very small bug fix
[claws.git] / src / codeconv.c
1 /*
2  * Sylpheed -- a GTK+ based, lightweight, and fast e-mail client
3  * Copyright (C) 1999,2000 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., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
18  */
19
20 #ifdef HAVE_CONFIG_H
21 #  include "config.h"
22 #endif
23
24 #include <glib.h>
25 #include <string.h>
26 #include <ctype.h>
27 #include <stdlib.h>
28
29 #if (HAVE_WCTYPE_H && HAVE_WCHAR_H)
30 #  include <wchar.h>
31 #  include <wctype.h>
32 #endif
33
34 #if HAVE_LOCALE_H
35 #  include <locale.h>
36 #endif
37
38 #if HAVE_LIBJCONV
39 #  include <jconv.h>
40 #endif
41
42 #include <kcc.h>
43
44 #include "intl.h"
45 #include "codeconv.h"
46 #include "unmime.h"
47 #include "base64.h"
48 #include "utils.h"
49 #include "prefs_common.h"
50
51 #define iskanji(c) \
52         (((c) & 0xff) >= 0xa1 && ((c) & 0xff) <= 0xfe)
53 #define iseucss(c) \
54         (((c) & 0xff) == 0x8e || ((c) & 0xff) == 0x8f)
55 #define isunprintablekanji(c) \
56         (((c) & 0xff) >= 0xa9 && ((c) & 0xff) <= 0xaf)
57
58 void conv_jistoeuc(gchar *outbuf, gint outlen, const gchar *inbuf)
59 {
60         KCC_filter(outbuf, "EUC", (gchar *)inbuf, "JISBB", 0, 0, 0);
61 }
62
63 void conv_euctojis(gchar *outbuf, gint outlen, const gchar *inbuf)
64 {
65         size_t inlen, len;
66
67         inlen = strlen(inbuf);
68         if (iskanji(inbuf[inlen - 1]) || iseucss(inbuf[inlen - 1])) {
69                 /* if tail end of the string is not ended with ascii,
70                    add dummy return code. */
71                 gchar *tmpin, *tmpout;
72
73                 /* length of original string + '\n' + '\0' */
74                 tmpin = alloca(inlen + 2);
75                 if (tmpin == NULL) {
76                         g_warning(_("can't allocate memory\n"));
77                         KCC_filter(outbuf, "JISBB", (gchar *)inbuf, "EUC",
78                                    0, 0, 0);
79                         return;
80                 }
81                 strcpy(tmpin, inbuf);
82                 tmpin[inlen] = '\n';
83                 tmpin[inlen + 1] = '\0';
84
85                 tmpout = alloca(outlen + 1);
86                 if (tmpout == NULL) {
87                         g_warning(_("can't allocate memory\n"));
88                         KCC_filter(outbuf, "JISBB", (gchar *)inbuf, "EUC",
89                                    0, 0, 0);
90                         return;
91                 }
92
93                 KCC_filter(tmpout, "JISBB", tmpin, "EUC", 0, 0, 0);
94                 len = strlen(tmpout);
95                 if (tmpout[len - 1] == '\n')
96                         tmpout[len - 1] = '\0';
97                 strncpy2(outbuf, tmpout, outlen);
98         } else
99                 KCC_filter(outbuf, "JISBB", (gchar *)inbuf, "EUC", 0, 0, 0);
100 }
101
102 void conv_sjistoeuc(gchar *outbuf, gint outlen, const gchar *inbuf)
103 {
104         KCC_filter(outbuf, "EUC", (gchar *)inbuf, "SJIS", 0, 0, 0);
105 }
106
107 void conv_anytoeuc(gchar *outbuf, gint outlen, const gchar *inbuf)
108 {
109         KCC_filter(outbuf, "EUC", (gchar *)inbuf, "AUTO", 0, 0, 0);
110 }
111
112 void conv_anytojis(gchar *outbuf, gint outlen, const gchar *inbuf)
113 {
114         KCC_filter(outbuf, "JISBB", (gchar *)inbuf, "AUTO", 0, 0, 0);
115 }
116
117 #define SUBST_CHAR      '_'
118
119 void conv_unreadable_eucjp(gchar *str)
120 {
121         register guchar *p = str;
122
123         while (*p != '\0') {
124                 if (isascii(*p)) {
125                         /* convert CR+LF -> LF */
126                         if (*p == '\r' && *(p + 1) == '\n')
127                                 memmove(p, p + 1, strlen(p));
128                         /* printable 7 bit code */
129                         p++;
130                 } else if (iskanji(*p)) {
131                         if (iskanji(*(p + 1)) && !isunprintablekanji(*p))
132                                 /* printable euc-jp code */
133                                 p += 2;
134                         else {
135                                 /* substitute unprintable code */
136                                 *p++ = SUBST_CHAR;
137                                 if (*p != '\0') {
138                                         if (isascii(*p))
139                                                 p++;
140                                         else
141                                                 *p++ = SUBST_CHAR;
142                                 }
143                         }
144                 } else if (iseucss(*p)) {
145                         if ((*(p + 1) & 0x80) != 0)
146                                 /* euc-jp hankaku kana */
147                                 p += 2;
148                         else
149                                 *p++ = SUBST_CHAR;
150                 } else
151                         /* substitute unprintable 1 byte code */
152                         *p++ = SUBST_CHAR;
153         }
154 }
155
156 void conv_unreadable_8bit(gchar *str)
157 {
158         register guchar *p = str;
159
160         while (*p != '\0') {
161                 /* convert CR+LF -> LF */
162                 if (*p == '\r' && *(p + 1) == '\n')
163                         memmove(p, p + 1, strlen(p));
164                 else if (!isascii(*p)) *p = SUBST_CHAR;
165                 p++;
166         }
167 }
168
169 void conv_unreadable_latin(gchar *str)
170 {
171         register guchar *p = str;
172
173         while (*p != '\0') {
174                 /* convert CR+LF -> LF */
175                 if (*p == '\r' && *(p + 1) == '\n')
176                         memmove(p, p + 1, strlen(p));
