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         else if (!strncasecmp(charset, CS_ISO_8859_1, 10))
 418                 code_conv = conv_latintodisp;
 419 #if !HAVE_LIBJCONV
 420         else if (!strncasecmp(charset, "ISO-8859-", 9))
 421                 code_conv = conv_latintodisp;
 422 #endif
 423         else if (!strcasecmp(charset, CS_SHIFT_JIS) ||
 424                  !strcasecmp(charset, "SHIFT-JIS")  ||
 425                  !strcasecmp(charset, "SJIS")       ||
 426                  !strcasecmp(charset, "X-SJIS"))
 427                 code_conv = conv_sjistodisp;
 428         else if (!strcasecmp(charset, CS_EUC_JP) ||
 429                  !strcasecmp(charset, "EUCJP"))
 430                 code_conv = conv_euctodisp;
 431         else
 432                 code_conv = conv_noconv;
 433
 434         return code_conv;
 435 }
 436
 437 static const struct {
 438         CharSet charset;
 439         gchar *const name;
 440 } charsets[] = {
 441         {C_US_ASCII,            CS_US_ASCII},
 442         {C_UTF_8,               CS_UTF_8},
 443         {C_ISO_8859_1,          CS_ISO_8859_1},
 444         {C_ISO_8859_2,          CS_ISO_8859_2},
 445         {C_ISO_8859_4,          CS_ISO_8859_4},
 446         {C_ISO_8859_5,          CS_ISO_8859_5},
 447         {C_ISO_8859_7,          CS_ISO_8859_7},
 448         {C_ISO_8859_8,          CS_ISO_8859_8},
 449         {C_ISO_8859_9,          CS_ISO_8859_9},
 450         {C_ISO_8859_13,         CS_ISO_8859_13},
 451         {C_ISO_8859_15,         CS_ISO_8859_15},
 452         {C_BALTIC,              CS_BALTIC},
 453         {C_CP1251,              CS_CP1251},
 454         {C_WINDOWS_1251,        CS_WINDOWS_1251},
 455         {C_KOI8_R,              CS_KOI8_R},
 456         {C_KOI8_U,              CS_KOI8_U},
 457         {C_ISO_2022_JP,         CS_ISO_2022_JP},
 458         {C_ISO_2022_JP_2,       CS_ISO_2022_JP_2},
 459         {C_EUC_JP,              CS_EUC_JP},
 460         {C_SHIFT_JIS,           CS_SHIFT_JIS},
 461         {C_ISO_2022_KR,         CS_ISO_2022_KR},
 462         {C_EUC_KR,              CS_EUC_KR},
 463         {C_ISO_2022_CN,         CS_ISO_2022_CN},
 464         {C_EUC_CN,              CS_EUC_CN},
 465         {C_GB2312,              CS_GB2312},
 466         {C_EUC_TW,              CS_EUC_TW},
 467         {C_BIG5,                CS_BIG5},
 468 };
 469
 470 #if !HAVE_LIBJCONV
 471 static const struct {
 472         gchar *const locale;
 473         CharSet charset;
 474 } locale_table[] = {
 475         {"ja_JP.eucJP"  , C_EUC_JP},
 476         {"ja_JP.ujis"   , C_EUC_JP},
 477         {"ja_JP.EUC"    , C_EUC_JP},
 478         {"ja_JP.SJIS"   , C_SHIFT_JIS},
 479         {"ja_JP.JIS"    , C_ISO_2022_JP},
 480         {"ja_JP"        , C_EUC_JP},
 481         {"ko_KR"        , C_EUC_KR},
 482         {"zh_CN.GB2312" , C_GB2312},
 483         {"zh_CN"        , C_GB2312},
 484         {"zh_TW.eucTW"  , C_EUC_TW},
 485         {"zh_TW.Big5"   , C_BIG5},
 486         {"zh_TW"        , C_BIG5},
 487
 488         {"ru_RU.KOI8-R" , C_KOI8_R},
 489         {"ru_RU.CP1251" , C_CP1251},
 490
 491         {"en_US"        , C_ISO_8859_1},
 492         {"ca_ES"        , C_ISO_8859_1},
 493         {"da_DK"        , C_ISO_8859_1},
 494         {"de_DE"        , C_ISO_8859_1},
 495         {"nl_NL"        , C_ISO_8859_1},
