2005-10-22 [colin] 1.9.15cvs93
[claws.git] / src / codeconv.c
1 /*
2  * Sylpheed -- a GTK+ based, lightweight, and fast e-mail client
3  * Copyright (C) 1999-2005 Hiroyuki Yamamoto
4  *
5  * This program is free software; you can redistribute it and/or modify
6  * it under the terms of the GNU General Public License as published by
7  * the Free Software Foundation; either version 2 of the License, or
8  * (at your option) any later version.
9  *
10  * This program is distributed in the hope that it will be useful,
11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
13  * GNU General Public License for more details.
14  *
15  * You should have received a copy of the GNU General Public License
16  * along with this program; if not, write to the Free Software
17  * Foundation, Inc., 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 <iconv.h>
38
39 #include "codeconv.h"
40 #include "unmime.h"
41 #include "base64.h"
42 #include "quoted-printable.h"
43 #include "utils.h"
44 #include "prefs_common.h"
45
46 typedef enum
47 {
48         JIS_ASCII,
49         JIS_KANJI,
50         JIS_HWKANA,
51         JIS_AUXKANJI
52 } JISState;
53
54 #define SUBST_CHAR      0x5f;
55 #define ESC             '\033'
56
57 #define iseuckanji(c) \
58         (((c) & 0xff) >= 0xa1 && ((c) & 0xff) <= 0xfe)
59 #define iseuchwkana1(c) \
60         (((c) & 0xff) == 0x8e)
61 #define iseuchwkana2(c) \
62         (((c) & 0xff) >= 0xa1 && ((c) & 0xff) <= 0xdf)
63 #define iseucaux(c) \
64         (((c) & 0xff) == 0x8f)
65 #define issjiskanji1(c) \
66         ((((c) & 0xff) >= 0x81 && ((c) & 0xff) <= 0x9f) || \
67          (((c) & 0xff) >= 0xe0 && ((c) & 0xff) <= 0xfc))
68 #define issjiskanji2(c) \
69         ((((c) & 0xff) >= 0x40 && ((c) & 0xff) <= 0x7e) || \
70          (((c) & 0xff) >= 0x80 && ((c) & 0xff) <= 0xfc))
71 #define issjishwkana(c) \
72         (((c) & 0xff) >= 0xa1 && ((c) & 0xff) <= 0xdf)
73
74 #define K_IN()                          \
75         if (state != JIS_KANJI) {       \
76                 *out++ = ESC;           \
77                 *out++ = '$';           \
78                 *out++ = 'B';           \
79                 state = JIS_KANJI;      \
80         }
81
82 #define K_OUT()                         \
83         if (state != JIS_ASCII) {       \
84                 *out++ = ESC;           \
85                 *out++ = '(';           \
86                 *out++ = 'B';           \
87                 state = JIS_ASCII;      \
88         }
89
90 #define HW_IN()                         \
91         if (state != JIS_HWKANA) {      \
92                 *out++ = ESC;           \
93                 *out++ = '(';           \
94                 *out++ = 'I';           \
95                 state = JIS_HWKANA;     \
96         }
97
98 #define AUX_IN()                        \
99         if (state != JIS_AUXKANJI) {    \
100                 *out++ = ESC;           \
101                 *out++ = '$';           \
102                 *out++ = '(';           \
103                 *out++ = 'D';           \
104                 state = JIS_AUXKANJI;   \
105         }
106
107 static void conv_jistoeuc(gchar *outbuf, gint outlen, const gchar *inbuf);
108 static void conv_euctojis(gchar *outbuf, gint outlen, const gchar *inbuf);
109 static void conv_sjistoeuc(gchar *outbuf, gint outlen, const gchar *inbuf);
110
111 static void conv_jistoutf8(gchar *outbuf, gint outlen, const gchar *inbuf);
112 static void conv_sjistoutf8(gchar *outbuf, gint outlen, const gchar *inbuf);
113 static void conv_euctoutf8(gchar *outbuf, gint outlen, const gchar *inbuf);
114 static void conv_anytoutf8(gchar *outbuf, gint outlen, const gchar *inbuf);
115
116 static void conv_utf8toeuc(gchar *outbuf, gint outlen, const gchar *inbuf);
117 static void conv_utf8tojis(gchar *outbuf, gint outlen, const gchar *inbuf);
118
119 static void conv_unreadable_8bit(gchar *str);
120
121 static void conv_jistodisp(gchar *outbuf, gint outlen, const gchar *inbuf);
122 static void conv_sjistodisp(gchar *outbuf, gint outlen, const gchar *inbuf);
123 static void conv_euctodisp(gchar *outbuf, gint outlen, const gchar *inbuf);
124
125 static void conv_anytodisp(gchar *outbuf, gint outlen, const gchar *inbuf);
126 static void conv_ustodisp(gchar *outbuf, gint outlen, const gchar *inbuf);
127 static void conv_noconv(gchar *outbuf, gint outlen, const gchar *inbuf);
128
129 static gboolean strict_mode = FALSE;
130
131 void codeconv_set_strict(gboolean mode)
132 {
133         strict_mode = mode;
134 }
135
136 static void conv_jistoeuc(gchar *outbuf, gint outlen, const gchar *inbuf)
137 {
138         const guchar *in = inbuf;
139         guchar *out = outbuf;
140         JISState state = JIS_ASCII;
141
142         while (*in != '\0') {
143                 if (*in == ESC) {
144                         in++;
145                         if (*in == '$') {
146                                 if (*(in + 1) == '@' || *(in + 1) == 'B') {
147                                         state = JIS_KANJI;
148                                         in += 2;
149                                 } else if (*(in + 1) == '(' &&
150                                            *(in + 2) == 'D') {
151                                         state = JIS_AUXKANJI;
152                                         in += 3;
153                                 } else {
154                                         /* unknown escape sequence */
155                                         state = JIS_ASCII;
156                                 }
157                         } else if (*in == '(') {
158                                 if (*(in + 1) == 'B' || *(in + 1) == 'J') {
159                                         state = JIS_ASCII;
160                                         in += 2;
161                                 } else if (*(in + 1) == 'I') {
162                                         state = JIS_HWKANA;
163                                         in += 2;
164                                 } else {
165                                         /* unknown escape sequence */
166                                         state = JIS_ASCII;
167                                 }
168                         } else {
169                                 /* unknown escape sequence */
170                                 state = JIS_ASCII;
171                         }
172                 } else if (*in == 0x0e) {
173                         state = JIS_HWKANA;
174                         in++;
175                 } else if (*in == 0x0f) {
176                         state = JIS_ASCII;
177                         in++;
178                 } else {
179                         switch (state) {
180                         case JIS_ASCII:
181                                 *out++ = *in++;
182                                 break;
183                         case JIS_KANJI:
184                                 *out++ = *in++ | 0x80;
185                                 if (*in == '\0') break;
186                                 *out++ = *in++ | 0x80;
187                                 break;
188                         case JIS_HWKANA:
189                                 *out++ = 0x8e;
190                                 *out++ = *in++ | 0x80;
191                                 break;
192                         case JIS_AUXKANJI:
193                                 *out++ = 0x8f;
194                                 *out++ = *in++ | 0x80;
195                                 if (*in == '\0') break;
196                                 *out++ = *in++ | 0x80;
197                                 break;
198                         }
199                 }
200         }
201
202         *out = '\0';
203 }
204
205 #define JIS_HWDAKUTEN           0x5e
206 #define JIS_HWHANDAKUTEN        0x5f
207
208 static gint conv_jis_hantozen(guchar *outbuf, guchar jis_code, guchar sound_sym)
209 {
210         static guint16 h2z_tbl[] = {
211                 /* 0x20 - 0x2f */
212                 0x0000, 0x2123, 0x2156, 0x2157, 0x2122, 0x2126, 0x2572, 0x2521,
213                 0x2523, 0x2525, 0x2527, 0x2529, 0x2563, 0x2565, 0x2567, 0x2543,
214                 /* 0x30 - 0x3f */
215                 0x213c, 0x2522, 0x2524, 0x2526, 0x2528, 0x252a, 0x252b, 0x252d,
216                 0x252f, 0x2531, 0x2533, 0x2535, 0x2537, 0x2539, 0x253b, 0x253d,
217                 /* 0x40 - 0x4f */