177                 else if ((*p & 0xff) >= 0x80 && (*p & 0xff) <= 0x9f)
178                         *p = SUBST_CHAR;
179                 p++;
180         }
181 }
182
183 #define NCV     '\0'
184
185 void conv_mb_alnum(gchar *str)
186 {
187         static guchar char_tbl[] = {
188                 /* 0xa0 - 0xaf */
189                 NCV, ' ', NCV, NCV, ',', '.', NCV, ':',
190                 ';', '?', '!', NCV, NCV, NCV, NCV, NCV,
191                 /* 0xb0 - 0xbf */
192                 NCV, NCV, NCV, NCV, NCV, NCV, NCV, NCV,
193                 NCV, NCV, NCV, NCV, NCV, NCV, NCV, NCV,
194                 /* 0xc0 - 0xcf */
195                 NCV, NCV, NCV, NCV, NCV, NCV, NCV, NCV,
196                 NCV, NCV, '(', ')', NCV, NCV, '[', ']',
197                 /* 0xd0 - 0xdf */
198                 '{', '}', NCV, NCV, NCV, NCV, NCV, NCV,
199                 NCV, NCV, NCV, NCV, '+', '-', NCV, NCV,
200                 /* 0xe0 - 0xef */
201                 NCV, '=', NCV, '<', '>', NCV, NCV, NCV,
202                 NCV, NCV, NCV, NCV, NCV, NCV, NCV, NCV
203         };
204
205         register guchar *p = str;
206         register gint len;
207
208         len = strlen(str);
209
210         while (len > 1) {
211                 if (*p == 0xa3) {
212                         register guchar ch = *(p + 1);
213
214                         if (ch >= 0xb0 && ch <= 0xfa) {
215                                 /* [a-zA-Z] */
216                                 *p = ch & 0x7f;
217                                 p++;
218                                 len--;
219                                 memmove(p, p + 1, len);
220                                 len--;
221                         } else  {
222                                 p += 2;
223                                 len -= 2;
224                         }
225                 } else if (*p == 0xa1) {
226                         register guchar ch = *(p + 1);
227
228                         if (ch >= 0xa0 && ch <= 0xef &&
229                             NCV != char_tbl[ch - 0xa0]) {
230                                 *p = char_tbl[ch - 0xa0];
231                                 p++;
232                                 len--;
233                                 memmove(p, p + 1, len);
234                                 len--;
235                         } else {
236                                 p += 2;
237                                 len -= 2;
238                         }
239                 } else if (iskanji(*p)) {
240                         p += 2;
241                         len -= 2;
242                 } else {
243                         p++;
244                         len--;
245                 }
246         }
247 }
248
249 void conv_jistodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
250 {
251         conv_jistoeuc(outbuf, outlen, inbuf);
252         conv_unreadable_eucjp(outbuf);
253 }
254
255 void conv_sjistodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
256 {
257         conv_sjistoeuc(outbuf, outlen, inbuf);
258         conv_unreadable_eucjp(outbuf);
259 }
260
261 void conv_euctodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
262 {
263         strncpy2(outbuf, inbuf, outlen);
264         conv_unreadable_eucjp(outbuf);
265 }
266
267 void conv_ustodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
268 {
269         strncpy2(outbuf, inbuf, outlen);
270         conv_unreadable_8bit(outbuf);
271 }
272
273 void conv_latintodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
274 {
275         strncpy2(outbuf, inbuf, outlen);
276         conv_unreadable_latin(outbuf);
277 }
278
279 void conv_noconv(gchar *outbuf, gint outlen, const gchar *inbuf)
280 {
281         strncpy2(outbuf, inbuf, outlen);
282 }
283
284 CodeConverter *conv_code_converter_new(const gchar *charset)
285 {
286         CodeConverter *conv;
287
288         conv = g_new0(CodeConverter, 1);
289 #if !HAVE_LIBJCONV
290         conv->code_conv_func = conv_get_code_conv_func(charset);
291 #endif
292         conv->charset_str = g_strdup(charset);
293         conv->charset = conv_get_charset_from_str(charset);
294
295         return conv;
296 }
297
298 void conv_code_converter_destroy(CodeConverter *conv)
299 {
300         g_free(conv->charset_str);
301         g_free(conv);
302 }
303
304 gint conv_convert(CodeConverter *conv, gchar *outbuf, gint outlen,
305                   const gchar *inbuf)
306 {
307 #if HAVE_LIBJCONV
308         gchar *str;
309
310         str = conv_codeset_strdup(inbuf, conv->charset_str, NULL);
311         if (!str)
312                 return -1;
313         else {
314                 strncpy2(outbuf, str, outlen);
315                 g_free(str);
316         }
317 #else /* !HAVE_LIBJCONV */
318         conv->code_conv_func(outbuf, outlen, inbuf);
319 #endif
320
321         return 0;
322 }
323
324 gchar *conv_codeset_strdup(const gchar *inbuf,
325                            const gchar *src_codeset, const gchar *dest_codeset)
326 {
327         gchar *buf;
328         size_t len;
329 #if HAVE_LIBJCONV
330         gint actual_codeset;
331         const gchar *const *codesets;
332         gint n_codesets;
333 #else /* !HAVE_LIBJCONV */
334         CharSet src_charset = C_AUTO, dest_charset = C_AUTO;
335 #endif
336
337         if (!dest_codeset) {
338                 CodeConvFunc func;
339
340                 func = conv_get_code_conv_func(src_codeset);
341                 if (func != conv_noconv) {
342                         if (func == conv_jistodisp || func == conv_sjistodisp)
343                                 len = strlen(inbuf) * 2 + 1;
344                         else