 496         {"et_EE"        , C_ISO_8859_1},
 497         {"fi_FI"        , C_ISO_8859_1},
 498         {"fr_FR"        , C_ISO_8859_1},
 499         {"is_IS"        , C_ISO_8859_1},
 500         {"it_IT"        , C_ISO_8859_1},
 501         {"no_NO"        , C_ISO_8859_1},
 502         {"pt_PT"        , C_ISO_8859_1},
 503         {"pt_BR"        , C_ISO_8859_1},
 504         {"es_ES"        , C_ISO_8859_1},
 505         {"sv_SE"        , C_ISO_8859_1},
 506
 507         {"hr_HR"        , C_ISO_8859_2},
 508         {"hu_HU"        , C_ISO_8859_2},
 509         {"pl_PL"        , C_ISO_8859_2},
 510         {"ro_RO"        , C_ISO_8859_2},
 511         {"sk_SK"        , C_ISO_8859_2},
 512         {"sl_SI"        , C_ISO_8859_2},
 513         {"ru_RU"        , C_ISO_8859_5},
 514         {"el_GR"        , C_ISO_8859_7},
 515         {"iw_IL"        , C_ISO_8859_8},
 516         {"tr_TR"        , C_ISO_8859_9},
 517
 518         {"lt_LT.iso88594"       , C_ISO_8859_4},
 519         {"lt_LT.ISO8859-4"      , C_ISO_8859_4},
 520         {"lt_LT.ISO_8859-4"     , C_ISO_8859_4},
 521         {"lt_LT"                , C_ISO_8859_13},
 522         {"lv_LV"                , C_ISO_8859_13},
 523
 524         {"C"            , C_US_ASCII},
 525         {"POSIX"        , C_US_ASCII},
 526 };
 527 #endif /* !HAVE_LIBJCONV */
 528
 529 const gchar *conv_get_charset_str(CharSet charset)
 530 {
 531         gint i;
 532
 533         for (i = 0; i < sizeof(charsets) / sizeof(charsets[0]); i++) {
 534                 if (charsets[i].charset == charset)
 535                         return charsets[i].name;
 536         }
 537
 538         return NULL;
 539 }
 540
 541 CharSet conv_get_charset_from_str(const gchar *charset)
 542 {
 543         gint i;
 544
 545         if (!charset) return C_AUTO;
 546
 547         for (i = 0; i < sizeof(charsets) / sizeof(charsets[0]); i++) {
 548                 if (!strcasecmp(charsets[i].name, charset))
 549                         return charsets[i].charset;
 550         }
 551
 552         return C_AUTO;
 553 }
 554
 555 CharSet conv_get_current_charset(void)
 556 {
 557         static CharSet cur_charset = -1;
 558         gint i;
 559
 560 #if HAVE_LIBJCONV
 561         const gchar *cur_codeset;
 562 #else
 563         const gchar *cur_locale;
 564 #endif
 565
 566         if (cur_charset != -1)
 567                 return cur_charset;
 568
 569 #if HAVE_LIBJCONV
 570         cur_codeset = jconv_info_get_current_codeset();
 571         for (i = 0; i < sizeof(charsets) / sizeof(charsets[0]); i++) {
 572                 if (!strcasecmp(cur_codeset, charsets[i].name)) {
 573                         cur_charset = charsets[i].charset;
 574                         return cur_charset;
 575                 }
 576         }
 577 #else
 578         cur_locale = g_getenv("LC_ALL");
 579         if (!cur_locale) cur_locale = g_getenv("LC_CTYPE");
 580         if (!cur_locale) cur_locale = g_getenv("LANG");
 581         if (!cur_locale) cur_locale = setlocale(LC_CTYPE, NULL);
 582
 583         debug_print("current locale: %s\n",
 584                     cur_locale ? cur_locale : "(none)");
 585
 586         if (!cur_locale) {
 587                 cur_charset = C_US_ASCII;
 588                 return cur_charset;
 589         }
 590
