO_8859_6  , C_ISO_8859_6},
1207         {"ar_SA"        , C_ISO_8859_6  , C_ISO_8859_6},
1208         {"ar_SD"        , C_ISO_8859_6  , C_ISO_8859_6},
1209         {"ar_SY"        , C_ISO_8859_6  , C_ISO_8859_6},
1210         {"ar_TN"        , C_ISO_8859_6  , C_ISO_8859_6},
1211         {"ar_YE"        , C_ISO_8859_6  , C_ISO_8859_6},
1212
1213         {"el_GR"        , C_ISO_8859_7  , C_ISO_8859_7},
1214         {"he_IL"        , C_ISO_8859_8  , C_ISO_8859_8},
1215         {"iw_IL"        , C_ISO_8859_8  , C_ISO_8859_8},
1216         {"tr_TR"        , C_ISO_8859_9  , C_ISO_8859_9},
1217
1218         {"lv_LV"        , C_ISO_8859_13 , C_ISO_8859_13},
1219         {"mi_NZ"        , C_ISO_8859_13 , C_ISO_8859_13},
1220
1221         {"cy_GB"        , C_ISO_8859_14 , C_ISO_8859_14},
1222
1223         {"ar_IN"        , C_UTF_8       , C_UTF_8},
1224         {"en_IN"        , C_UTF_8       , C_UTF_8},
1225         {"se_NO"        , C_UTF_8       , C_UTF_8},
1226         {"ta_IN"        , C_UTF_8       , C_UTF_8},
1227         {"te_IN"        , C_UTF_8       , C_UTF_8},
1228         {"ur_PK"        , C_UTF_8       , C_UTF_8},
1229
1230         {"th_TH"        , C_TIS_620     , C_TIS_620},
1231         /* {"th_TH"     , C_WINDOWS_874}, */
1232         /* {"th_TH"     , C_ISO_8859_11}, */
1233
1234         {"ka_GE"        , C_GEORGIAN_PS , C_GEORGIAN_PS},
1235         {"vi_VN.TCVN"   , C_TCVN5712_1  , C_TCVN5712_1},
1236
1237         {"C"                    , C_US_ASCII    , C_US_ASCII},
1238         {"POSIX"                , C_US_ASCII    , C_US_ASCII},
1239         {"ANSI_X3.4-1968"       , C_US_ASCII    , C_US_ASCII},
1240 };
1241
1242 static GHashTable *conv_get_charset_to_str_table(void)
1243 {
1244         static GHashTable *table;
1245         gint i;
1246
1247         if (table)
1248                 return table;
1249
1250         table = g_hash_table_new(NULL, g_direct_equal);
1251
1252         for (i = 0; i < sizeof(charsets) / sizeof(charsets[0]); i++) {
1253                 if (g_hash_table_lookup(table, GUINT_TO_POINTER(charsets[i].charset))
1254                     == NULL) {
1255                         g_hash_table_insert
1256                                 (table, GUINT_TO_POINTER(charsets[i].charset),
1257                                  charsets[i].name);
1258                 }
1259         }
1260
1261         return table;
1262 }
1263
1264 static GHashTable *conv_get_charset_from_str_table(void)
1265 {
1266         static GHashTable *table;
1267         gint i;
1268
1269         if (table)
1270                 return table;
1271
1272         table = g_hash_table_new(str_case_hash, str_case_equal);
1273
1274         for (i = 0; i < sizeof(charsets) / sizeof(charsets[0]); i++) {
1275                 g_hash_table_insert(table, charsets[i].name,
1276                                     GUINT_TO_POINTER(charsets[i].charset));
1277         }
1278
1279         return table;
1280 }
1281
1282 const gchar *conv_get_charset_str(CharSet charset)
1283 {
1284         GHashTable *table;
1285
1286         table = conv_get_charset_to_str_table();
1287         return g_hash_table_lookup(table, GUINT_TO_POINTER(charset));
1288 }
1289
1290 CharSet conv_get_charset_from_str(const gchar *charset)
1291 {
1292         GHashTable *table;
1293
1294         if (!charset) return C_AUTO;
1295
1296         table = conv_get_charset_from_str_table();
1297         return GPOINTER_TO_UINT(g_hash_table_lookup(table, charset));
1298 }
1299
1300 static CharSet conv_get_locale_charset(void)
1301 {
1302         static CharSet cur_charset = -1;
1303         const gchar *cur_locale;
1304         const gchar *p;
1305         gint i;
1306
1307         if (cur_charset != -1)
1308                 return cur_charset;
1309
1310         cur_locale = conv_get_current_locale();
1311         if (!cur_locale) {
1312                 cur_charset = C_US_ASCII;
1313                 return cur_charset;
1314         }
1315
1316         if (strcasestr(cur_locale, "UTF-8") ||
1317             strcasestr(cur_locale, "utf8")) {
1318                 cur_charset = C_UTF_8;