345                                 len = strlen(inbuf) + 1;
346                         buf = g_malloc(len);
347                         if (!buf) return NULL;
348                         func(buf, len, inbuf);
349                         buf = g_realloc(buf, strlen(buf) + 1);
350                         return buf;
351                 }
352         }
353
354 #if HAVE_LIBJCONV
355         if (src_codeset) {
356                 codesets = &src_codeset;
357                 n_codesets = 1;
358         } else
359                 codesets = jconv_info_get_pref_codesets(&n_codesets);
360         if (!dest_codeset)
361                 dest_codeset = conv_get_current_charset_str();
362
363         if (jconv_alloc_conv(inbuf, strlen(inbuf), &buf, &len,
364                              codesets, n_codesets,
365                              &actual_codeset, dest_codeset)
366             == 0)
367                 return buf;
368         else
369                 return NULL;
370 #else /* !HAVE_LIBJCONV */
371         if (src_codeset) {
372                 if (!strcasecmp(src_codeset, CS_EUC_JP) ||
373                     !strcasecmp(src_codeset, "EUCJP"))
374                         src_charset = C_EUC_JP;
375                 else if (!strcasecmp(src_codeset, CS_SHIFT_JIS) ||
376                          !strcasecmp(src_codeset, "SHIFT-JIS") ||
377                          !strcasecmp(src_codeset, "SJIS"))
378                         src_charset = C_SHIFT_JIS;
379                 if (dest_codeset && !strcasecmp(dest_codeset, CS_ISO_2022_JP))
380                         dest_charset = C_ISO_2022_JP;
381         }
382
383         if ((src_charset == C_EUC_JP || src_charset == C_SHIFT_JIS) &&
384             dest_charset == C_ISO_2022_JP) {
385                 len = (strlen(inbuf) + 1) * 3;
386                 buf = g_malloc(len);
387                 if (buf) {
388                         if (src_charset == C_EUC_JP)
389                                 conv_euctojis(buf, len, inbuf);
390                         else
391                                 conv_anytojis(buf, len, inbuf);
392                         buf = g_realloc(buf, strlen(buf) + 1);
393                 }
394         } else
395                 buf = g_strdup(inbuf);
396
397         return buf;
398 #endif /* !HAVE_LIBJCONV */
399 }
400
401 CodeConvFunc conv_get_code_conv_func(const gchar *charset)
402 {
403         CodeConvFunc code_conv;
404
405         if (!charset) {
406                 if (conv_get_outgoing_charset() == C_ISO_2022_JP)
407                         return conv_jistodisp;
408                 else
409                         return conv_noconv;
410         }
411
412         if (!strcasecmp(charset, CS_ISO_2022_JP) ||
413             !strcasecmp(charset, CS_ISO_2022_JP_2))
414                 code_conv = conv_jistodisp;
415         else if (!strcasecmp(charset, CS_US_ASCII))
416                 code_conv = conv_ustodisp;
417 #if !HAVE_LIBJCONV
418         else if (!strncasecmp(charset, "ISO-8859-", 9))
419                 code_conv = conv_latintodisp;
420 #endif /* !HAVE_LIBJCONV */
421         else if (!strcasecmp(charset, CS_SHIFT_JIS) ||
422                  !strcasecmp(charset, "SHIFT-JIS")  ||
423                  !strcasecmp(charset, "SJIS")       ||
424                  !strcasecmp(charset, "X-SJIS"))
425                 code_conv = conv_sjistodisp;
426         else if (!strcasecmp(charset, CS_EUC_JP) ||
427                  !strcasecmp(charset, "EUCJP"))
428                 code_conv = conv_euctodisp;
429         else
430                 code_conv = conv_noconv;
431
432         return code_conv;
433 }
434
435 static const struct {
436         CharSet charset;
437         gchar *const name;
438 } charsets[] = {
439         {C_US_ASCII,            CS_US_ASCII},
440         {C_UTF_8,               CS_UTF_8},
441         {C_ISO_8859_1,          CS_ISO_8859_1},
442         {C_ISO_8859_2,          CS_ISO_8859_2},
443         {C_ISO_8859_4,          CS_ISO_8859_4},
444         {C_ISO_8859_5,          CS_ISO_8859_5},
445         {C_ISO_8859_7,          CS_ISO_8859_7},
446         {C_ISO_8859_8,          CS_ISO_8859_8},
447         {C_ISO_8859_9,          CS_ISO_8859_9},
448         {C_ISO_8859_13,         CS_ISO_8859_13},
449         {C_ISO_8859_15,         CS_ISO_8859_15},
450         {C_BALTIC,              CS_BALTIC},
451         {C_CP1251,              CS_CP1251},
452         {C_WINDOWS_1251,        CS_WINDOWS_1251},
453         {C_KOI8_R,              CS_KOI8_R},
454         {C_KOI8_U,              CS_KOI8_U},
455         {C_ISO_2022_JP,         CS_ISO_2022_JP},
456         {C_ISO_2022_JP_2,       CS_ISO_2022_JP_2},
457         {C_EUC_JP,              CS_EUC_JP},
458         {C_SHIFT_JIS,           CS_SHIFT_JIS},
459         {C_ISO_2022_KR,         CS_ISO_2022_KR},
460         {C_EUC_KR,              CS_EUC_KR},
461         {C_ISO_2022_CN,         CS_ISO_2022_CN},
462         {C_EUC_CN,              CS_EUC_CN},
463         {C_GB2312,              CS_GB2312},
464         {C_EUC_TW,              CS_EUC_TW},
465         {C_BIG5,                CS_BIG5},
466 };
467
468 #if !HAVE_LIBJCONV
469 static const struct {
470         gchar *const locale;
471         CharSet charset;
472 } locale_table[] = {
473         {"ja_JP.eucJP"  , C_EUC_JP},
474         {"ja_JP.ujis"   , C_EUC_JP},
475         {"ja_JP.EUC"    , C_EUC_JP},
476         {"ja_JP.SJIS"   , C_SHIFT_JIS},
477         {"ja_JP.JIS"    , C_ISO_2022_JP},
478         {"ja_JP"        , C_EUC_JP},
479         {"ko_KR"        , C_EUC_KR},
480         {"zh_CN.GB2312" , C_GB2312},
481         {"zh_CN"        , C_GB2312},
482         {"zh_TW.eucTW"  , C_EUC_TW},