 591         if (strcasestr(cur_locale, "UTF-8")) {
 592                 cur_charset = C_UTF_8;
 593                 return cur_charset;
 594         }
 595
 596         for (i = 0; i < sizeof(locale_table) / sizeof(locale_table[0]); i++) {
 597                 const gchar *p;
 598
 599                 /* "ja_JP.EUC" matches with "ja_JP.eucJP" and "ja_JP.EUC" */
 600                 /* "ja_JP" matches with "ja_JP.xxxx" and "ja" */
 601                 if (!strncasecmp(cur_locale, locale_table[i].locale,
 602                                  strlen(locale_table[i].locale))) {
 603                         cur_charset = locale_table[i].charset;
 604                         return cur_charset;
 605                 } else if ((p = strchr(locale_table[i].locale, '_')) &&
 606                          !strchr(p + 1, '.')) {
 607                         if (strlen(cur_locale) == 2 &&
 608                             !strncasecmp(cur_locale, locale_table[i].locale, 2)) {
 609                                 cur_charset = locale_table[i].charset;
 610                                 return cur_charset;
 611                         }
 612                 }
 613         }
 614 #endif
 615
 616         cur_charset = C_AUTO;
 617         return cur_charset;
 618 }
 619
 620 const gchar *conv_get_current_charset_str(void)
 621 {
 622 #if HAVE_LIBJCONV
 623         return jconv_info_get_current_codeset();
 624 #else
 625         static const gchar *codeset = NULL;
 626
 627         if (!codeset)
 628                 codeset = conv_get_charset_str(conv_get_current_charset());
 629
 630         return codeset ? codeset : "US-ASCII";
 631 #endif
 632 }
 633
 634 CharSet conv_get_outgoing_charset(void)
 635 {
 636         static CharSet out_charset = -1;
 637
 638 #if HAVE_LIBJCONV
 639         gint i, j, n_pref_codesets;
 640         const gchar *const *pref_codesets;
 641 #else
 642         CharSet cur_charset;
 643 #endif
 644
 645         if (out_charset != -1)
 646                 return out_charset;
 647
 648 #if HAVE_LIBJCONV
 649         /* skip US-ASCII and UTF-8 */
 650         pref_codesets = jconv_info_get_pref_codesets(&n_pref_codesets);
 651         for (i = 0; i < n_pref_codesets; i++) {
 652                 for (j = 2; j < sizeof(charsets) / sizeof(charsets[0]); j++) {
 653                         if (!strcasecmp(pref_codesets[i], charsets[j].name)) {
 654                                 out_charset = charsets[j].charset;
 655                                 return out_charset;
 656                         }
 657                 }
 658         }
 659
 660         for (i = 0; i < n_pref_codesets; i++) {
 661                 if (!strcasecmp(pref_codesets[i], "UTF-8")) {
 662                         out_charset = C_UTF_8;
 663                         return out_charset;
 664                 }
 665         }
 666
 667         out_charset = C_AUTO;
 668 #else
 669         cur_charset = conv_get_current_charset();
 670         switch (cur_charset) {
 671         case C_EUC_JP:
 672         case C_SHIFT_JIS:
 673                 out_charset = C_ISO_2022_JP;
 674                 break;
 675         default:
 676                 out_charset = cur_charset;
 677         }
 678 #endif
 679
 680         return out_charset;
 681 }
 682
 683 const gchar *conv_get_outgoing_charset_str(void)
 684 {
 685         CharSet out_charset;