1319                 return cur_charset;
1320         }
1321
1322         if ((p = strcasestr(cur_locale, "@euro")) && p[5] == '\0') {
1323                 cur_charset = C_ISO_8859_15;
1324                 return cur_charset;
1325         }
1326
1327         for (i = 0; i < sizeof(locale_table) / sizeof(locale_table[0]); i++) {
1328                 const gchar *p;
1329
1330                 /* "ja_JP.EUC" matches with "ja_JP.eucJP", "ja_JP.EUC" and
1331                    "ja_JP". "ja_JP" matches with "ja_JP.xxxx" and "ja" */
1332                 if (!g_ascii_strncasecmp(cur_locale, locale_table[i].locale,
1333                                  strlen(locale_table[i].locale))) {
1334                         cur_charset = locale_table[i].charset;
1335                         return cur_charset;
1336                 } else if ((p = strchr(locale_table[i].locale, '_')) &&
1337                          !strchr(p + 1, '.')) {
1338                         if (strlen(cur_locale) == 2 &&
1339                             !g_ascii_strncasecmp(cur_locale, locale_table[i].locale, 2)) {
1340                                 cur_charset = locale_table[i].charset;
1341                                 return cur_charset;
1342                         }
1343                 }
1344         }
1345
1346         cur_charset = C_AUTO;
1347         return cur_charset;
1348 }
1349
1350 static CharSet conv_get_locale_charset_no_utf8(void)
1351 {
1352         static CharSet cur_charset = -1;
1353         const gchar *cur_locale;
1354         const gchar *p;
1355         gint i;
1356
1357         if (prefs_common.broken_are_utf8) {
1358                 cur_charset = C_UTF_8;
1359                 return cur_charset;
1360         }
1361
1362         cur_locale = conv_get_current_locale();
1363         if (!cur_locale) {
1364                 cur_charset = C_US_ASCII;
1365                 return cur_charset;
1366         }
1367
1368         if (strcasestr(cur_locale, "UTF-8") ||
1369             strcasestr(cur_locale, "utf8")) {
1370                 cur_charset = C_UTF_8;
1371                 return cur_charset;
1372         }
1373
1374         if ((p = strcasestr(cur_locale, "@euro")) && p[5] == '\0') {
1375                 cur_charset = C_ISO_8859_15;
1376                 return cur_charset;
1377         }
1378
1379         for (i = 0; i < sizeof(locale_table) / sizeof(locale_table[0]); i++) {
1380                 const gchar *p;
1381
1382                 /* "ja_JP.EUC" matches with "ja_JP.eucJP", "ja_JP.EUC" and
1383                    "ja_JP". "ja_JP" matches with "ja_JP.xxxx" and "ja" */
1384                 if (!g_ascii_strncasecmp(cur_locale, locale_table[i].locale,
1385                                  strlen(locale_table[i].locale))) {
1386                         cur_charset = locale_table[i].charset;
1387                         return cur_charset;
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                                 cur_charset = locale_table[i].charset;
1393                                 return cur_charset;
1394                         }
1395                 }
1396         }
1397
1398         cur_charset = C_AUTO;
1399         return cur_charset;
1400 }
1401
1402 const gchar *conv_get_locale_charset_str(void)
1403 {
1404         static const gchar *codeset = NULL;
1405
1406         if (!codeset)
1407                 codeset = conv_get_charset_str(conv_get_locale_charset());
1408
1409         return codeset ? codeset : CS_INTERNAL;
1410 }
1411
1412 const gchar *conv_get_locale_charset_str_no_utf8(void)
1413 {
1414         static const gchar *codeset = NULL;
1415
1416         if (!codeset)
1417                 codeset = conv_get_charset_str(conv_get_locale_charset_no_utf8());
1418
1419         return codeset ? codeset : CS_INTERNAL;
1420 }
1421
1422 static CharSet conv_get_outgoing_charset(void)
1423 {
1424         static CharSet out_charset = -1;
1425         const gchar *cur_locale;
1426         const gchar *p;
1427         gint i;
1428
1429         if (out_charset != -1)
1430                 return out_charset;
1431
1432         cur_locale = conv_get_current_locale();