218                 0x253f, 0x2541, 0x2544, 0x2546, 0x2548, 0x254a, 0x254b, 0x254c,
219                 0x254d, 0x254e, 0x254f, 0x2552, 0x2555, 0x2558, 0x255b, 0x255e,
220                 /* 0x50 - 0x5f */
221                 0x255f, 0x2560, 0x2561, 0x2562, 0x2564, 0x2566, 0x2568, 0x2569,
222                 0x256a, 0x256b, 0x256c, 0x256d, 0x256f, 0x2573, 0x212b, 0x212c
223         };
224
225         static guint16 dakuten_tbl[] = {
226                 /* 0x30 - 0x3f */
227                 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x252c, 0x252e,
228                 0x2530, 0x2532, 0x2534, 0x2536, 0x2538, 0x253a, 0x253c, 0x253e,
229                 /* 0x40 - 0x4f */
230                 0x2540, 0x2542, 0x2545, 0x2547, 0x2549, 0x0000, 0x0000, 0x0000,
231                 0x0000, 0x0000, 0x2550, 0x2553, 0x2556, 0x2559, 0x255c, 0x0000
232         };
233
234         static guint16 handakuten_tbl[] = {
235                 /* 0x4a - 0x4e */
236                 0x2551, 0x2554, 0x2557, 0x255a, 0x255d
237         };
238
239         guint16 out_code;
240
241         jis_code &= 0x7f;
242         sound_sym &= 0x7f;
243
244         if (jis_code < 0x21 || jis_code > 0x5f)
245                 return 0;
246
247         if (sound_sym == JIS_HWDAKUTEN &&
248             jis_code >= 0x36 && jis_code <= 0x4e) {
249                 out_code = dakuten_tbl[jis_code - 0x30];
250                 if (out_code != 0) {
251                         *outbuf = out_code >> 8;
252                         *(outbuf + 1) = out_code & 0xff;
253                         return 2;
254                 }
255         }
256
257         if (sound_sym == JIS_HWHANDAKUTEN &&
258             jis_code >= 0x4a && jis_code <= 0x4e) {
259                 out_code = handakuten_tbl[jis_code - 0x4a];
260                 *outbuf = out_code >> 8;
261                 *(outbuf + 1) = out_code & 0xff;
262                 return 2;
263         }
264
265         out_code = h2z_tbl[jis_code - 0x20];
266         *outbuf = out_code >> 8;
267         *(outbuf + 1) = out_code & 0xff;
268         return 1;
269 }
270
271 static void conv_euctojis(gchar *outbuf, gint outlen, const gchar *inbuf)
272 {
273         const guchar *in = inbuf;
274         guchar *out = outbuf;
275         JISState state = JIS_ASCII;
276
277         while (*in != '\0') {
278                 if (IS_ASCII(*in)) {
279                         K_OUT();
280                         *out++ = *in++;
281                 } else if (iseuckanji(*in)) {
282                         if (iseuckanji(*(in + 1))) {
283                                 K_IN();
284                                 *out++ = *in++ & 0x7f;
285                                 *out++ = *in++ & 0x7f;
286                         } else {
287                                 K_OUT();
288                                 *out++ = SUBST_CHAR;
289                                 in++;
290                                 if (*in != '\0' && !IS_ASCII(*in)) {
291                                         *out++ = SUBST_CHAR;
292                                         in++;
293                                 }
294                         }
295                 } else if (iseuchwkana1(*in)) {
296                         if (iseuchwkana2(*(in + 1))) {
297                                 if (prefs_common.allow_jisx0201_kana) {
298                                         HW_IN();
299                                         in++;
300                                         *out++ = *in++ & 0x7f;
301                                 } else {
302                                         guchar jis_ch[2];
303                                         gint len;
304
305                                         if (iseuchwkana1(*(in + 2)) &&
306                                             iseuchwkana2(*(in + 3)))
307                                                 len = conv_jis_hantozen
308                                                         (jis_ch,
309                                                          *(in + 1), *(in + 3));
310                                         else
311                                                 len = conv_jis_hantozen
312                                                         (jis_ch,
313                                                          *(in + 1), '\0');
314                                         if (len == 0)
315                                                 in += 2;
316                                         else {
317                                                 K_IN();
318                                                 in += len * 2;
319                                                 *out++ = jis_ch[0];
320                                                 *out++ = jis_ch[1];
321                                         }
322                                 }
323                         } else {
324                                 K_OUT();
325                                 in++;
326                                 if (*in != '\0' && !IS_ASCII(*in)) {
327                                         *out++ = SUBST_CHAR;
328                                         in++;
329                                 }
330                         }
331                 } else if (iseucaux(*in)) {
332                         in++;
333                         if (iseuckanji(*in) && iseuckanji(*(in + 1))) {
334                                 AUX_IN();
335                                 *out++ = *in++ & 0x7f;
336                                 *out++ = *in++ & 0x7f;
337                         } else {
338                                 K_OUT();
339                                 if (*in != '\0' && !IS_ASCII(*in)) {
340                                         *out++ = SUBST_CHAR;
341                                         in++;
342                                         if (*in != '\0' && !IS_ASCII(*in)) {
343                                                 *out++ = SUBST_CHAR;
344                                                 in++;
345                                         }
346                                 }
347                         }
348                 } else {
349                         K_OUT();
350                         *out++ = SUBST_CHAR;
351                         in++;
352                 }
353         }
354
355         K_OUT();
356         *out = '\0';
357 }
358
359 static void conv_sjistoeuc(gchar *outbuf, gint outlen, const gchar *inbuf)
360 {
361         const guchar *in = inbuf;
362         guchar *out = outbuf;
363
364         while (*in != '\0') {
365                 if (IS_ASCII(*in)) {
366                         *out++ = *in++;
367                 } else if (issjiskanji1(*in)) {
368                         if (issjiskanji2(*(in + 1))) {
369                                 guchar out1 = *in;
370                                 guchar out2 = *(in + 1);
371                                 guchar row;
372
373                                 row = out1 < 0xa0 ? 0x70 : 0xb0;
374                                 if (out2 < 0x9f) {
375                                         out1 = (out1 - row) * 2 - 1;
376                                         out2 -= out2 > 0x7f ? 0x20 : 0x1f;
377                                 } else {
378                                         out1 = (out1 - row) * 2;
379                                         out2 -= 0x7e;
380                                 }
381
382                                 *out++ = out1 | 0x80;
383                                 *out++ = out2 | 0x80;
384                                 in += 2;
385                         } else {
386                                 *out++ = SUBST_CHAR;
387                                 in++;
388                                 if (*in != '\0' && !IS_ASCII(*in)) {
389                                         *out++ = SUBST_CHAR;
390                                         in++;
391                                 }
392                         }
393                 } else if (issjishwkana(*in)) {
394                         *out++ = 0x8e;
395                         *out++ = *in++;
396                 } else {
397                         *out++ = SUBST_CHAR;
398                         in++;
399                 }
400         }
401
402         *out = '\0';
403 }
404
405 static void conv_jistoutf8(gchar *outbuf, gint outlen, const gchar *inbuf)
406 {
407         gchar *eucstr;
408
409         Xalloca(eucstr, outlen, return);
