Make sure we don't run out of the output buffer. Maybe fixes bug #3557
[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         gchar *out = outbuf;
159         JISState state = JIS_ASCII;
160
161         while (*in != '\0' && (out - outbuf) > outlen - 3) {
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         gchar *out = outbuf;
295         JISState state = JIS_ASCII;
296
297         while (*in != '\0' && (out - outbuf) < outlen - 3) {
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         gchar *out = outbuf;
384
385         while (*in != '\0' && (out - outbuf) < outlen - 3) {
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",
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",
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_MACCYR))
699                         return CS_MACCYR;
700                 if (!g_ascii_strcasecmp(encoding, CS_X_GBK))
701                         return CS_GBK;
702         }
703
704         return encoding;
705 }
706
707 CodeConverter *conv_code_converter_new(const gchar *src_charset)
708 {
709         CodeConverter *conv;
710
711         src_charset = conv_get_fallback_for_private_encoding(src_charset);
712
713         conv = g_new0(CodeConverter, 1);
714         conv->code_conv_func = conv_get_code_conv_func(src_charset, NULL);
715         conv->charset_str = g_strdup(src_charset);
716         conv->charset = conv_get_charset_from_str(src_charset);
717
718         return conv;
719 }
720
721 void conv_code_converter_destroy(CodeConverter *conv)
722 {
723         g_free(conv->charset_str);
724         g_free(conv);
725 }
726
727 gint conv_convert(CodeConverter *conv, gchar *outbuf, gint outlen,
728                   const gchar *inbuf)
729 {
730         if (conv->code_conv_func != conv_noconv)
731                 return conv->code_conv_func(outbuf, outlen, inbuf);
732         else {
733                 gchar *str;
734
735                 str = conv_iconv_strdup(inbuf, conv->charset_str, NULL);
736                 if (!str)
737                         return -1;
738                 else {
739                         strncpy2(outbuf, str, outlen);
740                         g_free(str);
741                 }
742         }
743
744         return 0;
745 }
746
747 gchar *conv_codeset_strdup(const gchar *inbuf,
748                            const gchar *src_code, const gchar *dest_code)
749 {
750         gchar *buf;
751         size_t len;
752         CodeConvFunc conv_func;
753
754         if (!strcmp2(src_code, dest_code)) {
755                 CharSet dest_charset = conv_get_charset_from_str(dest_code);
756                 if (strict_mode && dest_charset == C_UTF_8) {
757                         /* ensure valid UTF-8 if target is UTF-8 */
758                         if (!g_utf8_validate(inbuf, -1, NULL)) {
759                                 return NULL;
760                         }
761                 }
762                 /* otherwise, try for a lucky day */
763                 return g_strdup(inbuf);
764         }
765
766         src_code = conv_get_fallback_for_private_encoding(src_code);
767         conv_func = conv_get_code_conv_func(src_code, dest_code);
768         if (conv_func == conv_ustodisp && strict_mode && !is_ascii_str(inbuf))
769                 return NULL;
770
771         if (conv_func != conv_noconv) {
772                 len = (strlen(inbuf) + 1) * 3;
773                 buf = g_malloc(len);
774                 if (!buf) return NULL;
775
776                 if (conv_func(buf, len, inbuf) == 0 || !strict_mode)
777                         return g_realloc(buf, strlen(buf) + 1);
778                 else {
779                         g_free(buf);
780                         return NULL;
781                 }
782         }
783
784         return conv_iconv_strdup(inbuf, src_code, dest_code);
785 }
786
787 static CodeConvFunc conv_get_code_conv_func(const gchar *src_charset_str,
788                                      const gchar *dest_charset_str)
789 {
790         CodeConvFunc code_conv = conv_noconv;
791         CharSet src_charset;
792         CharSet dest_charset;
793
794         if (!src_charset_str)
795                 src_charset = conv_get_locale_charset();
796         else
797                 src_charset = conv_get_charset_from_str(src_charset_str);
798
799         /* auto detection mode */
800         if (!src_charset_str && !dest_charset_str) {
801                 if (conv_is_ja_locale())
802                         return conv_anytodisp;
803                 else
804                         return conv_noconv;
805         }
806
807         dest_charset = conv_get_charset_from_str(dest_charset_str);
808
809         if (dest_charset == C_US_ASCII)
810                 return conv_ustodisp;
811
812         switch (src_charset) {
813         case C_US_ASCII:
814         case C_ISO_8859_1:
815         case C_ISO_8859_2:
816         case C_ISO_8859_3:
817         case C_ISO_8859_4:
818         case C_ISO_8859_5:
819         case C_ISO_8859_6:
820         case C_ISO_8859_7:
821         case C_ISO_8859_8:
822         case C_ISO_8859_9:
823         case C_ISO_8859_10:
824         case C_ISO_8859_11:
825         case C_ISO_8859_13:
826         case C_ISO_8859_14:
827         case C_ISO_8859_15:
828                 break;
829         case C_ISO_2022_JP:
830         case C_ISO_2022_JP_2:
831         case C_ISO_2022_JP_3:
832                 if (dest_charset == C_AUTO)
833                         code_conv = conv_jistodisp;
834                 else if (dest_charset == C_EUC_JP)
835                         code_conv = conv_jistoeuc;
836                 else if (dest_charset == C_UTF_8)
837                         code_conv = conv_jistoutf8;
838                 break;
839         case C_SHIFT_JIS:
840                 if (dest_charset == C_AUTO)
841                         code_conv = conv_sjistodisp;
842                 else if (dest_charset == C_EUC_JP)
843                         code_conv = conv_sjistoeuc;