1433         if (!cur_locale) {
1434                 out_charset = C_AUTO;
1435                 return out_charset;
1436         }
1437
1438         if (strcasestr(cur_locale, "UTF-8") ||
1439             strcasestr(cur_locale, "utf8")) {
1440                 out_charset = C_UTF_8;
1441                 return out_charset;
1442         }
1443
1444         if ((p = strcasestr(cur_locale, "@euro")) && p[5] == '\0') {
1445                 out_charset = C_ISO_8859_15;
1446                 return out_charset;
1447         }
1448
1449         for (i = 0; i < sizeof(locale_table) / sizeof(locale_table[0]); i++) {
1450                 const gchar *p;
1451
1452                 if (!g_ascii_strncasecmp(cur_locale, locale_table[i].locale,
1453                                  strlen(locale_table[i].locale))) {
1454                         out_charset = locale_table[i].out_charset;
1455                         break;
1456                 } else if ((p = strchr(locale_table[i].locale, '_')) &&
1457                          !strchr(p + 1, '.')) {
1458                         if (strlen(cur_locale) == 2 &&
1459                             !g_ascii_strncasecmp(cur_locale, locale_table[i].locale, 2)) {
1460                                 out_charset = locale_table[i].out_charset;
1461                                 break;
1462                         }
1463                 }
1464         }
1465
1466         return out_charset;
1467 }
1468
1469 const gchar *conv_get_outgoing_charset_str(void)
1470 {
1471         CharSet out_charset;
1472         const gchar *str;
1473
1474         out_charset = conv_get_outgoing_charset();
1475         str = conv_get_charset_str(out_charset);
1476
1477         return str ? str : CS_UTF_8;
1478 }
1479
1480 const gchar *conv_get_current_locale(void)
1481 {
1482         const gchar *cur_locale;
1483
1484 #ifdef G_OS_WIN32
1485         cur_locale = g_win32_getlocale();
1486 #else
1487         cur_locale = g_getenv("LC_ALL");
1488         if (!cur_locale) cur_locale = g_getenv("LC_CTYPE");
1489         if (!cur_locale) cur_locale = g_getenv("LANG");
1490         if (!cur_locale) cur_locale = setlocale(LC_CTYPE, NULL);
1491 #endif /* G_OS_WIN32 */
1492
1493         debug_print("current locale: %s\n",
1494                     cur_locale ? cur_locale : "(none)");
1495
1496         return cur_locale;
1497 }
1498
1499 static gboolean conv_is_ja_locale(void)
1500 {
1501         static gint is_ja_locale = -1;
1502         const gchar *cur_locale;
1503
1504         if (is_ja_locale != -1)
1505                 return is_ja_locale != 0;
1506
1507         is_ja_locale = 0;
1508         cur_locale = conv_get_current_locale();
1509         if (cur_locale) {
1510                 if (g_ascii_strncasecmp(cur_locale, "ja", 2) == 0)
1511                         is_ja_locale = 1;
1512         }
1513
1514         return is_ja_locale != 0;
1515 }
1516
1517 gchar *conv_unmime_header(const gchar *str, const gchar *default_encoding,
1518                            gboolean addr_field)
1519 {
1520         gchar buf[BUFFSIZE];
1521
1522         if (is_ascii_str(str))
1523                 return unmime_header(str, addr_field);
1524
1525         if (default_encoding) {
1526                 gchar *utf8_buf;
1527
1528                 utf8_buf = conv_codeset_strdup
1529                         (str, default_encoding, CS_INTERNAL);
1530                 if (utf8_buf) {
1531                         gchar *decoded_str;
1532
1533                         decoded_str = unmime_header(utf8_buf, addr_field);
1534                         g_free(utf8_buf);
1535                         return decoded_str;
1536                 }
1537         }
1538
1539         if (conv_is_ja_locale())
1540                 conv_anytodisp(buf, sizeof(buf), str);
1541         else
1542                 conv_localetodisp(buf, sizeof(buf), str);
1543
1544         return unmime_header(buf, addr_field);
1545 }
1546
1547 #define MAX_LINELEN             76
1548 #define MAX_HARD_LINELEN        996
1549 #define MIMESEP_BEGIN           "=?"
1550 #define MIMESEP_END             "?="
1551