483         {"zh_TW.Big5"   , C_BIG5},
484         {"zh_TW"        , C_BIG5},
485
486         {"ru_RU.KOI8-R" , C_KOI8_R},
487         {"ru_RU.CP1251" , C_CP1251},
488
489         {"en_US"        , C_ISO_8859_1},
490         {"ca_ES"        , C_ISO_8859_1},
491         {"da_DK"        , C_ISO_8859_1},
492         {"de_DE"        , C_ISO_8859_1},
493         {"nl_NL"        , C_ISO_8859_1},
494         {"et_EE"        , C_ISO_8859_1},
495         {"fi_FI"        , C_ISO_8859_1},
496         {"fr_FR"        , C_ISO_8859_1},
497         {"is_IS"        , C_ISO_8859_1},
498         {"it_IT"        , C_ISO_8859_1},
499         {"no_NO"        , C_ISO_8859_1},
500         {"pt_PT"        , C_ISO_8859_1},
501         {"pt_BR"        , C_ISO_8859_1},
502         {"es_ES"        , C_ISO_8859_1},
503         {"sv_SE"        , C_ISO_8859_1},
504
505         {"hr_HR"        , C_ISO_8859_2},
506         {"hu_HU"        , C_ISO_8859_2},
507         {"pl_PL"        , C_ISO_8859_2},
508         {"ro_RO"        , C_ISO_8859_2},
509         {"sk_SK"        , C_ISO_8859_2},
510         {"sl_SI"        , C_ISO_8859_2},
511         {"ru_RU"        , C_ISO_8859_5},
512         {"el_GR"        , C_ISO_8859_7},
513         {"iw_IL"        , C_ISO_8859_8},
514         {"tr_TR"        , C_ISO_8859_9},
515
516         {"lt_LT.iso88594"       , C_ISO_8859_4},
517         {"lt_LT.ISO8859-4"      , C_ISO_8859_4},
518         {"lt_LT.ISO_8859-4"     , C_ISO_8859_4},
519         {"lt_LT"                , C_ISO_8859_13},
520         {"lv_LV"                , C_ISO_8859_13},
521
522         {"C"            , C_US_ASCII},
523         {"POSIX"        , C_US_ASCII},
524 };
525 #endif /* !HAVE_LIBJCONV */
526
527 const gchar *conv_get_charset_str(CharSet charset)
528 {
529         gint i;
530
531         for (i = 0; i < sizeof(charsets) / sizeof(charsets[0]); i++) {
532                 if (charsets[i].charset == charset)
533                         return charsets[i].name;
534         }
535
536         return NULL;
537 }
538
539 CharSet conv_get_charset_from_str(const gchar *charset)
540 {
541         gint i;
542
543         if (!charset) return C_AUTO;
544
545         for (i = 0; i < sizeof(charsets) / sizeof(charsets[0]); i++) {
546                 if (!strcasecmp(charsets[i].name, charset))
547                         return charsets[i].charset;
548         }
549
550         return C_AUTO;
551 }
552
553 CharSet conv_get_current_charset(void)
554 {
555         static CharSet cur_charset = -1;
556         gint i;
557
558 #if HAVE_LIBJCONV
559         const gchar *cur_codeset;
560 #else
561         const gchar *cur_locale;
562 #endif
563
564         if (cur_charset != -1)
565                 return cur_charset;
566
567 #if HAVE_LIBJCONV
568         cur_codeset = jconv_info_get_current_codeset();
569         for (i = 0; i < sizeof(charsets) / sizeof(charsets[0]); i++) {
570                 if (!strcasecmp(cur_codeset, charsets[i].name)) {
571                         cur_charset = charsets[i].charset;
572                         return cur_charset;
573                 }
574         }
575 #else
576         cur_locale = g_getenv("LC_ALL");
577         if (!cur_locale) cur_locale = g_getenv("LC_CTYPE");
578         if (!cur_locale) cur_locale = g_getenv("LANG");
579         if (!cur_locale) cur_locale = setlocale(LC_CTYPE, NULL);
580
581         debug_print("current locale: %s\n",
582                     cur_locale ? cur_locale : "(none)");
583
584         if (!cur_locale) {
585                 cur_charset = C_US_ASCII;
586                 return cur_charset;
587         }
588
589         if (strcasestr(cur_locale, "UTF-8")) {
590                 cur_charset = C_UTF_8;
591                 return cur_charset;
592         }
593
594         for (i = 0; i < sizeof(locale_table) / sizeof(locale_table[0]); i++) {
595                 const gchar *p;
596
597                 /* "ja_JP.EUC" matches with "ja_JP.eucJP" and "ja_JP.EUC" */
598                 /* "ja_JP" matches with "ja_JP.xxxx" and "ja" */
599                 if (!strncasecmp(cur_locale, locale_table[i].locale,
600                                  strlen(locale_table[i].locale))) {
601                         cur_charset = locale_table[i].charset;
602                         return cur_charset;
603                 } else if ((p = strchr(locale_table[i].locale, '_')) &&
604                          !strchr(p + 1, '.')) {
605                         if (strlen(cur_locale) == 2 &&
606                             !strncasecmp(cur_locale, locale_table[i].locale, 2)) {
607                                 cur_charset = locale_table[i].charset;
608                                 return cur_charset;
609                         }
610                 }
611         }
612 #endif
613
614         cur_charset = C_AUTO;
615         return cur_charset;
616 }
617
618 const gchar *conv_get_current_charset_str(void)
619 {
620 #if HAVE_LIBJCONV
621         return jconv_info_get_current_codeset();
622 #else
623         static const gchar *codeset = NULL;
624
625         if (!codeset)
626                 codeset = conv_get_charset_str(conv_get_current_charset());
627
628         return codeset ? codeset : "US-ASCII";
629 #endif
630 }
631
632 CharSet conv_get_outgoing_charset(void)
633 {
634         static CharSet out_charset = -1;
635
636 #if HAVE_LIBJCONV
637         gint i, j, n_pref_codesets;
638         const gchar *const *pref_codesets;
639 #else
640         CharSet cur_charset;
641 #endif
642
643         if (out_charset != -1)
644                 return out_charset;
645
646 #if HAVE_LIBJCONV
647         /* skip US-ASCII and UTF-8 */