844                 else if (dest_charset == C_UTF_8)
845                         code_conv = conv_sjistoutf8;
846                 break;
847         case C_EUC_JP:
848                 if (dest_charset == C_AUTO)
849                         code_conv = conv_euctodisp;
850                 else if (dest_charset == C_ISO_2022_JP   ||
851                          dest_charset == C_ISO_2022_JP_2 ||
852                          dest_charset == C_ISO_2022_JP_3)
853                         code_conv = conv_euctojis;
854                 else if (dest_charset == C_UTF_8)
855                         code_conv = conv_euctoutf8;
856                 break;
857         case C_UTF_8:
858                 if (dest_charset == C_EUC_JP)
859                         code_conv = conv_utf8toeuc;
860                 else if (dest_charset == C_ISO_2022_JP   ||
861                          dest_charset == C_ISO_2022_JP_2 ||
862                          dest_charset == C_ISO_2022_JP_3)
863                         code_conv = conv_utf8tojis;
864                 break;
865         default:
866                 break;
867         }
868
869         return code_conv;
870 }
871
872 static gchar *conv_iconv_strdup(const gchar *inbuf,
873                          const gchar *src_code, const gchar *dest_code)
874 {
875         iconv_t cd;
876         gchar *outbuf;
877
878         if (!src_code && !dest_code && 
879             g_utf8_validate(inbuf, -1, NULL))
880                 return g_strdup(inbuf);
881
882         if (!src_code)
883                 src_code = conv_get_outgoing_charset_str();
884         if (!dest_code)
885                 dest_code = CS_INTERNAL;
886
887         /* don't convert if src and dest codeset are identical */
888         if (!strcasecmp(src_code, dest_code))
889                 return g_strdup(inbuf);
890
891         /* don't convert if dest codeset is US-ASCII */
892         if (!strcasecmp(src_code, CS_US_ASCII))
893                 return g_strdup(inbuf);
894
895         /* don't convert if dest codeset is US-ASCII */
896         if (!strcasecmp(dest_code, CS_US_ASCII))
897                 return g_strdup(inbuf);
898
899         cd = iconv_open(dest_code, src_code);
900         if (cd == (iconv_t)-1)
901                 return NULL;
902
903         outbuf = conv_iconv_strdup_with_cd(inbuf, cd);
904
905         iconv_close(cd);
906
907         return outbuf;
908 }
909
910 gchar *conv_iconv_strdup_with_cd(const gchar *inbuf, iconv_t cd)
911 {
912         const gchar *inbuf_p;
913         gchar *outbuf;
914         gchar *outbuf_p;
915         size_t in_size;
916         size_t in_left;
917         size_t out_size;
918         size_t out_left;
919         size_t n_conv;
920         size_t len;
921
922         inbuf_p = inbuf;
923         in_size = strlen(inbuf);
924         in_left = in_size;
925         out_size = (in_size + 1) * 2;
926         outbuf = g_malloc(out_size);
927         outbuf_p = outbuf;
928         out_left = out_size;
929
930 #define EXPAND_BUF()                            \
931 {                                               \
932         len = outbuf_p - outbuf;                \
933         out_size *= 2;                          \
934         outbuf = g_realloc(outbuf, out_size);   \
935         outbuf_p = outbuf + len;                \
936         out_left = out_size - len;              \
937 }
938
939         while ((n_conv = iconv(cd, (ICONV_CONST gchar **)&inbuf_p, &in_left,
940                                &outbuf_p, &out_left)) == (size_t)-1) {
941                 if (EILSEQ == errno) {
942                         if (strict_mode) {
943                                 g_free(outbuf);
944                                 return NULL;
945                         }
946                         //g_print("iconv(): at %d: %s\n", in_size - in_left, g_strerror(errno));
947                         inbuf_p++;
948                         in_left--;
949                         if (out_left == 0) {
950                                 EXPAND_BUF();
951                         }
952                         *outbuf_p++ = SUBST_CHAR;
953                         out_left--;
954                 } else if (EINVAL == errno) {
955                         break;
956                 } else if (E2BIG == errno) {
957                         EXPAND_BUF();
958                 } else {
959                         g_warning("conv_iconv_strdup(): %s",
960                                   g_strerror(errno));
961                         break;
962                 }
963         }
964
965         while ((n_conv = iconv(cd, NULL, NULL, &outbuf_p, &out_left)) ==
966                (size_t)-1) {
967                 if (E2BIG == errno) {
968                         EXPAND_BUF();
969                 } else {
970                         g_warning("conv_iconv_strdup(): %s",
971                                   g_strerror(errno));
972                         break;
973                 }
974         }
975
976 #undef EXPAND_BUF
977
978         len = outbuf_p - outbuf;
979         outbuf = g_realloc(outbuf, len + 1);
980         outbuf[len] = '\0';
981
982         return outbuf;
983 }
984
985 static const struct {
986         CharSet charset;
987         gchar *const name;
988 } charsets[] = {
989         {C_US_ASCII,            CS_US_ASCII},
990         {C_US_ASCII,            CS_ANSI_X3_4_1968},
991         {C_UTF_8,               CS_UTF_8},
992         {C_UTF_7,               CS_UTF_7},
993         {C_ISO_8859_1,          CS_ISO_8859_1},
994         {C_ISO_8859_2,          CS_ISO_8859_2},
995         {C_ISO_8859_3,          CS_ISO_8859_3},
996         {C_ISO_8859_4,          CS_ISO_8859_4},
997         {C_ISO_8859_5,          CS_ISO_8859_5},
998         {C_ISO_8859_6,          CS_ISO_8859_6},
999         {C_ISO_8859_7,          CS_ISO_8859_7},
1000         {C_ISO_8859_8,          CS_ISO_8859_8},
1001         {C_ISO_8859_9,          CS_ISO_8859_9},
1002         {C_ISO_8859_10,         CS_ISO_8859_10},