410
411         conv_jistoeuc(eucstr, outlen, inbuf);
412         conv_euctoutf8(outbuf, outlen, eucstr);
413 }
414
415 static void conv_sjistoutf8(gchar *outbuf, gint outlen, const gchar *inbuf)
416 {
417         gchar *tmpstr;
418
419         tmpstr = conv_iconv_strdup(inbuf, CS_SHIFT_JIS, CS_UTF_8);
420         if (tmpstr) {
421                 strncpy2(outbuf, tmpstr, outlen);
422                 g_free(tmpstr);
423         } else
424                 strncpy2(outbuf, inbuf, outlen);
425 }
426
427 static void conv_euctoutf8(gchar *outbuf, gint outlen, const gchar *inbuf)
428 {
429         static iconv_t cd = (iconv_t)-1;
430         static gboolean iconv_ok = TRUE;
431         gchar *tmpstr;
432
433         if (cd == (iconv_t)-1) {
434                 if (!iconv_ok) {
435                         strncpy2(outbuf, inbuf, outlen);
436                         return;
437                 }
438                 cd = iconv_open(CS_UTF_8, CS_EUC_JP_MS);
439                 if (cd == (iconv_t)-1) {
440                         cd = iconv_open(CS_UTF_8, CS_EUC_JP);
441                         if (cd == (iconv_t)-1) {
442                                 g_warning("conv_euctoutf8(): %s\n",
443                                           g_strerror(errno));
444                                 iconv_ok = FALSE;
445                                 strncpy2(outbuf, inbuf, outlen);
446                                 return;
447                         }
448                 }
449         }
450
451         tmpstr = conv_iconv_strdup_with_cd(inbuf, cd);
452         if (tmpstr) {
453                 strncpy2(outbuf, tmpstr, outlen);
454                 g_free(tmpstr);
455         } else
456                 strncpy2(outbuf, inbuf, outlen);
457 }
458
459 static void conv_anytoutf8(gchar *outbuf, gint outlen, const gchar *inbuf)
460 {
461         switch (conv_guess_ja_encoding(inbuf)) {
462         case C_ISO_2022_JP:
463                 conv_jistoutf8(outbuf, outlen, inbuf);
464                 break;
465         case C_SHIFT_JIS:
466                 conv_sjistoutf8(outbuf, outlen, inbuf);
467                 break;
468         case C_EUC_JP:
469                 conv_euctoutf8(outbuf, outlen, inbuf);
470                 break;
471         default:
472                 strncpy2(outbuf, inbuf, outlen);
473                 break;
474         }
475 }
476
477 static void conv_utf8toeuc(gchar *outbuf, gint outlen, const gchar *inbuf)
478 {
479         static iconv_t cd = (iconv_t)-1;
480         static gboolean iconv_ok = TRUE;
481         gchar *tmpstr;
482
483         if (cd == (iconv_t)-1) {
484                 if (!iconv_ok) {
485                         strncpy2(outbuf, inbuf, outlen);
486                         return;
487                 }
488                 cd = iconv_open(CS_EUC_JP_MS, CS_UTF_8);
489                 if (cd == (iconv_t)-1) {
490                         cd = iconv_open(CS_EUC_JP, CS_UTF_8);
491                         if (cd == (iconv_t)-1) {
492                                 g_warning("conv_utf8toeuc(): %s\n",
493                                           g_strerror(errno));
494                                 iconv_ok = FALSE;
495                                 strncpy2(outbuf, inbuf, outlen);
496                                 return;
497                         }
498                 }
499         }
500
501         tmpstr = conv_iconv_strdup_with_cd(inbuf, cd);
502         if (tmpstr) {
503                 strncpy2(outbuf, tmpstr, outlen);
504                 g_free(tmpstr);
505         } else
506                 strncpy2(outbuf, inbuf, outlen);
507 }
508
509 static void conv_utf8tojis(gchar *outbuf, gint outlen, const gchar *inbuf)
510 {
511         gchar *eucstr;
512
513         Xalloca(eucstr, outlen, return);
514
515         conv_utf8toeuc(eucstr, outlen, inbuf);
516         conv_euctojis(outbuf, outlen, eucstr);
517 }
518
519 static void conv_unreadable_8bit(gchar *str)
520 {
521         register guchar *p = str;
522
523         while (*p != '\0') {
524                 /* convert CR+LF -> LF */
525                 if (*p == '\r' && *(p + 1) == '\n')
526                         memmove(p, p + 1, strlen(p));
527                 else if (!IS_ASCII(*p)) *p = SUBST_CHAR;
528                 p++;
529         }
530 }
531
532 #define NCV     '\0'
533
534 void conv_mb_alnum(gchar *str)
535 {
536         static guchar char_tbl[] = {
537                 /* 0xa0 - 0xaf */
538                 NCV, ' ', NCV, NCV, ',', '.', NCV, ':',
539                 ';', '?', '!', NCV, NCV, NCV, NCV, NCV,
540                 /* 0xb0 - 0xbf */
541                 NCV, NCV, NCV, NCV, NCV, NCV, NCV, NCV,
542                 NCV, NCV, NCV, NCV, NCV, NCV, NCV, NCV,
543                 /* 0xc0 - 0xcf */
544                 NCV, NCV, NCV, NCV, NCV, NCV, NCV, NCV,
545                 NCV, NCV, '(', ')', NCV, NCV, '[', ']',
546                 /* 0xd0 - 0xdf */
547                 '{', '}', NCV, NCV, NCV, NCV, NCV, NCV,
548                 NCV, NCV, NCV, NCV, '+', '-', NCV, NCV,
549                 /* 0xe0 - 0xef */
550                 NCV, '=', NCV, '<', '>', NCV, NCV, NCV,
551                 NCV, NCV, NCV, NCV, NCV, NCV, NCV, NCV
552         };
553
554         register guchar *p = str;
555         register gint len;
556
557         len = strlen(str);
558
559         while (len > 1) {
560                 if (*p == 0xa3) {
561                         register guchar ch = *(p + 1);
562
563                         if (ch >= 0xb0 && ch <= 0xfa) {
564                                 /* [a-zA-Z] */
565                                 *p = ch & 0x7f;
566                                 p++;
567                                 len--;
568                                 memmove(p, p + 1, len);
569                                 len--;
570                         } else  {
571                                 p += 2;
572                                 len -= 2;
573                         }
574                 } else if (*p == 0xa1) {
575                         register guchar ch = *(p + 1);
576
577                         if (ch >= 0xa0 && ch <= 0xef &&
578                             NCV != char_tbl[ch - 0xa0]) {
579                                 *p = char_tbl[ch - 0xa0];
580                                 p++;
581                                 len--;
582                                 memmove(p, p + 1, len);
583                                 len--;
584                         } else {
585                                 p += 2;
586                                 len -= 2;
587                         }
588                 } else if (iseuckanji(*p)) {
589                         p += 2;
590                         len -= 2;
591                 } else {
592                         p++;
593                         len--;
594                 }
595         }
596 }
597
598 CharSet conv_guess_ja_encoding(const gchar *str)
599 {
600         const guchar *p = str;
601         CharSet guessed = C_US_ASCII;
602
603         while (*p != '\0') {
604                 if (*p == ESC && (*(p + 1) == '$' || *(p + 1) == '(')) {
605                         if (guessed == C_US_ASCII)
606                                 return C_ISO_2022_JP;
607                         p += 2;
608                 } else if (IS_ASCII(*p)) {
609                         p++;
610                 } else if (iseuckanji(*p) && iseuckanji(*(p + 1))) {
611                         if (*p >= 0xfd && *p <= 0xfe)
612                                 return C_EUC_JP;
613                         else if (guessed == C_SHIFT_JIS) {
614                                 if ((issjiskanji1(*p) &&
615                                      issjiskanji2(*(p + 1))) ||
616                                     issjishwkana(*p))
617                                         guessed = C_SHIFT_JIS;
618                                 else
619                                         guessed = C_EUC_JP;
620                         } else
621                                 guessed = C_EUC_JP;
622                         p += 2;
623                 } else if (issjiskanji1(*p) && issjiskanji2(*(p + 1))) {
624                         if (iseuchwkana1(*p) && iseuchwkana2(*(p + 1)))
625                                 guessed = C_SHIFT_JIS;
626                         else
627                                 return C_SHIFT_JIS;
628                         p += 2;