648         pref_codesets = jconv_info_get_pref_codesets(&n_pref_codesets);
649         for (i = 0; i < n_pref_codesets; i++) {
650                 for (j = 2; j < sizeof(charsets) / sizeof(charsets[0]); j++) {
651                         if (!strcasecmp(pref_codesets[i], charsets[j].name)) {
652                                 out_charset = charsets[j].charset;
653                                 return out_charset;
654                         }
655                 }
656         }
657
658         for (i = 0; i < n_pref_codesets; i++) {
659                 if (!strcasecmp(pref_codesets[i], "UTF-8")) {
660                         out_charset = C_UTF_8;
661                         return out_charset;
662                 }
663         }
664
665         out_charset = C_AUTO;
666 #else
667         cur_charset = conv_get_current_charset();
668         switch (cur_charset) {
669         case C_EUC_JP:
670         case C_SHIFT_JIS:
671                 out_charset = C_ISO_2022_JP;
672                 break;
673         default:
674                 out_charset = cur_charset;
675         }
676 #endif
677
678         return out_charset;
679 }
680
681 const gchar *conv_get_outgoing_charset_str(void)
682 {
683         CharSet out_charset;
684         const gchar *str;
685
686         if (prefs_common.outgoing_charset) {
687                 if (!isalpha(prefs_common.outgoing_charset[0])) {
688                         g_free(prefs_common.outgoing_charset);
689                         prefs_common.outgoing_charset = g_strdup(CS_AUTO);
690                 } else if (strcmp(prefs_common.outgoing_charset, CS_AUTO) != 0)
691                         return prefs_common.outgoing_charset;
692         }
693
694         out_charset = conv_get_outgoing_charset();
695         str = conv_get_charset_str(out_charset);
696
697         return str ? str : "US-ASCII";
698 }
699
700 void conv_unmime_header_overwrite(gchar *str)
701 {
702         gchar *buf;
703         gint outlen;
704         CharSet cur_charset;
705
706         cur_charset = conv_get_current_charset();
707
708 #if HAVE_LIBJCONV
709         Xstrdup_a(buf, str, {return;});
710         outlen = strlen(str) + 1;
711         UnMimeHeaderConv(buf, str, outlen);
712         if (cur_charset == C_EUC_JP)
713                 conv_unreadable_eucjp(str);
714 #else
715         if (cur_charset == C_EUC_JP) {
716                 gchar *tmp;
717                 gint len;
718
719                 Xstrdup_a(buf, str, {return;});
720                 outlen = strlen(str) + 1;
721                 UnMimeHeader(buf);
722                 len = strlen(buf) * 2 + 1;
723                 Xalloca(tmp, len, {strncpy2(str, buf, outlen); return;});
724                 conv_jistodisp(tmp, len, buf);
725                 strncpy2(str, tmp, outlen);
726         } else
727                 UnMimeHeader(str);
728 #endif
729 }
730
731 void conv_unmime_header(gchar *outbuf, gint outlen, const gchar *str,
732                         const gchar *charset)
733 {
734         gchar *buf;
735         CharSet cur_charset;
736
737         Xstrdup_a(buf, str, {return;});
738         cur_charset = conv_get_current_charset();
739
740 #if HAVE_LIBJCONV
741         UnMimeHeaderConv(buf, outbuf, outlen);
742         if (cur_charset == C_EUC_JP)
743                 conv_unreadable_eucjp(outbuf);
744 #else
745         UnMimeHeader(buf);
746         if (cur_charset == C_EUC_JP) {
747                 gchar *tmp;
748                 gint len;
749
750                 len = strlen(buf) * 2 + 1;
751                 Xalloca(tmp, len, {strncpy2(outbuf, buf, outlen); return;});
752                 conv_jistodisp(tmp, len, buf);
753                 strncpy2(outbuf, tmp, outlen);
754         } else
755                 strncpy2(outbuf, buf, outlen);
756 #endif
757 }
758
759 #define MAX_ENCLEN      75
760 #define MAX_LINELEN     76
761
762 #define B64LEN(len)     ((len) / 3 * 4 + ((len) % 3 ? 4 : 0))
763
764 #if HAVE_LIBJCONV
765 void conv_encode_header(gchar *dest, gint len, const gchar *src,
766                         gint header_len)
767 {
768         wchar_t *wsrc;
769         wchar_t *wsrcp;
770         gchar *destp;
771         size_t line_len, mimehdr_len, mimehdr_begin_len;
772         gchar *mimehdr_init = "=?";
773         gchar *mimehdr_end = "?=";
774         gchar *mimehdr_enctype = "?B?";
775         const gchar *mimehdr_charset;
776
777         /* g_print("src = %s\n", src); */
778         mimehdr_charset = conv_get_outgoing_charset_str();
779
780         /* convert to wide-character string */
781         wsrcp = wsrc = strdup_mbstowcs(src);
782
783         mimehdr_len = strlen(mimehdr_init) + strlen(mimehdr_end) +
784                 strlen(mimehdr_charset) + strlen(mimehdr_enctype);
785         mimehdr_begin_len = strlen(mimehdr_init) +
786                 strlen(mimehdr_charset) + strlen(mimehdr_enctype);
787         /* line_len = 1; */
788         line_len = header_len;
789         destp = dest;
790         *dest = '\0';
791
792         g_return_if_fail(wsrc != NULL);
793
794         while (*wsrcp) {
795                 wchar_t *wp, *wtmp, *wtmpp;
796                 gint nspc = 0;
797
798                 /* irresponsible buffer overrun check */
799                 if ((len - (destp - dest)) < (MAX_LINELEN + 1) * 2) break;
800
801                 /* encode string including space
802                    if non-ASCII string follows */
803                 if (is_next_nonascii(wsrcp)) {
804                         wp = wsrcp;
805                         while ((wp = find_wspace(wp)) != NULL)
806                                 if (!is_next_nonascii(wp)) break;
807                 } else
808                         wp = find_wspace(wsrcp);