1003         {C_ISO_8859_11,         CS_ISO_8859_11},
1004         {C_ISO_8859_13,         CS_ISO_8859_13},
1005         {C_ISO_8859_14,         CS_ISO_8859_14},
1006         {C_ISO_8859_15,         CS_ISO_8859_15},
1007         {C_BALTIC,              CS_BALTIC},
1008         {C_CP1250,              CS_CP1250},
1009         {C_CP1251,              CS_CP1251},
1010         {C_CP1252,              CS_CP1252},
1011         {C_CP1253,              CS_CP1253},
1012         {C_CP1254,              CS_CP1254},
1013         {C_CP1255,              CS_CP1255},
1014         {C_CP1256,              CS_CP1256},
1015         {C_CP1257,              CS_CP1257},
1016         {C_CP1258,              CS_CP1258},
1017         {C_WINDOWS_1250,        CS_WINDOWS_1250},
1018         {C_WINDOWS_1251,        CS_WINDOWS_1251},
1019         {C_WINDOWS_1252,        CS_WINDOWS_1252},
1020         {C_WINDOWS_1253,        CS_WINDOWS_1253},
1021         {C_WINDOWS_1254,        CS_WINDOWS_1254},
1022         {C_WINDOWS_1255,        CS_WINDOWS_1255},
1023         {C_WINDOWS_1256,        CS_WINDOWS_1256},
1024         {C_WINDOWS_1257,        CS_WINDOWS_1257},
1025         {C_WINDOWS_1258,        CS_WINDOWS_1258},
1026         {C_KOI8_R,              CS_KOI8_R},
1027         {C_MACCYR,              CS_MACCYR},
1028         {C_KOI8_T,              CS_KOI8_T},
1029         {C_KOI8_U,              CS_KOI8_U},
1030         {C_ISO_2022_JP,         CS_ISO_2022_JP},
1031         {C_ISO_2022_JP_2,       CS_ISO_2022_JP_2},
1032         {C_ISO_2022_JP_3,       CS_ISO_2022_JP_3},
1033         {C_EUC_JP,              CS_EUC_JP},
1034         {C_EUC_JP,              CS_EUCJP},
1035         {C_EUC_JP_MS,           CS_EUC_JP_MS},
1036         {C_SHIFT_JIS,           CS_SHIFT_JIS},
1037         {C_SHIFT_JIS,           CS_SHIFT__JIS},
1038         {C_SHIFT_JIS,           CS_SJIS},
1039         {C_ISO_2022_KR,         CS_ISO_2022_KR},
1040         {C_EUC_KR,              CS_EUC_KR},
1041         {C_ISO_2022_CN,         CS_ISO_2022_CN},
1042         {C_EUC_CN,              CS_EUC_CN},
1043         {C_GB18030,             CS_GB18030},
1044         {C_GB2312,              CS_GB2312},
1045         {C_GBK,                 CS_GBK},
1046         {C_EUC_TW,              CS_EUC_TW},
1047         {C_BIG5,                CS_BIG5},
1048         {C_BIG5_HKSCS,          CS_BIG5_HKSCS},
1049         {C_TIS_620,             CS_TIS_620},
1050         {C_WINDOWS_874,         CS_WINDOWS_874},
1051         {C_GEORGIAN_PS,         CS_GEORGIAN_PS},
1052         {C_TCVN5712_1,          CS_TCVN5712_1},
1053 };
1054
1055 static const struct {
1056         gchar *const locale;
1057         CharSet charset;
1058         CharSet out_charset;
1059 } locale_table[] = {
1060         {"ja_JP.eucJP"  , C_EUC_JP      , C_ISO_2022_JP},
1061         {"ja_JP.EUC-JP" , C_EUC_JP      , C_ISO_2022_JP},
1062         {"ja_JP.EUC"    , C_EUC_JP      , C_ISO_2022_JP},
1063         {"ja_JP.ujis"   , C_EUC_JP      , C_ISO_2022_JP},
1064         {"ja_JP.SJIS"   , C_SHIFT_JIS   , C_ISO_2022_JP},
1065         {"ja_JP.JIS"    , C_ISO_2022_JP , C_ISO_2022_JP},
1066 #ifdef G_OS_WIN32
1067         {"ja_JP"        , C_SHIFT_JIS   , C_ISO_2022_JP},
1068 #else
1069         {"ja_JP"        , C_EUC_JP      , C_ISO_2022_JP},
1070 #endif
1071         {"ko_KR.EUC-KR" , C_EUC_KR      , C_EUC_KR},
1072         {"ko_KR"        , C_EUC_KR      , C_EUC_KR},
1073         {"zh_CN.GB18030"        , C_GB18030     , C_GB18030},
1074         {"zh_CN.GB2312" , C_GB2312      , C_GB2312},
1075         {"zh_CN.GBK"    , C_GBK         , C_GBK},
1076         {"zh_CN"        , C_GB18030     , C_GB18030},
1077         {"zh_HK"        , C_BIG5_HKSCS  , C_BIG5_HKSCS},
1078         {"zh_TW.eucTW"  , C_EUC_TW      , C_BIG5},
1079         {"zh_TW.EUC-TW" , C_EUC_TW      , C_BIG5},
1080         {"zh_TW.Big5"   , C_BIG5        , C_BIG5},
1081         {"zh_TW"        , C_BIG5        , C_BIG5},
1082
1083         {"ru_RU.KOI8-R" , C_KOI8_R      , C_KOI8_R},
1084         {"ru_RU.KOI8R"  , C_KOI8_R      , C_KOI8_R},
1085         {"ru_RU.CP1251" , C_WINDOWS_1251, C_KOI8_R},
1086 #ifdef G_OS_WIN32
1087         {"ru_RU"        , C_WINDOWS_1251, C_KOI8_R},
1088 #else
1089         {"ru_RU"        , C_ISO_8859_5  , C_KOI8_R},
1090 #endif
1091         {"tg_TJ"        , C_KOI8_T      , C_KOI8_T},
1092         {"ru_UA"        , C_KOI8_U      , C_KOI8_U},
1093         {"uk_UA.CP1251" , C_WINDOWS_1251, C_KOI8_U},
1094         {"uk_UA"        , C_KOI8_U      , C_KOI8_U},
1095
1096         {"be_BY"        , C_WINDOWS_1251, C_WINDOWS_1251},
1097         {"bg_BG"        , C_WINDOWS_1251, C_WINDOWS_1251},
1098
1099         {"yi_US"        , C_WINDOWS_1255, C_WINDOWS_1255},
1100
1101         {"af_ZA"        , C_ISO_8859_1  , C_ISO_8859_1},
1102         {"br_FR"        , C_ISO_8859_1  , C_ISO_8859_1},
1103         {"ca_ES"        , C_ISO_8859_1  , C_ISO_8859_1},
1104         {"da_DK"        , C_ISO_8859_1  , C_ISO_8859_1},
1105         {"de_AT"        , C_ISO_8859_1  , C_ISO_8859_1},
1106         {"de_BE"        , C_ISO_8859_1  , C_ISO_8859_1},
1107         {"de_CH"        , C_ISO_8859_1  , C_ISO_8859_1},
1108         {"de_DE"        , C_ISO_8859_1  , C_ISO_8859_1},
1109         {"de_LU"        , C_ISO_8859_1  , C_ISO_8859_1},
1110         {"en_AU"        , C_ISO_8859_1  , C_ISO_8859_1},
1111         {"en_BW"        , C_ISO_8859_1  , C_ISO_8859_1},
1112         {"en_CA"        , C_ISO_8859_1  , C_ISO_8859_1},
1113         {"en_DK"        , C_ISO_8859_1  , C_ISO_8859_1},
1114         {"en_GB"        , C_ISO_8859_1  , C_ISO_8859_1},
1115         {"en_HK"        , C_ISO_8859_1  , C_ISO_8859_1},
1116         {"en_IE"        , C_ISO_8859_1  , C_ISO_8859_1},
1117         {"en_NZ"        , C_ISO_8859_1  , C_ISO_8859_1},
1118         {"en_PH"        , C_ISO_8859_1  , C_ISO_8859_1},
1119         {"en_SG"        , C_ISO_8859_1  , C_ISO_8859_1},
1120         {"en_US"        , C_ISO_8859_1  , C_ISO_8859_1},
1121         {"en_ZA"        , C_ISO_8859_1  , C_ISO_8859_1},