1552 #define LBREAK_IF_REQUIRED(cond, is_plain_text)                         \
1553 {                                                                       \
1554         if (len - (destp - (guchar *)dest) < MAX_LINELEN + 2) {         \
1555                 *destp = '\0';                                          \
1556                 return;                                                 \
1557         }                                                               \
1558                                                                         \
1559         if ((cond) && *srcp) {                                          \
1560                 if (destp > (guchar *)dest && left < MAX_LINELEN - 1) { \
1561                         if (isspace(*(destp - 1)))                      \
1562                                 destp--;                                \
1563                         else if (is_plain_text && isspace(*srcp))       \
1564                                 srcp++;                                 \
1565                         if (*srcp) {                                    \
1566                                 *destp++ = '\n';                        \
1567                                 *destp++ = ' ';                         \
1568                                 left = MAX_LINELEN - 1;                 \
1569                         }                                               \
1570                 } else if (destp == (guchar *)dest && left < 7) {       \
1571                         if (isspace(*(destp - 1)))                      \
1572                                 destp--;                                \
1573                         else if (is_plain_text && isspace(*srcp))       \
1574                                 srcp++;                                 \
1575                         if (*srcp) {                                    \
1576                                 *destp++ = '\n';                        \
1577                                 *destp++ = ' ';                         \
1578                                 left = MAX_LINELEN - 1;                 \
1579                         }                                               \
1580                 }                                                       \
1581         }                                                               \
1582 }
1583
1584 #define B64LEN(len) ((len) / 3 * 4 + ((len) % 3 ? 4 : 0))
1585
1586 void conv_encode_header_full(gchar *dest, gint len, const gchar *src,
1587                         gint header_len, gboolean addr_field,
1588                         const gchar *out_encoding_)
1589 {
1590         const gchar *cur_encoding;
1591         const gchar *out_encoding;
1592         gint mimestr_len;
1593         gchar *mimesep_enc;
1594         gint left;
1595         const guchar *srcp = src;
1596         guchar *destp = dest;
1597         gboolean use_base64;
1598
1599         cm_return_if_fail(g_utf8_validate(src, -1, NULL) == TRUE);
1600         cm_return_if_fail(destp != NULL);
1601
1602         if (MB_CUR_MAX > 1) {
1603                 use_base64 = TRUE;
1604                 mimesep_enc = "?B?";
1605         } else {
1606                 use_base64 = FALSE;
1607                 mimesep_enc = "?Q?";
1608         }
1609
1610         cur_encoding = CS_INTERNAL;
1611
1612         if (out_encoding_)
1613                 out_encoding = out_encoding_;
1614         else
1615                 out_encoding = conv_get_outgoing_charset_str();
1616
1617         if (!strcmp(out_encoding, CS_US_ASCII))
1618                 out_encoding = CS_ISO_8859_1;
1619
1620         mimestr_len = strlen(MIMESEP_BEGIN) + strlen(out_encoding) +
1621                 strlen(mimesep_enc) + strlen(MIMESEP_END);
1622
1623         left = MAX_LINELEN - header_len;
1624
1625         while (*srcp) {
1626                 LBREAK_IF_REQUIRED(left <= 0, TRUE);
1627
1628                 while (isspace(*srcp)) {
1629                         *destp++ = *srcp++;
1630                         left--;
1631                         LBREAK_IF_REQUIRED(left <= 0, TRUE);