629                 } else if (issjishwkana(*p)) {
630                         guessed = C_SHIFT_JIS;
631                         p++;
632                 } else {
633                         p++;
634                 }
635         }
636
637         return guessed;
638 }
639
640 static void conv_jistodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
641 {
642         conv_jistoutf8(outbuf, outlen, inbuf);
643 }
644
645 static void conv_sjistodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
646 {
647         conv_sjistoutf8(outbuf, outlen, inbuf);
648 }
649
650 static void conv_euctodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
651 {
652         conv_euctoutf8(outbuf, outlen, inbuf);
653 }
654
655 void conv_utf8todisp(gchar *outbuf, gint outlen, const gchar *inbuf)
656 {
657         if (g_utf8_validate(inbuf, -1, NULL) == TRUE)
658                 strncpy2(outbuf, inbuf, outlen);
659         else
660                 conv_ustodisp(outbuf, outlen, inbuf);
661 }
662
663 static void conv_anytodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
664 {
665         conv_anytoutf8(outbuf, outlen, inbuf);
666         if (g_utf8_validate(outbuf, -1, NULL) != TRUE)
667                 conv_unreadable_8bit(outbuf);
668 }
669
670 static void conv_ustodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
671 {
672         strncpy2(outbuf, inbuf, outlen);
673         conv_unreadable_8bit(outbuf);
674 }
675
676 void conv_localetodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
677 {
678         gchar *tmpstr;
679
680         codeconv_set_strict(TRUE);
681         tmpstr = conv_iconv_strdup(inbuf, conv_get_locale_charset_str(),
682                                    CS_INTERNAL);
683         codeconv_set_strict(FALSE);
684         if (tmpstr && g_utf8_validate(tmpstr, -1, NULL)) {
685                 strncpy2(outbuf, tmpstr, outlen);
686                 g_free(tmpstr);
687                 return;
688         } else if (tmpstr && !g_utf8_validate(tmpstr, -1, NULL)) {
689                 g_free(tmpstr);
690                 codeconv_set_strict(TRUE);
691                 tmpstr = conv_iconv_strdup(inbuf, 
692                                 conv_get_locale_charset_str_no_utf8(),
693                                 CS_INTERNAL);
694                 codeconv_set_strict(FALSE);
695         }
696         if (tmpstr && g_utf8_validate(tmpstr, -1, NULL)) {
697                 strncpy2(outbuf, tmpstr, outlen);
698                 g_free(tmpstr);
699                 return;
700         } else {
701                 if (tmpstr)
702                         g_free(tmpstr);
703                 conv_utf8todisp(outbuf, outlen, inbuf);
704         }
705 }
706
707 static void conv_noconv(gchar *outbuf, gint outlen, const gchar *inbuf)
708 {
709         strncpy2(outbuf, inbuf, outlen);
710 }
711
712 static const gchar *
713 conv_get_fallback_for_private_encoding(const gchar *encoding)
714 {
715         if (encoding && (encoding[0] == 'X' || encoding[0] == 'x') &&
716             encoding[1] == '-') {
717                 if (!g_ascii_strcasecmp(encoding, CS_X_GBK))
718                         return CS_GBK;
719         }
720
721         return encoding;
722 }
723
724 CodeConverter *conv_code_converter_new(const gchar *src_charset)
725 {
726         CodeConverter *conv;
727
728         src_charset = conv_get_fallback_for_private_encoding(src_charset);
729
730         conv = g_new0(CodeConverter, 1);
731         conv->code_conv_func = conv_get_code_conv_func(src_charset, NULL);
732         conv->charset_str = g_strdup(src_charset);
733         conv->charset = conv_get_charset_from_str(src_charset);
734
735         return conv;
736 }
737
738 void conv_code_converter_destroy(CodeConverter *conv)
739 {
740         g_free(conv->charset_str);
741         g_free(conv);
742 }
743
744 gint conv_convert(CodeConverter *conv, gchar *outbuf, gint outlen,
745                   const gchar *inbuf)
746 {
747         if (conv->code_conv_func != conv_noconv)
748                 conv->code_conv_func(outbuf, outlen, inbuf);
749         else {
750                 gchar *str;
751
752                 str = conv_iconv_strdup(inbuf, conv->charset_str, NULL);
753                 if (!str)
754                         return -1;
755                 else {
756                         strncpy2(outbuf, str, outlen);
757                         g_free(str);
758                 }
759         }
760
761         return 0;
762 }
763
764 gchar *conv_codeset_strdup(const gchar *inbuf,
765                            const gchar *src_code, const gchar *dest_code)
766 {
767         gchar *buf;
768         size_t len;
769         CodeConvFunc conv_func;
770
771         src_code = conv_get_fallback_for_private_encoding(src_code);
772         conv_func = conv_get_code_conv_func(src_code, dest_code);
773         if (conv_func != conv_noconv) {
774                 len = (strlen(inbuf) + 1) * 3;
775                 buf = g_malloc(len);
776                 if (!buf) return NULL;
777
778                 conv_func(buf, len, inbuf);
779                 return g_realloc(buf, strlen(buf) + 1);
780         }
781
782         return conv_iconv_strdup(inbuf, src_code, dest_code);
783 }
784
785 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 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)
877                 src_code = conv_get_outgoing_charset_str();
878         if (!dest_code)
879                 dest_code = CS_INTERNAL;
880
881         /* don't convert if src and dest codeset are identical */
882         if (!strcasecmp(src_code, dest_code))
883                 return g_strdup(inbuf);
884
885         /* don't convert if current codeset is US-ASCII */
886         if (!strcasecmp(dest_code, CS_US_ASCII))
887                 return g_strdup(inbuf);
888
889         cd = iconv_open(dest_code, src_code);
890         if (cd == (iconv_t)-1)
891                 return NULL;
892
893         outbuf = conv_iconv_strdup_with_cd(inbuf, cd);
894
895         iconv_close(cd);
896
897         return outbuf;
898 }
899
900 gchar *conv_iconv_strdup_with_cd(const gchar *inbuf, iconv_t cd)
901 {
902         const gchar *inbuf_p;
903         gchar *outbuf;
904         gchar *outbuf_p;
905         size_t in_size;
906         size_t in_left;
907         size_t out_size;
908         size_t out_left;
909         size_t n_conv;
910         size_t len;
911
912         inbuf_p = inbuf;
913         in_size = strlen(inbuf);
914         in_left = in_size;
915         out_size = (in_size + 1) * 2;
916         outbuf = g_malloc(out_size);
917         outbuf_p = outbuf;
918         out_left = out_size;
919
920 #define EXPAND_BUF()                            \
921 {                                               \
922         len = outbuf_p - outbuf;                \
923         out_size *= 2;                          \
924         outbuf = g_realloc(outbuf, out_size);   \
925         outbuf_p = outbuf + len;                \
926         out_left = out_size - len;              \
927 }
928
929         while ((n_conv = iconv(cd, (ICONV_CONST gchar **)&inbuf_p, &in_left,
930                                &outbuf_p, &out_left)) == (size_t)-1) {
931                 if (EILSEQ == errno) {
932                         if (strict_mode) {
933                                 g_free(outbuf);
934                                 return NULL;
935                         }
936                         //g_print("iconv(): at %d: %s\n", in_size - in_left, g_strerror(errno));
937                         inbuf_p++;
938                         in_left--;
939                         if (out_left == 0) {
940                                 EXPAND_BUF();
941                         }
942                         *outbuf_p++ = SUBST_CHAR;
943                         out_left--;
944                 } else if (EINVAL == errno) {
945                         break;
946                 } else if (E2BIG == errno) {
947                         EXPAND_BUF();
948                 } else {
949                         g_warning("conv_iconv_strdup(): %s\n",
950                                   g_strerror(errno));
951                         break;
952                 }
953         }
954
955         while ((n_conv = iconv(cd, NULL, NULL, &outbuf_p, &out_left)) ==
956                (size_t)-1) {
957                 if (E2BIG == errno) {
958                         EXPAND_BUF();
959                 } else {