1122         {"en_ZW"        , C_ISO_8859_1  , C_ISO_8859_1},
1123         {"es_AR"        , C_ISO_8859_1  , C_ISO_8859_1},
1124         {"es_BO"        , C_ISO_8859_1  , C_ISO_8859_1},
1125         {"es_CL"        , C_ISO_8859_1  , C_ISO_8859_1},
1126         {"es_CO"        , C_ISO_8859_1  , C_ISO_8859_1},
1127         {"es_CR"        , C_ISO_8859_1  , C_ISO_8859_1},
1128         {"es_DO"        , C_ISO_8859_1  , C_ISO_8859_1},
1129         {"es_EC"        , C_ISO_8859_1  , C_ISO_8859_1},
1130         {"es_ES"        , C_ISO_8859_1  , C_ISO_8859_1},
1131         {"es_GT"        , C_ISO_8859_1  , C_ISO_8859_1},
1132         {"es_HN"        , C_ISO_8859_1  , C_ISO_8859_1},
1133         {"es_MX"        , C_ISO_8859_1  , C_ISO_8859_1},
1134         {"es_NI"        , C_ISO_8859_1  , C_ISO_8859_1},
1135         {"es_PA"        , C_ISO_8859_1  , C_ISO_8859_1},
1136         {"es_PE"        , C_ISO_8859_1  , C_ISO_8859_1},
1137         {"es_PR"        , C_ISO_8859_1  , C_ISO_8859_1},
1138         {"es_PY"        , C_ISO_8859_1  , C_ISO_8859_1},
1139         {"es_SV"        , C_ISO_8859_1  , C_ISO_8859_1},
1140         {"es_US"        , C_ISO_8859_1  , C_ISO_8859_1},
1141         {"es_UY"        , C_ISO_8859_1  , C_ISO_8859_1},
1142         {"es_VE"        , C_ISO_8859_1  , C_ISO_8859_1},
1143         {"et_EE"        , C_ISO_8859_1  , C_ISO_8859_1},
1144         {"eu_ES"        , C_ISO_8859_1  , C_ISO_8859_1},
1145         {"fi_FI"        , C_ISO_8859_1  , C_ISO_8859_1},
1146         {"fo_FO"        , C_ISO_8859_1  , C_ISO_8859_1},
1147         {"fr_BE"        , C_ISO_8859_1  , C_ISO_8859_1},
1148         {"fr_CA"        , C_ISO_8859_1  , C_ISO_8859_1},
1149         {"fr_CH"        , C_ISO_8859_1  , C_ISO_8859_1},
1150         {"fr_FR"        , C_ISO_8859_1  , C_ISO_8859_1},
1151         {"fr_LU"        , C_ISO_8859_1  , C_ISO_8859_1},
1152         {"ga_IE"        , C_ISO_8859_1  , C_ISO_8859_1},
1153         {"gl_ES"        , C_ISO_8859_1  , C_ISO_8859_1},
1154         {"gv_GB"        , C_ISO_8859_1  , C_ISO_8859_1},
1155         {"id_ID"        , C_ISO_8859_1  , C_ISO_8859_1},
1156         {"is_IS"        , C_ISO_8859_1  , C_ISO_8859_1},
1157         {"it_CH"        , C_ISO_8859_1  , C_ISO_8859_1},
1158         {"it_IT"        , C_ISO_8859_1  , C_ISO_8859_1},
1159         {"kl_GL"        , C_ISO_8859_1  , C_ISO_8859_1},
1160         {"kw_GB"        , C_ISO_8859_1  , C_ISO_8859_1},
1161         {"ms_MY"        , C_ISO_8859_1  , C_ISO_8859_1},
1162         {"nl_BE"        , C_ISO_8859_1  , C_ISO_8859_1},
1163         {"nl_NL"        , C_ISO_8859_1  , C_ISO_8859_1},
1164         {"nb_NO"        , C_ISO_8859_1  , C_ISO_8859_1},
1165         {"nn_NO"        , C_ISO_8859_1  , C_ISO_8859_1},
1166         {"no_NO"        , C_ISO_8859_1  , C_ISO_8859_1},
1167         {"oc_FR"        , C_ISO_8859_1  , C_ISO_8859_1},
1168         {"pt_BR"        , C_ISO_8859_1  , C_ISO_8859_1},
1169         {"pt_PT"        , C_ISO_8859_1  , C_ISO_8859_1},
1170         {"sq_AL"        , C_ISO_8859_1  , C_ISO_8859_1},
1171         {"sv_FI"        , C_ISO_8859_1  , C_ISO_8859_1},
1172         {"sv_SE"        , C_ISO_8859_1  , C_ISO_8859_1},
1173         {"tl_PH"        , C_ISO_8859_1  , C_ISO_8859_1},
1174         {"uz_UZ"        , C_ISO_8859_1  , C_ISO_8859_1},
1175         {"wa_BE"        , C_ISO_8859_1  , C_ISO_8859_1},
1176
1177         {"bs_BA"        , C_ISO_8859_2  , C_ISO_8859_2},
1178         {"cs_CZ"        , C_ISO_8859_2  , C_ISO_8859_2},
1179         {"hr_HR"        , C_ISO_8859_2  , C_ISO_8859_2},
1180         {"hu_HU"        , C_ISO_8859_2  , C_ISO_8859_2},
1181         {"pl_PL"        , C_ISO_8859_2  , C_ISO_8859_2},
1182         {"ro_RO"        , C_ISO_8859_2  , C_ISO_8859_2},
1183         {"sk_SK"        , C_ISO_8859_2  , C_ISO_8859_2},
1184         {"sl_SI"        , C_ISO_8859_2  , C_ISO_8859_2},
1185
1186         {"sr_YU@cyrillic"       , C_ISO_8859_5  , C_ISO_8859_5},
1187         {"sr_YU"                , C_ISO_8859_2  , C_ISO_8859_2},
1188
1189         {"mt_MT"                , C_ISO_8859_3  , C_ISO_8859_3},
1190
1191         {"lt_LT.iso88594"       , C_ISO_8859_4  , C_ISO_8859_4},
1192         {"lt_LT.ISO8859-4"      , C_ISO_8859_4  , C_ISO_8859_4},
1193         {"lt_LT.ISO_8859-4"     , C_ISO_8859_4  , C_ISO_8859_4},
1194         {"lt_LT"                , C_ISO_8859_13 , C_ISO_8859_13},
1195
1196         {"mk_MK"        , C_ISO_8859_5  , C_ISO_8859_5},
1197
1198         {"ar_AE"        , C_ISO_8859_6  , C_ISO_8859_6},
1199         {"ar_BH"        , C_ISO_8859_6  , C_ISO_8859_6},
1200         {"ar_DZ"        , C_ISO_8859_6  , C_ISO_8859_6},
1201         {"ar_EG"        , C_ISO_8859_6  , C_ISO_8859_6},
1202         {"ar_IQ"        , C_ISO_8859_6  , C_ISO_8859_6},
1203         {"ar_JO"        , C_ISO_8859_6  , C_ISO_8859_6},
1204         {"ar_KW"        , C_ISO_8859_6  , C_ISO_8859_6},
1205         {"ar_LB"        , C_ISO_8859_6  , C_ISO_8859_6},
1206         {"ar_LY"        , C_ISO_8859_6  , C_ISO_8859_6},
1207         {"ar_MA"        , C_ISO_8859_6  , C_ISO_8859_6},
1208         {"ar_OM"        , C_ISO_8859_6  , C_ISO_8859_6},
1209         {"ar_QA"        , C_ISO_8859_6  , C_ISO_8859_6},
1210         {"ar_SA"        , C_ISO_8859_6  , C_ISO_8859_6},
1211         {"ar_SD"        , C_ISO_8859_6  , C_ISO_8859_6},
1212         {"ar_SY"        , C_ISO_8859_6  , C_ISO_8859_6},
1213         {"ar_TN"        , C_ISO_8859_6  , C_ISO_8859_6},
1214         {"ar_YE"        , C_ISO_8859_6  , C_ISO_8859_6},
1215
1216         {"el_GR"        , C_ISO_8859_7  , C_ISO_8859_7},
1217         {"he_IL"        , C_ISO_8859_8  , C_ISO_8859_8},
1218         {"iw_IL"        , C_ISO_8859_8  , C_ISO_8859_8},
1219         {"tr_TR"        , C_ISO_8859_9  , C_ISO_8859_9},
1220
1221         {"lv_LV"        , C_ISO_8859_13 , C_ISO_8859_13},
1222         {"mi_NZ"        , C_ISO_8859_13 , C_ISO_8859_13},
1223
1224         {"cy_GB"        , C_ISO_8859_14 , C_ISO_8859_14},
1225
1226         {"ar_IN"        , C_UTF_8       , C_UTF_8},
1227         {"en_IN"        , C_UTF_8       , C_UTF_8},