1632                 }
1633
1634                 /* output as it is if the next word is ASCII string */
1635                 if (!is_next_nonascii(srcp)) {
1636                         gint word_len;
1637
1638                         word_len = get_next_word_len(srcp);
1639                         LBREAK_IF_REQUIRED(left < word_len, TRUE);
1640                         while (word_len > 0) {
1641                                 LBREAK_IF_REQUIRED(left + (MAX_HARD_LINELEN - MAX_LINELEN) <= 0, TRUE)
1642                                 *destp++ = *srcp++;
1643                                 left--;
1644                                 word_len--;
1645                         }
1646
1647                         continue;
1648                 }
1649
1650                 /* don't include parentheses and quotes in encoded strings */
1651                 if (addr_field && (*srcp == '(' || *srcp == ')' || *srcp == '"')) {
1652                         LBREAK_IF_REQUIRED(left < 2, FALSE);
1653                         *destp++ = *srcp++;
1654                         left--;
1655                 }
1656
1657                 while (1) {
1658                         gint mb_len = 0;
1659                         gint cur_len = 0;
1660                         gchar *part_str;
1661                         gchar *out_str;
1662                         gchar *enc_str;
1663                         const guchar *p = srcp;
1664                         gint out_str_len;
1665                         gint out_enc_str_len;
1666                         gint mime_block_len;
1667                         gboolean cont = FALSE;
1668
1669                         while (*p != '\0') {
1670                                 if (isspace(*p) && !is_next_nonascii(p + 1))
1671                                         break;
1672                                 /* don't include parentheses in encoded
1673                                    strings */
1674                                 if (addr_field && (*p == '(' || *p == ')' || *p == '"'))
1675                                         break;
1676
1677                                 mb_len = g_utf8_skip[*p];
1678
1679                                 Xstrndup_a(part_str, srcp, cur_len + mb_len, );
1680                                 out_str = conv_codeset_strdup
1681                                         (part_str, cur_encoding, out_encoding);
1682                                 if (!out_str) {
1683                                         if (strict_mode) {
1684                                                 *dest = '\0';
1685                                                 return;
1686                                         } else {
1687                                                 g_warning("conv_encode_header(): code conversion failed\n");
1688                                                 conv_unreadable_8bit(part_str);
1689                                                 out_str = g_strdup(part_str);
1690                                         }
1691                                 }
1692                                 out_str_len = strlen(out_str);
1693
1694                                 if (use_base64)
1695                                         out_enc_str_len = B64LEN(out_str_len);
1696                                 else
1697                                         out_enc_str_len =
1698                                                 qp_get_q_encoding_len(out_str);
1699
1700                                 g_free(out_str);
1701
1702                                 if (mimestr_len + out_enc_str_len <= left) {
1703                                         cur_len += mb_len;
1704                                         p += mb_len;
1705                                 } else if (cur_len == 0) {
1706                                         left = 0;
1707                                         LBREAK_IF_REQUIRED(1, FALSE);
1708                                         continue;
1709                                 } else {
1710                                         cont = TRUE;
1711                                         break;