960                         g_warning("conv_iconv_strdup(): %s\n",
961                                   g_strerror(errno));
962                         break;
963                 }
964         }
965
966 #undef EXPAND_BUF
967
968         len = outbuf_p - outbuf;
969         outbuf = g_realloc(outbuf, len + 1);
970         outbuf[len] = '\0';
971
972         return outbuf;
973 }
974
975 static const struct {
976         CharSet charset;
977         gchar *const name;
978 } charsets[] = {
979         {C_US_ASCII,            CS_US_ASCII},
980         {C_US_ASCII,            CS_ANSI_X3_4_1968},
981         {C_UTF_8,               CS_UTF_8},
982         {C_UTF_7,               CS_UTF_7},
983         {C_ISO_8859_1,          CS_ISO_8859_1},
984         {C_ISO_8859_2,          CS_ISO_8859_2},
985         {C_ISO_8859_3,          CS_ISO_8859_3},
986         {C_ISO_8859_4,          CS_ISO_8859_4},
987         {C_ISO_8859_5,          CS_ISO_8859_5},
988         {C_ISO_8859_6,          CS_ISO_8859_6},
989         {C_ISO_8859_7,          CS_ISO_8859_7},
990         {C_ISO_8859_8,          CS_ISO_8859_8},
991         {C_ISO_8859_9,          CS_ISO_8859_9},
992         {C_ISO_8859_10,         CS_ISO_8859_10},
993         {C_ISO_8859_11,         CS_ISO_8859_11},
994         {C_ISO_8859_13,         CS_ISO_8859_13},
995         {C_ISO_8859_14,         CS_ISO_8859_14},
996         {C_ISO_8859_15,         CS_ISO_8859_15},
997         {C_BALTIC,              CS_BALTIC},
998         {C_CP1250,              CS_CP1250},
999         {C_CP1251,              CS_CP1251},
1000         {C_CP1252,              CS_CP1252},
1001         {C_CP1253,              CS_CP1253},
1002         {C_CP1254,              CS_CP1254},
1003         {C_CP1255,              CS_CP1255},
1004         {C_CP1256,              CS_CP1256},
1005         {C_CP1257,              CS_CP1257},
1006         {C_CP1258,              CS_CP1258},
1007         {C_WINDOWS_1250,        CS_WINDOWS_1250},
1008         {C_WINDOWS_1251,        CS_WINDOWS_1251},
1009         {C_WINDOWS_1252,        CS_WINDOWS_1252},
1010         {C_WINDOWS_1253,        CS_WINDOWS_1253},
1011         {C_WINDOWS_1254,        CS_WINDOWS_1254},
1012         {C_WINDOWS_1255,        CS_WINDOWS_1255},
1013         {C_WINDOWS_1256,        CS_WINDOWS_1256},
1014         {C_WINDOWS_1257,        CS_WINDOWS_1257},
1015         {C_WINDOWS_1258,        CS_WINDOWS_1258},
1016         {C_KOI8_R,              CS_KOI8_R},
1017         {C_KOI8_T,              CS_KOI8_T},
1018         {C_KOI8_U,              CS_KOI8_U},
1019         {C_ISO_2022_JP,         CS_ISO_2022_JP},
1020         {C_ISO_2022_JP_2,       CS_ISO_2022_JP_2},
1021         {C_ISO_2022_JP_3,       CS_ISO_2022_JP_3},
1022         {C_EUC_JP,              CS_EUC_JP},
1023         {C_EUC_JP,              CS_EUCJP},
1024         {C_EUC_JP_MS,           CS_EUC_JP_MS},
1025         {C_SHIFT_JIS,           CS_SHIFT_JIS},
1026         {C_SHIFT_JIS,           CS_SHIFT__JIS},
1027         {C_SHIFT_JIS,           CS_SJIS},
1028         {C_ISO_2022_KR,         CS_ISO_2022_KR},
1029         {C_EUC_KR,              CS_EUC_KR},
1030         {C_ISO_2022_CN,         CS_ISO_2022_CN},
1031         {C_EUC_CN,              CS_EUC_CN},
1032         {C_GB2312,              CS_GB2312},
1033         {C_GBK,                 CS_GBK},
1034         {C_EUC_TW,              CS_EUC_TW},
1035         {C_BIG5,                CS_BIG5},
1036         {C_BIG5_HKSCS,          CS_BIG5_HKSCS},
1037         {C_TIS_620,             CS_TIS_620},
1038         {C_WINDOWS_874,         CS_WINDOWS_874},
1039         {C_GEORGIAN_PS,         CS_GEORGIAN_PS},
1040         {C_TCVN5712_1,          CS_TCVN5712_1},
1041 };
1042
1043 static const struct {
1044         gchar *const locale;
1045         CharSet charset;
1046         CharSet out_charset;
1047 } locale_table[] = {
1048         {"ja_JP.eucJP"  , C_EUC_JP      , C_ISO_2022_JP},
1049         {"ja_JP.EUC-JP" , C_EUC_JP      , C_ISO_2022_JP},
1050         {"ja_JP.EUC"    , C_EUC_JP      , C_ISO_2022_JP},
1051         {"ja_JP.ujis"   , C_EUC_JP      , C_ISO_2022_JP},
1052         {"ja_JP.SJIS"   , C_SHIFT_JIS   , C_ISO_2022_JP},
1053         {"ja_JP.JIS"    , C_ISO_2022_JP , C_ISO_2022_JP},
1054 #ifdef G_OS_WIN32
1055         {"ja_JP"        , C_SHIFT_JIS   , C_ISO_2022_JP},
1056 #else
1057         {"ja_JP"        , C_EUC_JP      , C_ISO_2022_JP},
1058 #endif
1059         {"ko_KR.EUC-KR" , C_EUC_KR      , C_EUC_KR},
1060         {"ko_KR"        , C_EUC_KR      , C_EUC_KR},
1061         {"zh_CN.GB2312" , C_GB2312      , C_GB2312},
1062         {"zh_CN.GBK"    , C_GBK         , C_GBK},
1063         {"zh_CN"        , C_GB2312      , C_GB2312},
1064         {"zh_HK"        , C_BIG5_HKSCS  , C_BIG5_HKSCS},
1065         {"zh_TW.eucTW"  , C_EUC_TW      , C_BIG5},
1066         {"zh_TW.EUC-TW" , C_EUC_TW      , C_BIG5},
1067         {"zh_TW.Big5"   , C_BIG5        , C_BIG5},
1068         {"zh_TW"        , C_BIG5        , C_BIG5},
1069
1070         {"ru_RU.KOI8-R" , C_KOI8_R      , C_KOI8_R},
1071         {"ru_RU.KOI8R"  , C_KOI8_R      , C_KOI8_R},
1072         {"ru_RU.CP1251" , C_WINDOWS_1251, C_KOI8_R},
1073         {"ru_RU"        , C_ISO_8859_5  , C_KOI8_R},
1074         {"tg_TJ"        , C_KOI8_T      , C_KOI8_T},
1075         {"ru_UA"        , C_KOI8_U      , C_KOI8_U},
1076         {"uk_UA.CP1251" , C_WINDOWS_1251, C_KOI8_U},
1077         {"uk_UA"        , C_KOI8_U      , C_KOI8_U},
1078
1079         {"be_BY"        , C_WINDOWS_1251, C_WINDOWS_1251},
1080         {"bg_BG"        , C_WINDOWS_1251, C_WINDOWS_1251},
1081
1082         {"yi_US"        , C_WINDOWS_1255, C_WINDOWS_1255},
1083
1084         {"af_ZA"        , C_ISO_8859_1  , C_ISO_8859_1},
1085         {"br_FR"        , C_ISO_8859_1  , C_ISO_8859_1},
1086         {"ca_ES"        , C_ISO_8859_1  , C_ISO_8859_1},
1087         {"da_DK"        , C_ISO_8859_1  , C_ISO_8859_1},
1088         {"de_AT"        , C_ISO_8859_1  , C_ISO_8859_1},
1089         {"de_BE"        , C_ISO_8859_1  , C_ISO_8859_1},
1090         {"de_CH"        , C_ISO_8859_1  , C_ISO_8859_1},
1091         {"de_DE"        , C_ISO_8859_1  , C_ISO_8859_1},
1092         {"de_LU"        , C_ISO_8859_1  , C_ISO_8859_1},
1093         {"en_AU"        , C_ISO_8859_1  , C_ISO_8859_1},
1094         {"en_BW"        , C_ISO_8859_1  , C_ISO_8859_1},
1095         {"en_CA"        , C_ISO_8859_1  , C_ISO_8859_1},
1096         {"en_DK"        , C_ISO_8859_1  , C_ISO_8859_1},
1097         {"en_GB"        , C_ISO_8859_1  , C_ISO_8859_1},
1098         {"en_HK"        , C_ISO_8859_1  , C_ISO_8859_1},
1099         {"en_IE"        , C_ISO_8859_1  , C_ISO_8859_1},
1100         {"en_NZ"        , C_ISO_8859_1  , C_ISO_8859_1},
1101         {"en_PH"        , C_ISO_8859_1  , C_ISO_8859_1},
1102         {"en_SG"        , C_ISO_8859_1  , C_ISO_8859_1},
1103         {"en_US"        , C_ISO_8859_1  , C_ISO_8859_1},
1104         {"en_ZA"        , C_ISO_8859_1  , C_ISO_8859_1},
1105         {"en_ZW"        , C_ISO_8859_1  , C_ISO_8859_1},
1106         {"es_AR"        , C_ISO_8859_1  , C_ISO_8859_1},
1107         {"es_BO"        , C_ISO_8859_1  , C_ISO_8859_1},
1108         {"es_CL"        , C_ISO_8859_1  , C_ISO_8859_1},
1109         {"es_CO"        , C_ISO_8859_1  , C_ISO_8859_1},
1110         {"es_CR"        , C_ISO_8859_1  , C_ISO_8859_1},
1111         {"es_DO"        , C_ISO_8859_1  , C_ISO_8859_1},
1112         {"es_EC"        , C_ISO_8859_1  , C_ISO_8859_1},
1113         {"es_ES"        , C_ISO_8859_1  , C_ISO_8859_1},
1114         {"es_GT"        , C_ISO_8859_1  , C_ISO_8859_1},
1115         {"es_HN"        , C_ISO_8859_1  , C_ISO_8859_1},
1116         {"es_MX"        , C_ISO_8859_1  , C_ISO_8859_1},
1117         {"es_NI"        , C_ISO_8859_1  , C_ISO_8859_1},
1118         {"es_PA"        , C_ISO_8859_1  , C_ISO_8859_1},
1119         {"es_PE"        , C_ISO_8859_1  , C_ISO_8859_1},
1120         {"es_PR"        , C_ISO_8859_1  , C_ISO_8859_1},
1121         {"es_PY"        , C_ISO_8859_1  , C_ISO_8859_1},
1122         {"es_SV"        , C_ISO_8859_1  , C_ISO_8859_1},
1123         {"es_US"        , C_ISO_8859_1  , C_ISO_8859_1},
1124         {"es_UY"        , C_ISO_8859_1  , C_ISO_8859_1},
1125         {"es_VE"        , C_ISO_8859_1  , C_ISO_8859_1},
1126         {"et_EE"        , C_ISO_8859_1  , C_ISO_8859_1},
1127         {"eu_ES"        , C_ISO_8859_1  , C_ISO_8859_1},
1128         {"fi_FI"        , C_ISO_8859_1  , C_ISO_8859_1},