1228         {"se_NO"        , C_UTF_8       , C_UTF_8},
1229         {"ta_IN"        , C_UTF_8       , C_UTF_8},
1230         {"te_IN"        , C_UTF_8       , C_UTF_8},
1231         {"ur_PK"        , C_UTF_8       , C_UTF_8},
1232
1233         {"th_TH"        , C_TIS_620     , C_TIS_620},
1234         /* {"th_TH"     , C_WINDOWS_874}, */
1235         /* {"th_TH"     , C_ISO_8859_11}, */
1236
1237         {"ka_GE"        , C_GEORGIAN_PS , C_GEORGIAN_PS},
1238         {"vi_VN.TCVN"   , C_TCVN5712_1  , C_TCVN5712_1},
1239
1240         {"C"                    , C_US_ASCII    , C_US_ASCII},
1241         {"POSIX"                , C_US_ASCII    , C_US_ASCII},
1242         {"ANSI_X3.4-1968"       , C_US_ASCII    , C_US_ASCII},
1243 };
1244
1245 static GHashTable *conv_get_charset_to_str_table(void)
1246 {
1247         static GHashTable *table;
1248         gint i;
1249
1250         if (table)
1251                 return table;
1252
1253         table = g_hash_table_new(NULL, g_direct_equal);
1254
1255         for (i = 0; i < sizeof(charsets) / sizeof(charsets[0]); i++) {
1256                 if (g_hash_table_lookup(table, GUINT_TO_POINTER(charsets[i].charset))
1257                     == NULL) {
1258                         g_hash_table_insert
1259                                 (table, GUINT_TO_POINTER(charsets[i].charset),
1260                                  charsets[i].name);
1261                 }
1262         }
1263
1264         return table;
1265 }
1266
1267 static GHashTable *conv_get_charset_from_str_table(void)
1268 {
1269         static GHashTable *table;
1270         gint i;
1271
1272         if (table)
1273                 return table;
1274
1275         table = g_hash_table_new(str_case_hash, str_case_equal);
1276
1277         for (i = 0; i < sizeof(charsets) / sizeof(charsets[0]); i++) {
1278                 g_hash_table_insert(table, charsets[i].name,
1279                                     GUINT_TO_POINTER(charsets[i].charset));
1280         }
1281
1282         return table;
1283 }
1284
1285 const gchar *conv_get_charset_str(CharSet charset)
1286 {
1287         GHashTable *table;
1288
1289         table = conv_get_charset_to_str_table();
1290         return g_hash_table_lookup(table, GUINT_TO_POINTER(charset));
1291 }
1292
1293 CharSet conv_get_charset_from_str(const gchar *charset)
1294 {
1295         GHashTable *table;
1296
1297         if (!charset) return C_AUTO;
1298
1299         table = conv_get_charset_from_str_table();
1300         return GPOINTER_TO_UINT(g_hash_table_lookup(table, charset));
1301 }
1302
1303 static CharSet conv_get_locale_charset(void)
1304 {
1305         static CharSet cur_charset = -1;
1306         const gchar *cur_locale;
1307         const gchar *p;
1308         gint i;
1309
1310         if (cur_charset != -1)
1311                 return cur_charset;
1312
1313         cur_locale = conv_get_current_locale();
1314         if (!cur_locale) {
1315                 cur_charset = C_US_ASCII;
1316                 return cur_charset;
1317         }
1318
1319         if (strcasestr(cur_locale, "UTF-8") ||
1320             strcasestr(cur_locale, "utf8")) {
1321                 cur_charset = C_UTF_8;
1322                 return cur_charset;
1323         }
1324
1325         if ((p = strcasestr(cur_locale, "@euro")) && p[5] == '\0') {
1326                 cur_charset = C_ISO_8859_15;
1327                 return cur_charset;
1328         }
1329
1330         for (i = 0; i < sizeof(locale_table) / sizeof(locale_table[0]); i++) {
1331                 const gchar *p;
1332
1333                 /* "ja_JP.EUC" matches with "ja_JP.eucJP", "ja_JP.EUC" and
1334                    "ja_JP". "ja_JP" matches with "ja_JP.xxxx" and "ja" */
1335                 if (!g_ascii_strncasecmp(cur_locale, locale_table[i].locale,
1336                                  strlen(locale_table[i].locale))) {
1337                         cur_charset = locale_table[i].charset;
1338                         return cur_charset;
1339                 } else if ((p = strchr(locale_table[i].locale, '_')) &&
1340                          !strchr(p + 1, '.')) {
1341                         if (strlen(cur_locale) == 2 &&
1342                             !g_ascii_strncasecmp(cur_locale, locale_table[i].locale, 2)) {
1343                                 cur_charset = locale_table[i].charset;
1344                                 return cur_charset;
1345                         }
1346                 }
1347         }
1348
1349         cur_charset = C_AUTO;
1350         return cur_charset;
1351 }
1352
1353 static CharSet conv_get_locale_charset_no_utf8(void)
1354 {
1355         static CharSet cur_charset = -1;
1356         const gchar *cur_locale;
1357         const gchar *p;
1358         gint i;
1359
1360         if (prefs_common.broken_are_utf8) {
1361                 cur_charset = C_UTF_8;
1362                 return cur_charset;
1363         }
1364
1365         cur_locale = conv_get_current_locale();
1366         if (!cur_locale) {
1367                 cur_charset = C_US_ASCII;
1368                 return cur_charset;
1369         }
1370
1371         if (strcasestr(cur_locale, "UTF-8") ||
1372             strcasestr(cur_locale, "utf8")) {
1373                 cur_charset = C_UTF_8;
1374                 return cur_charset;
1375         }
1376
1377         if ((p = strcasestr(cur_locale, "@euro")) && p[5] == '\0') {
1378                 cur_charset = C_ISO_8859_15;
1379                 return cur_charset;
1380         }
1381
1382         for (i = 0; i < sizeof(locale_table) / sizeof(locale_table[0]); i++) {
1383                 const gchar *p;
1384
1385                 /* "ja_JP.EUC" matches with "ja_JP.eucJP", "ja_JP.EUC" and
1386                    "ja_JP". "ja_JP" matches with "ja_JP.xxxx" and "ja" */
1387                 if (!g_ascii_strncasecmp(cur_locale, locale_table[i].locale,
1388                                  strlen(locale_table[i].locale))) {
1389                         cur_charset = locale_table[i].charset;