1712                                 }
1713                         }
1714
1715                         if (cur_len > 0) {
1716                                 Xstrndup_a(part_str, srcp, cur_len, );
1717                                 out_str = conv_codeset_strdup
1718                                         (part_str, cur_encoding, out_encoding);
1719                                 if (!out_str) {
1720                                         g_warning("conv_encode_header(): code conversion failed\n");
1721                                         conv_unreadable_8bit(part_str);
1722                                         out_str = g_strdup(part_str);
1723                                 }
1724                                 out_str_len = strlen(out_str);
1725
1726                                 if (use_base64)
1727                                         out_enc_str_len = B64LEN(out_str_len);
1728                                 else
1729                                         out_enc_str_len =
1730                                                 qp_get_q_encoding_len(out_str);
1731
1732                                 if (use_base64)
1733                                         enc_str = g_base64_encode(out_str, out_str_len);
1734                                 else {
1735                                         Xalloca(enc_str, out_enc_str_len + 1, );
1736                                         qp_q_encode(enc_str, out_str);
1737                                 }
1738
1739                                 g_free(out_str);
1740
1741                                 /* output MIME-encoded string block */
1742                                 mime_block_len = mimestr_len + strlen(enc_str);
1743                                 g_snprintf(destp, mime_block_len + 1,
1744                                            MIMESEP_BEGIN "%s%s%s" MIMESEP_END,
1745                                            out_encoding, mimesep_enc, enc_str);
1746
1747                                 if (use_base64)
1748                                         g_free(enc_str);
1749
1750                                 destp += mime_block_len;
1751                                 srcp += cur_len;
1752
1753                                 left -= mime_block_len;
1754                         }
1755
1756                         LBREAK_IF_REQUIRED(cont, FALSE);
1757
1758                         if (cur_len == 0)
1759                                 break;
1760                 }
1761         }
1762
1763         *destp = '\0';
1764 }
1765
1766 void conv_encode_header(gchar *dest, gint len, const gchar *src,
1767                         gint header_len, gboolean addr_field)
1768 {
1769         conv_encode_header_full(dest,len,src,header_len,addr_field,NULL);
1770 }
1771
1772 #undef LBREAK_IF_REQUIRED
1773 #undef B64LEN
1774
1775 gchar *conv_filename_from_utf8(const gchar *utf8_file)
1776 {
1777         gchar *fs_file;
1778         GError *error = NULL;
1779
1780         fs_file = g_filename_from_utf8(utf8_file, -1, NULL, NULL, &error);
1781         if (error) {
1782                 debug_print("failed to convert encoding of file name: %s\n",
1783                           error->message);
1784                 g_error_free(error);
1785         }
1786         if (!fs_file)
1787                 fs_file = g_strdup(utf8_file);
1788
1789         return fs_file;
1790 }
1791
1792 gchar *conv_filename_to_utf8(const gchar *fs_file)
1793 {
1794         gchar *utf8_file = NULL;
1795         GError *error = NULL;
1796
1797         utf8_file = g_filename_to_utf8(fs_file, -1, NULL, NULL, &error);
1798         if (error) {
1799                 g_warning("failed to convert encoding of file name: %s\n",
1800                           error->message);
1801                 g_error_free(error);
1802         }
1803
1804         if (!utf8_file || !g_utf8_validate(utf8_file, -1, NULL)) {
1805                 g_free(utf8_file);
1806                 utf8_file = g_strdup(fs_file);
1807                 conv_unreadable_8bit(utf8_file);
1808         }
1809
1810         return utf8_file;
1811 }