809
810                 if (wp != NULL) {
811                         wtmp = wcsndup(wsrcp, wp - wsrcp);
812                         wsrcp = wp + 1;
813                         while (iswspace(wsrcp[nspc])) nspc++;
814                 } else {
815                         wtmp = wcsdup(wsrcp);
816                         wsrcp += wcslen(wsrcp);
817                 }
818
819                 wtmpp = wtmp;
820
821                 do {
822                         gint tlen = 0, str_ascii = 1;
823                         gchar *tmp; /* internal codeset */
824                         gchar *raw; /* converted, but not base64 encoded */
825                         register gchar *tmpp;
826                         gint raw_len;
827
828                         tmpp = tmp = g_malloc(wcslen(wtmpp) * MB_CUR_MAX + 1);
829                         *tmp = '\0';
830                         raw = g_strdup("");
831                         raw_len = 0;
832
833                         while (*wtmpp != (wchar_t)0) {
834                                 gint mbl;
835                                 gint dummy;
836                                 gchar *raw_new = NULL;
837                                 int raw_new_len = 0;
838                                 const gchar *src_codeset;
839
840                                 if (*wtmpp < 32 || *wtmpp >= 127)
841                                         str_ascii = 0;
842                                 mbl = wctomb(tmpp, *wtmpp);
843                                 if (mbl == -1) {
844                                         g_warning("invalid wide character\n");
845                                         wtmpp++;
846                                         continue;
847                                 }
848                                 /* g_free(raw); */
849                                 src_codeset = conv_get_current_charset_str();
850                                 /* printf ("tmp = %s, tlen = %d, mbl\n", 
851                                         tmp, tlen, mbl); */
852                                 if (jconv_alloc_conv(tmp, tlen + mbl,
853                                                      &raw_new, &raw_new_len,
854                                                      &src_codeset, 1,
855                                                      &dummy, mimehdr_charset)
856                                     != 0) {
857                                         g_warning("can't convert\n");
858                                         tmpp[0] = '\0';
859                                         continue;
860                                 }
861                                 if (!str_ascii) {
862                                         gint dlen = mimehdr_len +
863                                                 B64LEN(raw_len);
864                                         if ((line_len + dlen +
865                                              (*(wtmpp + 1) ? 0 : nspc) +
866                                              (line_len > 1 ? 1 : 0))
867                                             > MAX_LINELEN) {
868                                                 g_free(raw_new);
869                                                 if (tlen == 0) {
870                                                         *destp++ = '\n';
871                                                         *destp++ = ' ';
872                                                         line_len = 1;
873                                                         str_ascii = 1;
874                                                         continue;
875                                                 } else {
876                                                         *tmpp = '\0';
877                                                         break;
878                                                 }
879                                         }
880                                 } else {
881                                         if ((line_len + tlen + mbl +
882                                              (*(wtmpp + 1) ? 0 : nspc) +
883                                              (line_len > 1 ? 1 : 0))
884                                             > MAX_LINELEN) {
885                                                 g_free(raw_new);
886                                                 if (1 + tlen + mbl +
887                                                     (*(wtmpp + 1) ? 0 : nspc)
888                                                     >= MAX_LINELEN) {
889                                                         *tmpp = '\0';
890                                                         break;
891                                                 }
892                                                 *destp++ = '\n';
893                                                 *destp++ = ' ';
894                                                 line_len = 1;
895                                                 continue;
896                                         }
897                                 }
898
899                                 tmpp += mbl;
900                                 *tmpp = '\0';
901
902                                 tlen += mbl;
903
904                                 g_free(raw);
905                                 raw = raw_new;
906                                 raw_len = raw_new_len;
907
908                                 wtmpp++;
909                         }
910                         /*
911                           g_print("tmp = %s, tlen = %d, mb_seqlen = %d\n",
912                                 tmp, tlen, mb_seqlen);