1129         {"fo_FO"        , C_ISO_8859_1  , C_ISO_8859_1},
1130         {"fr_BE"        , C_ISO_8859_1  , C_ISO_8859_1},
1131         {"fr_CA"        , C_ISO_8859_1  , C_ISO_8859_1},
1132         {"fr_CH"        , C_ISO_8859_1  , C_ISO_8859_1},
1133         {"fr_FR"        , C_ISO_8859_1  , C_ISO_8859_1},
1134         {"fr_LU"        , C_ISO_8859_1  , C_ISO_8859_1},
1135         {"ga_IE"        , C_ISO_8859_1  , C_ISO_8859_1},
1136         {"gl_ES"        , C_ISO_8859_1  , C_ISO_8859_1},
1137         {"gv_GB"        , C_ISO_8859_1  , C_ISO_8859_1},
1138         {"id_ID"        , C_ISO_8859_1  , C_ISO_8859_1},
1139         {"is_IS"        , C_ISO_8859_1  , C_ISO_8859_1},
1140         {"it_CH"        , C_ISO_8859_1  , C_ISO_8859_1},
1141         {"it_IT"        , C_ISO_8859_1  , C_ISO_8859_1},
1142         {"kl_GL"        , C_ISO_8859_1  , C_ISO_8859_1},
1143         {"kw_GB"        , C_ISO_8859_1  , C_ISO_8859_1},
1144         {"ms_MY"        , C_ISO_8859_1  , C_ISO_8859_1},
1145         {"nl_BE"        , C_ISO_8859_1  , C_ISO_8859_1},
1146         {"nl_NL"        , C_ISO_8859_1  , C_ISO_8859_1},
1147         {"nn_NO"        , C_ISO_8859_1  , C_ISO_8859_1},
1148         {"no_NO"        , C_ISO_8859_1  , C_ISO_8859_1},
1149         {"oc_FR"        , C_ISO_8859_1  , C_ISO_8859_1},
1150         {"pt_BR"        , C_ISO_8859_1  , C_ISO_8859_1},
1151         {"pt_PT"        , C_ISO_8859_1  , C_ISO_8859_1},
1152         {"sq_AL"        , C_ISO_8859_1  , C_ISO_8859_1},
1153         {"sv_FI"        , C_ISO_8859_1  , C_ISO_8859_1},
1154         {"sv_SE"        , C_ISO_8859_1  , C_ISO_8859_1},
1155         {"tl_PH"        , C_ISO_8859_1  , C_ISO_8859_1},
1156         {"uz_UZ"        , C_ISO_8859_1  , C_ISO_8859_1},
1157         {"wa_BE"        , C_ISO_8859_1  , C_ISO_8859_1},
1158
1159         {"bs_BA"        , C_ISO_8859_2  , C_ISO_8859_2},
1160         {"cs_CZ"        , C_ISO_8859_2  , C_ISO_8859_2},
1161         {"hr_HR"        , C_ISO_8859_2  , C_ISO_8859_2},
1162         {"hu_HU"        , C_ISO_8859_2  , C_ISO_8859_2},
1163         {"pl_PL"        , C_ISO_8859_2  , C_ISO_8859_2},
1164         {"ro_RO"        , C_ISO_8859_2  , C_ISO_8859_2},
1165         {"sk_SK"        , C_ISO_8859_2  , C_ISO_8859_2},
1166         {"sl_SI"        , C_ISO_8859_2  , C_ISO_8859_2},
1167
1168         {"sr_YU@cyrillic"       , C_ISO_8859_5  , C_ISO_8859_5},
1169         {"sr_YU"                , C_ISO_8859_2  , C_ISO_8859_2},
1170
1171         {"mt_MT"                , C_ISO_8859_3  , C_ISO_8859_3},
1172
1173         {"lt_LT.iso88594"       , C_ISO_8859_4  , C_ISO_8859_4},
1174         {"lt_LT.ISO8859-4"      , C_ISO_8859_4  , C_ISO_8859_4},
1175         {"lt_LT.ISO_8859-4"     , C_ISO_8859_4  , C_ISO_8859_4},
1176         {"lt_LT"                , C_ISO_8859_13 , C_ISO_8859_13},
1177
1178         {"mk_MK"        , C_ISO_8859_5  , C_ISO_8859_5},
1179
1180         {"ar_AE"        , C_ISO_8859_6  , C_ISO_8859_6},
1181         {"ar_BH"        , C_ISO_8859_6  , C_ISO_8859_6},
1182         {"ar_DZ"        , C_ISO_8859_6  , C_ISO_8859_6},
1183         {"ar_EG"        , C_ISO_8859_6  , C_ISO_8859_6},
1184         {"ar_IQ"        , C_ISO_8859_6  , C_ISO_8859_6},
1185         {"ar_JO"        , C_ISO_8859_6  , C_ISO_8859_6},
1186         {"ar_KW"        , C_ISO_8859_6  , C_ISO_8859_6},
1187         {"ar_LB"        , C_ISO_8859_6  , C_ISO_8859_6},
1188         {"ar_LY"        , C_ISO_8859_6  , C_ISO_8859_6},
1189         {"ar_MA"        , C_ISO_8859_6  , C_ISO_8859_6},
1190         {"ar_OM"        , C_ISO_8859_6  , C_ISO_8859_6},
1191         {"ar_QA"        , C_ISO_8859_6  , C_ISO_8859_6},
1192         {"ar_SA"        , C_ISO_8859_6  , C_ISO_8859_6},
1193         {"ar_SD"        , C_ISO_8859_6  , C_ISO_8859_6},
1194         {"ar_SY"        , C_ISO_8859_6  , C_ISO_8859_6},
1195         {"ar_TN"        , C_ISO_8859_6  , C_ISO_8859_6},
1196         {"ar_YE"        , C_ISO_8859_6  , C_ISO_8859_6},
1197
1198         {"el_GR"        , C_ISO_8859_7  , C_ISO_8859_7},
1199         {"he_IL"        , C_ISO_8859_8  , C_ISO_8859_8},
1200         {"iw_IL"        , C_ISO_8859_8  , C_ISO_8859_8},
1201         {"tr_TR"        , C_ISO_8859_9  , C_ISO_8859_9},
1202
1203         {"lv_LV"        , C_ISO_8859_13 , C_ISO_8859_13},
1204         {"mi_NZ"        , C_ISO_8859_13 , C_ISO_8859_13},
1205
1206         {"cy_GB"        , C_ISO_8859_14 , C_ISO_8859_14},
1207
1208         {"ar_IN"        , C_UTF_8       , C_UTF_8},
1209         {"en_IN"        , C_UTF_8       , C_UTF_8},
1210         {"se_NO"        , C_UTF_8       , C_UTF_8},
1211         {"ta_IN"        , C_UTF_8       , C_UTF_8},
1212         {"te_IN"        , C_UTF_8       , C_UTF_8},
1213         {"ur_PK"        , C_UTF_8       , C_UTF_8},
1214
1215         {"th_TH"        , C_TIS_620     , C_TIS_620},
1216         /* {"th_TH"     , C_WINDOWS_874}, */
1217         /* {"th_TH"     , C_ISO_8859_11}, */
1218
1219         {"ka_GE"        , C_GEORGIAN_PS , C_GEORGIAN_PS},
1220         {"vi_VN.TCVN"   , C_TCVN5712_1  , C_TCVN5712_1},
1221
1222         {"C"                    , C_US_ASCII    , C_US_ASCII},
1223         {"POSIX"                , C_US_ASCII    , C_US_ASCII},
1224         {"ANSI_X3.4-1968"       , C_US_ASCII    , C_US_ASCII},
1225 };
1226
1227 static GHashTable *conv_get_charset_to_str_table(void)
1228 {
1229         static GHashTable *table;
1230         gint i;
1231
1232         if (table)
1233                 return table;
1234
1235         table = g_hash_table_new(NULL, g_direct_equal);
1236
1237         for (i = 0; i < sizeof(charsets) / sizeof(charsets[0]); i++) {
1238                 if (g_hash_table_lookup(table, GUINT_TO_POINTER(charsets[i].charset))
1239                     == NULL) {
1240                         g_hash_table_insert
1241                                 (table, GUINT_TO_POINTER(charsets[i].charset),
1242                                  charsets[i].name);
1243                 }
1244         }
1245
1246         return table;
1247 }
1248
1249 static GHashTable *conv_get_charset_from_str_table(void)
1250 {
1251         static GHashTable *table;
1252         gint i;
1253
1254         if (table)
1255                 return table;
1256
1257         table = g_hash_table_new(str_case_hash, str_case_equal);
1258
1259         for (i = 0; i < sizeof(charsets) / sizeof(charsets[0]); i++) {
1260                 g_hash_table_insert(table, charsets[i].name,
1261                                     GUINT_TO_POINTER(charsets[i].charset));
1262         }
1263
1264         return table;
1265 }
1266
1267 const gchar *conv_get_charset_str(CharSet charset)
1268 {
1269         GHashTable *table;
1270
1271         table = conv_get_charset_to_str_table();
1272         return g_hash_table_lookup(table, GUINT_TO_POINTER(charset));
1273 }
1274
1275 CharSet conv_get_charset_from_str(const gchar *charset)
1276 {
1277         GHashTable *table;
1278
1279         if (!charset) return C_AUTO;
1280
1281         table = conv_get_charset_from_str_table();
1282         return GPOINTER_TO_UINT(g_hash_table_lookup(table, charset));
1283 }
1284
1285 CharSet conv_get_locale_charset(void)
1286 {
1287         static CharSet cur_charset = -1;
1288         const gchar *cur_locale;
1289         const gchar *p;
1290         gint i;
1291
1292         if (cur_charset != -1)
1293                 return cur_charset;
1294
1295         cur_locale = conv_get_current_locale();
1296         if (!cur_locale) {
1297                 cur_charset = C_US_ASCII;
1298                 return cur_charset;
1299         }
1300
1301         if (strcasestr(cur_locale, "UTF-8")) {
1302                 cur_charset = C_UTF_8;
1303                 return cur_charset;
1304         }
1305
1306         if ((p = strcasestr(cur_locale, "@euro")) && p[5] == '\0') {
1307                 cur_charset = C_ISO_8859_15;
1308                 return cur_charset;
1309         }
1310
1311         for (i = 0; i < sizeof(locale_table) / sizeof(locale_table[0]); i++) {
1312                 const gchar *p;
1313
1314                 /* "ja_JP.EUC" matches with "ja_JP.eucJP", "ja_JP.EUC" and
1315                    "ja_JP". "ja_JP" matches with "ja_JP.xxxx" and "ja" */
1316                 if (!g_ascii_strncasecmp(cur_locale, locale_table[i].locale,
1317                                  strlen(locale_table[i].locale))) {
1318                         cur_charset = locale_table[i].charset;
1319                         return cur_charset;