1390                         return cur_charset;
1391                 } else if ((p = strchr(locale_table[i].locale, '_')) &&
1392                          !strchr(p + 1, '.')) {
1393                         if (strlen(cur_locale) == 2 &&
1394                             !g_ascii_strncasecmp(cur_locale, locale_table[i].locale, 2)) {
1395                                 cur_charset = locale_table[i].charset;
1396                                 return cur_charset;
1397                         }
1398                 }
1399         }
1400
1401         cur_charset = C_AUTO;
1402         return cur_charset;
1403 }
1404
1405 const gchar *conv_get_locale_charset_str(void)
1406 {
1407         static const gchar *codeset = NULL;
1408
1409         if (!codeset)
1410                 codeset = conv_get_charset_str(conv_get_locale_charset());
1411
1412         return codeset ? codeset : CS_INTERNAL;
1413 }
1414
1415 const gchar *conv_get_locale_charset_str_no_utf8(void)
1416 {
1417         static const gchar *codeset = NULL;
1418
1419         if (!codeset)
1420                 codeset = conv_get_charset_str(conv_get_locale_charset_no_utf8());
1421
1422         return codeset ? codeset : CS_INTERNAL;
1423 }
1424
1425 static CharSet conv_get_outgoing_charset(void)
1426 {
1427         static CharSet out_charset = -1;
1428         const gchar *cur_locale;
1429         const gchar *p;
1430         gint i;
1431
1432         if (out_charset != -1)
1433                 return out_charset;
1434
1435         cur_locale = conv_get_current_locale();
1436         if (!cur_locale) {
1437                 out_charset = C_AUTO;
1438                 return out_charset;
1439         }
1440
1441         if (strcasestr(cur_locale, "UTF-8") ||
1442             strcasestr(cur_locale, "utf8")) {
1443                 out_charset = C_UTF_8;
1444                 return out_charset;
1445         }
1446
1447         if ((p = strcasestr(cur_locale, "@euro")) && p[5] == '\0') {
1448                 out_charset = C_ISO_8859_15;
1449                 return out_charset;
1450         }
1451
1452         for (i = 0; i < sizeof(locale_table) / sizeof(locale_table[0]); i++) {
1453                 const gchar *p;
1454
1455                 if (!g_ascii_strncasecmp(cur_locale, locale_table[i].locale,
1456                                  strlen(locale_table[i].locale))) {
1457                         out_charset = locale_table[i].out_charset;
1458                         break;
1459                 } else if ((p = strchr(locale_table[i].locale, '_')) &&
1460                          !strchr(p + 1, '.')) {
1461                         if (strlen(cur_locale) == 2 &&
1462                             !g_ascii_strncasecmp(cur_locale, locale_table[i].locale, 2)) {
1463                                 out_charset = locale_table[i].out_charset;
1464                                 break;
1465                         }
1466                 }
1467         }
1468
1469         return out_charset;
1470 }
1471
1472 const gchar *conv_get_outgoing_charset_str(void)
1473 {
1474         CharSet out_charset;
1475         const gchar *str;
1476
1477         out_charset = conv_get_outgoing_charset();
1478         str = conv_get_charset_str(out_charset);
1479
1480         return str ? str : CS_UTF_8;
1481 }
1482
1483 const gchar *conv_get_current_locale(void)
1484 {
1485         const gchar *cur_locale;
1486
1487 #ifdef G_OS_WIN32
1488         cur_locale = g_win32_getlocale();
1489 #else
1490         cur_locale = g_getenv("LC_ALL");
1491         if (!cur_locale) cur_locale = g_getenv("LC_CTYPE");
1492         if (!cur_locale) cur_locale = g_getenv("LANG");
1493         if (!cur_locale) cur_locale = setlocale(LC_CTYPE, NULL);
1494 #endif /* G_OS_WIN32 */
1495
1496         debug_print("current locale: %s\n",
1497                     cur_locale ? cur_locale : "(none)");
1498
1499         return cur_locale;
1500 }
1501
1502 static gboolean conv_is_ja_locale(void)
1503 {
1504         static gint is_ja_locale = -1;
1505         const gchar *cur_locale;
1506
1507         if (is_ja_locale != -1)
1508                 return is_ja_locale != 0;
1509
1510         is_ja_locale = 0;
1511         cur_locale = conv_get_current_locale();
1512         if (cur_locale) {
1513                 if (g_ascii_strncasecmp(cur_locale, "ja", 2) == 0)
1514                         is_ja_locale = 1;
1515         }
1516
1517         return is_ja_locale != 0;
1518 }
1519
1520 gchar *conv_unmime_header(const gchar *str, const gchar *default_encoding,
1521                            gboolean addr_field)
1522 {
1523         gchar buf[BUFFSIZE];
1524
1525         if (is_ascii_str(str))
1526                 return unmime_header(str, addr_field);
1527
1528         if (default_encoding) {
1529                 gchar *utf8_buf;
1530
1531                 utf8_buf = conv_codeset_strdup
1532                         (str, default_encoding, CS_INTERNAL);
1533                 if (utf8_buf) {
1534                         gchar *decoded_str;
1535
1536                         decoded_str = unmime_header(utf8_buf, addr_field);
1537                         g_free(utf8_buf);
1538                         return decoded_str;
1539                 }
1540         }
1541
1542         if (conv_is_ja_locale())
1543                 conv_anytodisp(buf, sizeof(buf), str);
1544         else
1545                 conv_localetodisp(buf, sizeof(buf), str);
1546
1547         return unmime_header(buf, addr_field);
1548 }
1549
1550 #define MAX_LINELEN             76
1551 #define MAX_HARD_LINELEN        996
1552 #define MIMESEP_BEGIN           "=?"
1553 #define MIMESEP_END             "?="
1554
1555 #define LBREAK_IF_REQUIRED(cond, is_plain_text)                         \
1556 {                                                                       \
1557         if (len - (destp - (guchar *)dest) < MAX_LINELEN + 2) {         \
1558                 *destp = '\0';                                          \