913                         */                              
914
915                         if (tlen == 0 || raw_len == 0) {
916                                 g_free(tmp);
917                                 g_free(raw);
918                                 continue;
919                         }
920
921                         if (line_len > 1 && destp > dest) {
922                                 *destp++ = ' ';
923                                 *destp = '\0';
924                                 line_len++;
925                         }
926
927                         if (!str_ascii) {
928                                 g_snprintf(destp, len - strlen(dest), "%s%s%s",
929                                            mimehdr_init, mimehdr_charset,
930                                            mimehdr_enctype);
931                                 destp += mimehdr_begin_len;
932                                 line_len += mimehdr_begin_len;
933
934                                 to64frombits(destp, raw, raw_len);
935                                 line_len += strlen(destp);
936                                 destp += strlen(destp);
937
938                                 strcpy(destp, mimehdr_end);
939                                 destp += strlen(mimehdr_end);
940                                 line_len += strlen(mimehdr_end);
941                         } else {
942                                 strcpy(destp, tmp);
943                                 line_len += strlen(destp);
944                                 destp += strlen(destp);
945                         }
946
947                         g_free(tmp);
948                         g_free(raw);
949                         /* g_print("line_len = %d\n\n", line_len); */
950                 } while (*wtmpp != (wchar_t)0);
951
952                 while (iswspace(*wsrcp)) {
953                         gint mbl;
954
955                         mbl = wctomb(destp, *wsrcp++);
956                         if (mbl != -1) {
957                                 destp += mbl;
958                                 line_len += mbl;
959                         }
960                 }
961                 *destp = '\0';
962
963                 g_free(wtmp);
964         }
965
966         g_free(wsrc);
967
968         /* g_print("dest = %s\n", dest);  */
969 }
970 #else /* !HAVE_LIBJCONV */
971
972 #define JIS_SEQLEN      3
973
974 void conv_encode_header(gchar *dest, gint len, const gchar *src,
975                         gint header_len)
976 {
977         wchar_t *wsrc;
978         wchar_t *wsrcp;
979         gchar *destp;
980         size_t line_len, mimehdr_len, mimehdr_begin_len;
981         gchar *mimehdr_init = "=?";
982         gchar *mimehdr_end = "?=";
983         gchar *mimehdr_enctype = "?B?";
984         const gchar *mimehdr_charset;
985
986         /* g_print("src = %s\n", src); */
987         mimehdr_charset = conv_get_outgoing_charset_str();
988         if (strcmp(mimehdr_charset, "ISO-2022-JP") != 0) {
989                 /* currently only supports Japanese */
990                 strncpy2(dest, src, len);
991                 return;
992         }
993
994         /* convert to wide-character string */
995         wsrcp = wsrc = strdup_mbstowcs(src);
996
997         mimehdr_len = strlen(mimehdr_init) + strlen(mimehdr_end) +
998                       strlen(mimehdr_charset) + strlen(mimehdr_enctype);
999         mimehdr_begin_len = strlen(mimehdr_init) +
1000                             strlen(mimehdr_charset) + strlen(mimehdr_enctype);
1001         /* line_len = 1; */
1002         line_len = header_len;
1003         destp = dest;
1004         *dest = '\0';
1005
1006         g_return_if_fail(wsrc != NULL);
1007
1008         while (*wsrcp) {
1009                 wchar_t *wp, *wtmp, *wtmpp;
1010                 gint nspc = 0;
1011
1012                 /* irresponsible buffer overrun check */
1013                 if ((len - (destp - dest)) < (MAX_LINELEN + 1) * 2) break;
1014
1015                 /* encode string including space
1016                    if non-ASCII string follows */
1017                 if (is_next_nonascii(wsrcp)) {
1018                         wp = wsrcp;
1019                         while ((wp = find_wspace(wp)) != NULL)
1020                                 if (!is_next_nonascii(wp)) break;
1021                 } else
1022                         wp = find_wspace(wsrcp);
1023
1024                 if (wp != NULL) {
1025                         wtmp = wcsndup(wsrcp, wp - wsrcp);
1026                         wsrcp = wp + 1;
1027                         while (iswspace(wsrcp[nspc])) nspc++;
1028                 } else {
1029                         wtmp = wcsdup(wsrcp);
1030                         wsrcp += wcslen(wsrcp);
1031                 }
1032
1033                 wtmpp = wtmp;
1034
1035                 do {
1036                         gint prev_mbl = 1, tlen = 0, mb_seqlen = 0;
1037                         gchar *tmp;
1038                         register gchar *tmpp;
1039
1040                         tmpp = tmp = g_malloc(wcslen(wtmpp) * MB_CUR_MAX + 1);
1041                         *tmp = '\0';
1042
1043                         while (*wtmpp != (wchar_t)0) {
1044                                 gint mbl;
1045
1046                                 mbl = wctomb(tmpp, *wtmpp);
1047                                 if (mbl == -1) {
1048                                         g_warning("invalid wide character\n");
1049                                         wtmpp++;
1050                                         continue;
1051                                 }
1052