1320                 } else if ((p = strchr(locale_table[i].locale, '_')) &&
1321                          !strchr(p + 1, '.')) {
1322                         if (strlen(cur_locale) == 2 &&
1323                             !g_ascii_strncasecmp(cur_locale, locale_table[i].locale, 2)) {
1324                                 cur_charset = locale_table[i].charset;
1325                                 return cur_charset;
1326                         }
1327                 }
1328         }
1329
1330         cur_charset = C_AUTO;
1331         return cur_charset;
1332 }
1333
1334 static CharSet conv_get_locale_charset_no_utf8(void)
1335 {
1336         static CharSet cur_charset = -1;
1337         const gchar *cur_locale;
1338         const gchar *p;
1339         gchar *tmp;
1340         gint i;
1341
1342         if (prefs_common.broken_are_utf8)
1343                 return conv_get_locale_charset();
1344
1345         if (cur_charset != -1)
1346                 return cur_charset;
1347
1348         cur_locale = conv_get_current_locale();
1349         if (!cur_locale) {
1350                 cur_charset = C_US_ASCII;
1351                 return cur_charset;
1352         }
1353
1354         if (strcasestr(cur_locale, "UTF-8")) {
1355                 tmp = g_strdup(cur_locale);
1356                 *(strcasestr(tmp, ".UTF-8")) = '\0';
1357                 cur_locale = tmp;
1358         }
1359
1360         if ((p = strcasestr(cur_locale, "@euro")) && p[5] == '\0') {
1361                 cur_charset = C_ISO_8859_15;
1362                 return cur_charset;
1363         }
1364
1365         for (i = 0; i < sizeof(locale_table) / sizeof(locale_table[0]); i++) {
1366                 const gchar *p;
1367
1368                 /* "ja_JP.EUC" matches with "ja_JP.eucJP", "ja_JP.EUC" and
1369                    "ja_JP". "ja_JP" matches with "ja_JP.xxxx" and "ja" */
1370                 if (!g_ascii_strncasecmp(cur_locale, locale_table[i].locale,
1371                                  strlen(locale_table[i].locale))) {
1372                         cur_charset = locale_table[i].charset;
1373                         return cur_charset;
1374                 } else if ((p = strchr(locale_table[i].locale, '_')) &&
1375                          !strchr(p + 1, '.')) {
1376                         if (strlen(cur_locale) == 2 &&
1377                             !g_ascii_strncasecmp(cur_locale, locale_table[i].locale, 2)) {
1378                                 cur_charset = locale_table[i].charset;
1379                                 return cur_charset;
1380                         }
1381                 }
1382         }
1383
1384         cur_charset = C_AUTO;
1385         return cur_charset;
1386 }
1387
1388 const gchar *conv_get_locale_charset_str(void)
1389 {
1390         static const gchar *codeset = NULL;
1391
1392         if (!codeset)
1393                 codeset = conv_get_charset_str(conv_get_locale_charset());
1394
1395         return codeset ? codeset : CS_INTERNAL;
1396 }
1397
1398 const gchar *conv_get_locale_charset_str_no_utf8(void)
1399 {
1400         static const gchar *codeset = NULL;
1401
1402         if (!codeset)
1403                 codeset = conv_get_charset_str(conv_get_locale_charset_no_utf8());
1404
1405         return codeset ? codeset : CS_INTERNAL;
1406 }
1407
1408 CharSet conv_get_internal_charset(void)
1409 {
1410         return C_INTERNAL;
1411 }
1412
1413 const gchar *conv_get_internal_charset_str(void)
1414 {
1415         return CS_INTERNAL;
1416 }
1417
1418 CharSet conv_get_outgoing_charset(void)
1419 {
1420         static CharSet out_charset = -1;
1421         const gchar *cur_locale;
1422         const gchar *p;
1423         gint i;
1424
1425         if (out_charset != -1)
1426                 return out_charset;
1427
1428         cur_locale = conv_get_current_locale();
1429         if (!cur_locale) {
1430                 out_charset = C_AUTO;
1431                 return out_charset;
1432         }
1433
1434         if (strcasestr(cur_locale, "UTF-8")) {
1435                 out_charset = C_UTF_8;
1436                 return out_charset;
1437         }
1438
1439         if ((p = strcasestr(cur_locale, "@euro")) && p[5] == '\0') {
1440                 out_charset = C_ISO_8859_15;
1441                 return out_charset;
1442         }
1443
1444         for (i = 0; i < sizeof(locale_table) / sizeof(locale_table[0]); i++) {
1445                 const gchar *p;
1446
1447                 if (!g_ascii_strncasecmp(cur_locale, locale_table[i].locale,
1448                                  strlen(locale_table[i].locale))) {
1449                         out_charset = locale_table[i].out_charset;
1450                         break;
1451                 } else if ((p = strchr(locale_table[i].locale, '_')) &&
1452                          !strchr(p + 1, '.')) {
1453                         if (strlen(cur_locale) == 2 &&
1454                             !g_ascii_strncasecmp(cur_locale, locale_table[i].locale, 2)) {
1455                                 out_charset = locale_table[i].out_charset;
1456                                 break;
1457                         }
1458                 }
1459         }
1460
1461         return out_charset;
1462 }
1463
1464 const gchar *conv_get_outgoing_charset_str(void)
1465 {
1466         CharSet out_charset;
1467         const gchar *str;
1468
1469         out_charset = conv_get_outgoing_charset();
1470         str = conv_get_charset_str(out_charset);
1471
1472         return str ? str : CS_UTF_8;
1473 }
1474
1475 gboolean conv_is_multibyte_encoding(CharSet encoding)
1476 {
1477         switch (encoding) {
1478         case C_EUC_JP:
1479         case C_EUC_JP_MS:
1480         case C_EUC_KR:
1481         case C_EUC_TW:
1482         case C_EUC_CN:
1483         case C_ISO_2022_JP:
1484         case C_ISO_2022_JP_2:
1485         case C_ISO_2022_JP_3:
1486         case C_ISO_2022_KR:
1487         case C_ISO_2022_CN:
1488         case C_SHIFT_JIS:
1489         case C_GB2312:
1490         case C_GBK:
1491         case C_BIG5:
1492         case C_UTF_8:
1493         case C_UTF_7:
1494                 return TRUE;
1495         default:
1496                 return FALSE;
1497         }
1498 }
1499
1500 const gchar *conv_get_current_locale(void)
1501 {
1502         const gchar *cur_locale;
1503
1504 #ifdef G_OS_WIN32
1505         cur_locale = g_win32_getlocale();
1506 #else
1507         cur_locale = g_getenv("LC_ALL");
1508         if (!cur_locale) cur_locale = g_getenv("LC_CTYPE");
1509         if (!cur_locale) cur_locale = g_getenv("LANG");
1510         if (!cur_locale) cur_locale = setlocale(LC_CTYPE, NULL);
1511 #endif /* G_OS_WIN32 */
1512
1513         debug_print("current locale: %s\n",
1514                     cur_locale ? cur_locale : "(none)");
1515
1516         return cur_locale;
1517 }
1518
1519 gboolean conv_is_ja_locale(void)
1520 {
1521         static gint is_ja_locale = -1;
1522         const gchar *cur_locale;
1523
1524         if (is_ja_locale != -1)
1525                 return is_ja_locale != 0;
1526
1527         is_ja_locale = 0;
1528         cur_locale = conv_get_current_locale();
1529         if (cur_locale) {
1530                 if (g_ascii_strncasecmp(cur_locale, "ja", 2) == 0)
1531                         is_ja_locale = 1;
1532         }
1533
1534         return is_ja_locale != 0;
1535 }
1536
1537 gchar *conv_unmime_header(const gchar *str, const gchar *default_encoding)
1538 {
1539         gchar buf[BUFFSIZE];
1540
1541         if (is_ascii_str(str))
1542                 return unmime_header(str);
1543
1544         if (default_encoding) {
1545                 gchar *utf8_buf;
1546
1547                 utf8_buf = conv_codeset_strdup
1548                         (str, default_encoding, CS_INTERNAL);
1549                 if (utf8_buf) {
1550                         gchar *decoded_str;
1551
1552                         decoded_str = unmime_header(utf8_buf);
1553                         g_free(utf8_buf);
1554                         return decoded_str;
1555                 }
1556         }
1557
1558         if (conv_is_ja_locale())
1559                 conv_anytodisp(buf, sizeof(buf), str);
1560         else
1561                 conv_localetodisp(buf, sizeof(buf), str);
1562
1563         return unmime_header(buf);
1564 }
1565
1566 #define MAX_LINELEN             76
1567 #define MAX_HARD_LINELEN        996
1568 #define MIMESEP_BEGIN           "=?"