1559                 return;                                                 \
1560         }                                                               \
1561                                                                         \
1562         if ((cond) && *srcp) {                                          \
1563                 if (destp > (guchar *)dest && left < MAX_LINELEN - 1) { \
1564                         if (isspace(*(destp - 1)))                      \
1565                                 destp--;                                \
1566                         else if (is_plain_text && isspace(*srcp))       \
1567                                 srcp++;                                 \
1568                         if (*srcp) {                                    \
1569                                 *destp++ = '\n';                        \
1570                                 *destp++ = ' ';                         \
1571                                 left = MAX_LINELEN - 1;                 \
1572                         }                                               \
1573                 } else if (destp == (guchar *)dest && left < 7) {       \
1574                         if (isspace(*(destp - 1)))                      \
1575                                 destp--;                                \
1576                         else if (is_plain_text && isspace(*srcp))       \
1577                                 srcp++;                                 \
1578                         if (*srcp) {                                    \
1579                                 *destp++ = '\n';                        \
1580                                 *destp++ = ' ';                         \
1581                                 left = MAX_LINELEN - 1;                 \
1582                         }                                               \
1583                 }                                                       \
1584         }                                                               \
1585 }
1586
1587 #define B64LEN(len) ((len) / 3 * 4 + ((len) % 3 ? 4 : 0))
1588
1589 void conv_encode_header_full(gchar *dest, gint len, const gchar *src,
1590                         gint header_len, gboolean addr_field,
1591                         const gchar *out_encoding_)
1592 {
1593         const gchar *cur_encoding;
1594         const gchar *out_encoding;
1595         gint mimestr_len;
1596         gchar *mimesep_enc;
1597         gint left;
1598         const guchar *srcp = src;
1599         guchar *destp = dest;
1600         gboolean use_base64;
1601
1602         cm_return_if_fail(g_utf8_validate(src, -1, NULL) == TRUE);
1603         cm_return_if_fail(destp != NULL);
1604
1605         if (MB_CUR_MAX > 1) {
1606                 use_base64 = TRUE;
1607                 mimesep_enc = "?B?";
1608         } else {
1609                 use_base64 = FALSE;
1610                 mimesep_enc = "?Q?";
1611         }
1612
1613         cur_encoding = CS_INTERNAL;
1614
1615         if (out_encoding_)
1616                 out_encoding = out_encoding_;
1617         else
1618                 out_encoding = conv_get_outgoing_charset_str();
1619
1620         if (!strcmp(out_encoding, CS_US_ASCII))
1621                 out_encoding = CS_ISO_8859_1;
1622
1623         mimestr_len = strlen(MIMESEP_BEGIN) + strlen(out_encoding) +
1624                 strlen(mimesep_enc) + strlen(MIMESEP_END);
1625
1626         left = MAX_LINELEN - header_len;
1627
1628         while (*srcp) {
1629                 LBREAK_IF_REQUIRED(left <= 0, TRUE);
1630
1631                 while (isspace(*srcp)) {
1632                         *destp++ = *srcp++;
1633                         left--;
1634                         LBREAK_IF_REQUIRED(left <= 0, TRUE);
1635                 }
1636
1637                 /* output as it is if the next word is ASCII string */
1638                 if (!is_next_nonascii(srcp)) {
1639                         gint word_len;
1640
1641                         word_len = get_next_word_len(srcp);
1642                         LBREAK_IF_REQUIRED(left < word_len, TRUE);
1643                         while (word_len > 0) {
1644                                 LBREAK_IF_REQUIRED(left + (MAX_HARD_LINELEN - MAX_LINELEN) <= 0, TRUE)
1645                                 *destp++ = *srcp++;
1646                                 left--;
1647                                 word_len--;
1648                         }
1649
1650                         continue;
1651                 }
1652
1653                 /* don't include parentheses and quotes in encoded strings */
1654                 if (addr_field && (*srcp == '(' || *srcp == ')' || *srcp == '"')) {
1655                         LBREAK_IF_REQUIRED(left < 2, FALSE);
1656                         *destp++ = *srcp++;
1657                         left--;
1658                 }
1659
1660                 while (1) {
1661                         gint mb_len = 0;
1662                         gint cur_len = 0;
1663                         gchar *part_str;
1664                         gchar *out_str;
1665                         gchar *enc_str;
1666                         const guchar *p = srcp;
1667                         gint out_str_len;
1668                         gint out_enc_str_len;
1669                         gint mime_block_len;
1670                         gboolean cont = FALSE;
1671
1672                         while (*p != '\0') {
1673                                 if (isspace(*p) && !is_next_nonascii(p + 1))
1674                                         break;
1675                                 /* don't include parentheses in encoded
1676                                    strings */
1677                                 if (addr_field && (*p == '(' || *p == ')' || *p == '"'))
1678                                         break;
1679
1680                                 mb_len = g_utf8_skip[*p];
1681
1682                                 Xstrndup_a(part_str, srcp, cur_len + mb_len, );