1053                                 /* length of KI + KO */
1054                                 if (prev_mbl == 1 && mbl == 2)
1055                                         mb_seqlen += JIS_SEQLEN * 2;
1056
1057                                 if (mb_seqlen) {
1058                                         gint dlen = mimehdr_len +
1059                                                 B64LEN(tlen + mb_seqlen + mbl);
1060
1061                                         if ((line_len + dlen +
1062                                              (*(wtmpp + 1) ? 0 : nspc) +
1063                                              (line_len > 1 ? 1 : 0))
1064                                             > MAX_LINELEN) {
1065                                                 if (tlen == 0) {
1066                                                         *destp++ = '\n';
1067                                                         *destp++ = ' ';
1068                                                         line_len = 1;
1069                                                         mb_seqlen = 0;
1070                                                         continue;
1071                                                 } else {
1072                                                         *tmpp = '\0';
1073                                                         break;
1074                                                 }
1075                                         }
1076                                 } else {
1077                                         if ((line_len + tlen + mbl +
1078                                              (*(wtmpp + 1) ? 0 : nspc) +
1079                                              (line_len > 1 ? 1 : 0))
1080                                             > MAX_LINELEN) {
1081                                                 if (1 + tlen + mbl +
1082                                                     (*(wtmpp + 1) ? 0 : nspc)
1083                                                     >= MAX_LINELEN) {
1084                                                         *tmpp = '\0';
1085                                                         break;
1086                                                 }
1087                                                 *destp++ = '\n';
1088                                                 *destp++ = ' ';
1089                                                 line_len = 1;
1090                                                 continue;
1091                                         }
1092                                 }
1093
1094                                 tmpp += mbl;
1095                                 *tmpp = '\0';
1096
1097                                 tlen += mbl;
1098                                 prev_mbl = mbl;
1099
1100                                 wtmpp++;
1101                         }
1102                         /* g_print("tmp = %s, tlen = %d, mb_seqlen = %d\n",
1103                                 tmp, tlen, mb_seqlen);
1104                         */                              
1105
1106                         if (tlen == 0) {
1107                                 g_free(tmp);
1108                                 continue;
1109                         }
1110
1111                         if (line_len > 1 && destp > dest) {
1112                                 *destp++ = ' ';
1113                                 *destp = '\0';
1114                                 line_len++;
1115                         }
1116
1117                         if (mb_seqlen) {
1118                                 gchar *tmp_jis;
1119
1120                                 tmp_jis = g_new(gchar, tlen + mb_seqlen + 1);
1121                                 conv_euctojis(tmp_jis,
1122                                               tlen + mb_seqlen + 1, tmp);
1123                                 g_snprintf(destp, len - strlen(dest), "%s%s%s",
1124                                            mimehdr_init, mimehdr_charset,
1125                                            mimehdr_enctype);
1126                                 destp += mimehdr_begin_len;
1127                                 line_len += mimehdr_begin_len;
1128
1129                                 to64frombits(destp, tmp_jis, strlen(tmp_jis));
1130                                 line_len += strlen(destp);
1131                                 destp += strlen(destp);
1132
1133                                 strcpy(destp, mimehdr_end);
1134                                 destp += strlen(mimehdr_end);
1135                                 line_len += strlen(mimehdr_end);
1136
1137                                 g_free(tmp_jis);
1138                         } else {
1139                                 strcpy(destp, tmp);
1140                                 line_len += strlen(destp);
1141                                 destp += strlen(destp);
1142                         }
1143
1144                         g_free(tmp);
1145                         /* g_print("line_len = %d\n\n", line_len); */
1146                 } while (*wtmpp != (wchar_t)0);
1147
1148                 while (iswspace(*wsrcp)) {
1149                         gint mbl;
1150
1151                         mbl = wctomb(destp, *wsrcp++);
1152                         if (mbl != -1) {
1153                                 destp += mbl;
1154                                 line_len += mbl;
1155                         }
1156                 }
1157                 *destp = '\0';
1158
1159                 g_free(wtmp);
1160         }
1161
1162         g_free(wsrc);
1163
1164         /* g_print("dest = %s\n", dest); */
1165 }
1166 #endif /* HAVE_LIBJCONV */