1569 #define MIMESEP_END             "?="
1570
1571 #define LBREAK_IF_REQUIRED(cond, is_plain_text)                         \
1572 {                                                                       \
1573         if (len - (destp - (guchar *)dest) < MAX_LINELEN + 2) {         \
1574                 *destp = '\0';                                          \
1575                 return;                                                 \
1576         }                                                               \
1577                                                                         \
1578         if ((cond) && *srcp) {                                          \
1579                 if (destp > (guchar *)dest && left < MAX_LINELEN - 1) { \
1580                         if (isspace(*(destp - 1)))                      \
1581                                 destp--;                                \
1582                         else if (is_plain_text && isspace(*srcp))       \
1583                                 srcp++;                                 \
1584                         if (*srcp) {                                    \
1585                                 *destp++ = '\n';                        \
1586                                 *destp++ = ' ';                         \
1587                                 left = MAX_LINELEN - 1;                 \
1588                         }                                               \
1589                 }                                                       \
1590         }                                                               \
1591 }
1592
1593 void conv_encode_header_full(gchar *dest, gint len, const gchar *src,
1594                         gint header_len, gboolean addr_field,
1595                         const gchar *out_encoding_)
1596 {
1597         const gchar *cur_encoding;
1598         const gchar *out_encoding;
1599         gint mimestr_len;
1600         gchar *mimesep_enc;
1601         gint left;
1602         const guchar *srcp = src;
1603         guchar *destp = dest;
1604         gboolean use_base64;
1605
1606         g_return_if_fail(g_utf8_validate(src, -1, NULL) == TRUE);
1607
1608         if (MB_CUR_MAX > 1) {
1609                 use_base64 = TRUE;
1610                 mimesep_enc = "?B?";
1611         } else {
1612                 use_base64 = FALSE;
1613                 mimesep_enc = "?Q?";
1614         }
1615
1616         cur_encoding = CS_INTERNAL;
1617
1618         if (out_encoding_)
1619                 out_encoding = out_encoding_;
1620         else
1621                 out_encoding = conv_get_outgoing_charset_str();
1622
1623         if (!strcmp(out_encoding, CS_US_ASCII))
1624                 out_encoding = CS_ISO_8859_1;
1625
1626         mimestr_len = strlen(MIMESEP_BEGIN) + strlen(out_encoding) +
1627                 strlen(mimesep_enc) + strlen(MIMESEP_END);
1628
1629         left = MAX_LINELEN - header_len;
1630
1631         while (*srcp) {
1632                 LBREAK_IF_REQUIRED(left <= 0, TRUE);
1633
1634                 while (isspace(*srcp)) {
1635                         *destp++ = *srcp++;
1636                         left--;
1637                         LBREAK_IF_REQUIRED(left <= 0, TRUE);
1638                 }
1639
1640                 /* output as it is if the next word is ASCII string */
1641                 if (!is_next_nonascii(srcp)) {
1642                         gint word_len;
1643
1644                         word_len = get_next_word_len(srcp);
1645                         LBREAK_IF_REQUIRED(left < word_len, TRUE);
1646                         while (word_len > 0) {
1647                                 LBREAK_IF_REQUIRED(left + (MAX_HARD_LINELEN - MAX_LINELEN) <= 0, TRUE)
1648                                 *destp++ = *srcp++;
1649                                 left--;
1650                                 word_len--;
1651                         }
1652
1653                         continue;
1654                 }
1655
1656                 /* don't include parentheses in encoded strings */
1657                 if (addr_field && (*srcp == '(' || *srcp == ')')) {
1658                         LBREAK_IF_REQUIRED(left < 2, FALSE);
1659                         *destp++ = *srcp++;
1660                         left--;
1661                 }
1662
1663                 while (1) {
1664                         gint mb_len = 0;
1665                         gint cur_len = 0;
1666                         gchar *part_str;
1667                         gchar *out_str;
1668                         gchar *enc_str;
1669                         const guchar *p = srcp;
1670                         gint out_str_len;
1671                         gint out_enc_str_len;
1672                         gint mime_block_len;
1673                         gboolean cont = FALSE;
1674
1675                         while (*p != '\0') {
1676                                 if (isspace(*p) && !is_next_nonascii(p + 1))
1677                                         break;
1678                                 /* don't include parentheses in encoded
1679                                    strings */
1680                                 if (addr_field && (*p == '(' || *p == ')'))
1681                                         break;
1682
1683                                 mb_len = g_utf8_skip[*p];
1684
1685                                 Xstrndup_a(part_str, srcp, cur_len + mb_len, );
1686                                 out_str = conv_codeset_strdup
1687                                         (part_str, cur_encoding, out_encoding);
1688                                 if (!out_str) {
1689                                         if (strict_mode) {
1690                                                 *dest = '\0';
1691                                                 return;
1692                                         } else {
1693                                                 g_warning("conv_encode_header(): code conversion failed\n");
1694                                                 conv_unreadable_8bit(part_str);
1695                                                 out_str = g_strdup(part_str);
1696                                         }
1697                                 }
1698                                 out_str_len = strlen(out_str);
1699
1700                                 if (use_base64)
1701                                         out_enc_str_len = B64LEN(out_str_len);
1702                                 else
1703                                         out_enc_str_len =
1704                                                 qp_get_q_encoding_len(out_str);
1705
1706                                 g_free(out_str);
1707
1708                                 if (mimestr_len + out_enc_str_len <= left) {
1709                                         cur_len += mb_len;
1710                                         p += mb_len;
1711                                 } else if (cur_len == 0) {
1712                                         LBREAK_IF_REQUIRED(1, FALSE);
1713                                         continue;
1714                                 } else {
1715                                         cont = TRUE;
1716                                         break;
1717                                 }
1718                         }
1719
1720                         if (cur_len > 0) {
1721                                 Xstrndup_a(part_str, srcp, cur_len, );
1722                                 out_str = conv_codeset_strdup
1723                                         (part_str, cur_encoding, out_encoding);
1724                                 if (!out_str) {
1725                                         g_warning("conv_encode_header(): code conversion failed\n");
1726                                         conv_unreadable_8bit(part_str);
1727                                         out_str = g_strdup(part_str);
1728                                 }
1729                                 out_str_len = strlen(out_str);
1730
1731                                 if (use_base64)
1732                                         out_enc_str_len = B64LEN(out_str_len);
1733                                 else
1734                                         out_enc_str_len =
1735                                                 qp_get_q_encoding_len(out_str);
1736
1737                                 Xalloca(enc_str, out_enc_str_len + 1, );
1738                                 if (use_base64)
1739                                         base64_encode(enc_str, out_str, out_str_len);
1740                                 else
1741                                         qp_q_encode(enc_str, out_str);
1742
1743                                 g_free(out_str);
1744
1745                                 /* output MIME-encoded string block */
1746                                 mime_block_len = mimestr_len + strlen(enc_str);
1747                                 g_snprintf(destp, mime_block_len + 1,
1748                                            MIMESEP_BEGIN "%s%s%s" MIMESEP_END,
1749                                            out_encoding, mimesep_enc, enc_str);
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 gchar *conv_filename_from_utf8(const gchar *utf8_file)
1774 {
1775         gchar *fs_file;
1776         GError *error = NULL;
1777
1778         fs_file = g_filename_from_utf8(utf8_file, -1, NULL, NULL, &error);
1779         if (error) {
1780                 g_warning("failed to convert encoding of file name: %s\n",
1781                           error->message);
1782                 g_error_free(error);
1783         }
1784         if (!fs_file)
1785                 fs_file = g_strdup(utf8_file);
1786
1787         return fs_file;
1788 }
1789
1790 gchar *conv_filename_to_utf8(const gchar *fs_file)
1791 {
1792         gchar *utf8_file = NULL;
1793         GError *error = NULL;
1794
1795         utf8_file = g_filename_to_utf8(fs_file, -1, NULL, NULL, &error);
1796         if (error) {
1797                 g_warning("failed to convert encoding of file name: %s\n",
1798                           error->message);
1799                 g_error_free(error);
1800         }
1801
1802         if (!utf8_file || !g_utf8_validate(utf8_file, -1, NULL)) {
1803                 g_free(utf8_file);
1804                 utf8_file = g_strdup(fs_file);
1805                 conv_unreadable_8bit(utf8_file);
1806         }
1807
1808         return utf8_file;
1809 }