1683                                 out_str = conv_codeset_strdup
1684                                         (part_str, cur_encoding, out_encoding);
1685                                 if (!out_str) {
1686                                         if (strict_mode) {
1687                                                 *dest = '\0';
1688                                                 return;
1689                                         } else {
1690                                                 g_warning("conv_encode_header(): code conversion failed");
1691                                                 conv_unreadable_8bit(part_str);
1692                                                 out_str = g_strdup(part_str);
1693                                         }
1694                                 }
1695                                 out_str_len = strlen(out_str);
1696
1697                                 if (use_base64)
1698                                         out_enc_str_len = B64LEN(out_str_len);
1699                                 else
1700                                         out_enc_str_len =
1701                                                 qp_get_q_encoding_len(out_str);
1702
1703                                 g_free(out_str);
1704
1705                                 if (mimestr_len + out_enc_str_len <= left) {
1706                                         cur_len += mb_len;
1707                                         p += mb_len;
1708                                 } else if (cur_len == 0) {
1709                                         left = 0;
1710                                         LBREAK_IF_REQUIRED(1, FALSE);
1711                                         continue;
1712                                 } else {
1713                                         cont = TRUE;
1714                                         break;
1715                                 }
1716                         }
1717
1718                         if (cur_len > 0) {
1719                                 Xstrndup_a(part_str, srcp, cur_len, );
1720                                 out_str = conv_codeset_strdup
1721                                         (part_str, cur_encoding, out_encoding);
1722                                 if (!out_str) {
1723                                         g_warning("conv_encode_header(): code conversion failed");
1724                                         conv_unreadable_8bit(part_str);
1725                                         out_str = g_strdup(part_str);
1726                                 }
1727                                 out_str_len = strlen(out_str);
1728
1729                                 if (use_base64)
1730                                         out_enc_str_len = B64LEN(out_str_len);
1731                                 else
1732                                         out_enc_str_len =
1733                                                 qp_get_q_encoding_len(out_str);
1734
1735                                 if (use_base64)
1736                                         enc_str = g_base64_encode(out_str, out_str_len);
1737                                 else {
1738                                         Xalloca(enc_str, out_enc_str_len + 1, );
1739                                         qp_q_encode(enc_str, out_str);
1740                                 }
1741
1742                                 g_free(out_str);
1743
1744                                 /* output MIME-encoded string block */
1745                                 mime_block_len = mimestr_len + strlen(enc_str);
1746                                 g_snprintf(destp, mime_block_len + 1,
1747                                            MIMESEP_BEGIN "%s%s%s" MIMESEP_END,
1748                                            out_encoding, mimesep_enc, enc_str);
1749
1750                                 if (use_base64)
1751                                         g_free(enc_str);
1752
1753                                 destp += mime_block_len;
1754                                 srcp += cur_len;
1755
1756                                 left -= mime_block_len;
1757                         }
1758
1759                         LBREAK_IF_REQUIRED(cont, FALSE);
1760
1761                         if (cur_len == 0)
1762                                 break;
1763                 }
1764         }
1765
1766         *destp = '\0';
1767 }
1768
1769 void conv_encode_header(gchar *dest, gint len, const gchar *src,
1770                         gint header_len, gboolean addr_field)
1771 {
1772         conv_encode_header_full(dest,len,src,header_len,addr_field,NULL);
1773 }
1774
1775 #undef LBREAK_IF_REQUIRED
1776 #undef B64LEN
1777
1778 gchar *conv_filename_from_utf8(const gchar *utf8_file)
1779 {
1780         gchar *fs_file;
1781         GError *error = NULL;
1782
1783         fs_file = g_filename_from_utf8(utf8_file, -1, NULL, NULL, &error);
1784         if (error) {
1785                 debug_print("failed to convert encoding of file name: %s\n",
1786                           error->message);
1787                 g_error_free(error);
1788         }
1789         if (!fs_file)
1790                 fs_file = g_strdup(utf8_file);
1791
1792         return fs_file;
1793 }
1794
1795 gchar *conv_filename_to_utf8(const gchar *fs_file)
1796 {
1797         gchar *utf8_file = NULL;
1798         GError *error = NULL;
1799
1800         utf8_file = g_filename_to_utf8(fs_file, -1, NULL, NULL, &error);
1801         if (error) {
1802                 g_warning("failed to convert encoding of file name: %s",
1803                           error->message);
1804                 g_error_free(error);
1805         }
1806
1807         if (!utf8_file || !g_utf8_validate(utf8_file, -1, NULL)) {
1808                 g_free(utf8_file);
1809                 utf8_file = g_strdup(fs_file);
1810                 conv_unreadable_8bit(utf8_file);
1811         }
1812
1813         return utf8_file;
1814 }