 686         const gchar *str;
 687
 688         if (prefs_common.outgoing_charset) {
 689                 if (!isalpha(prefs_common.outgoing_charset[0])) {
 690                         g_free(prefs_common.outgoing_charset);
 691                         prefs_common.outgoing_charset = g_strdup(CS_AUTO);
 692                 } else if (strcmp(prefs_common.outgoing_charset, CS_AUTO) != 0)
 693                         return prefs_common.outgoing_charset;
 694         }
 695
 696         out_charset = conv_get_outgoing_charset();
 697         str = conv_get_charset_str(out_charset);
 698
 699         return str ? str : "US-ASCII";
 700 }
 701
 702 void conv_unmime_header_overwrite(gchar *str)
 703 {
 704         gchar *buf;
 705         gint outlen;
 706         CharSet cur_charset;
 707
 708         cur_charset = conv_get_current_charset();
 709
 710 #if HAVE_LIBJCONV
 711         Xstrdup_a(buf, str, {return;});
 712         outlen = strlen(str) + 1;
 713         UnMimeHeaderConv(buf, str, outlen);
 714         if (cur_charset == C_EUC_JP)
 715                 conv_unreadable_eucjp(str);
 716 #else
 717         if (cur_charset == C_EUC_JP) {
 718                 gchar *tmp;
 719                 gint len;
 720
 721                 Xstrdup_a(buf, str, {return;});
 722                 outlen = strlen(str) + 1;
 723                 UnMimeHeader(buf);
 724                 len = strlen(buf) * 2 + 1;
 725                 Xalloca(tmp, len, {strncpy2(str, buf, outlen); return;});
 726                 conv_jistodisp(tmp, len, buf);
 727                 strncpy2(str, tmp, outlen);
 728         } else
 729                 UnMimeHeader(str);
 730 #endif
 731 }
 732
 733 void conv_unmime_header(gchar *outbuf, gint outlen, const gchar *str,
 734                         const gchar *charset)
 735 {
 736         gchar *buf;
 737         CharSet cur_charset;
 738
 739         Xstrdup_a(buf, str, {return;});
 740         cur_charset = conv_get_current_charset();
 741
 742 #if HAVE_LIBJCONV
 743         UnMimeHeaderConv(buf, outbuf, outlen);
 744         if (cur_charset == C_EUC_JP)
 745                 conv_unreadable_eucjp(outbuf);
 746 #else
 747         UnMimeHeader(buf);
 748         if (cur_charset == C_EUC_JP) {
 749                 gchar *tmp;
 750                 gint len;
 751
 752                 len = strlen(buf) * 2 + 1;
 753                 Xalloca(tmp, len, {strncpy2(outbuf, buf, outlen); return;});
 754                 conv_jistodisp(tmp, len, buf);
 755                 strncpy2(outbuf, tmp, outlen);
 756         } else
 757                 strncpy2(outbuf, buf, outlen);
 758 #endif
 759 }
 760
 761 #define MAX_ENCLEN      75
 762 #define MAX_LINELEN     76
 763
 764 #define B64LEN(len)     ((len) / 3 * 4 + ((len) % 3 ? 4 : 0))
 765
 766 #if HAVE_LIBJCONV
 767 void conv_encode_header(gchar *dest, gint len, const gchar *src,
 768                         gint header_len)
 769 {
 770         wchar_t *wsrc;
 771         wchar_t *wsrcp;
 772         gchar *destp;
 773         size_t line_len, mimehdr_len, mimehdr_begin_len;
 774         gchar *mimehdr_init = "=?";
 775         gchar *mimehdr_end = "?=";
 776         gchar *mimehdr_enctype = "?B?";
 777         const gchar *mimehdr_charset;
 778
 779         /* g_print("src = %s\n", src); */
 780         mimehdr_charset = conv_get_outgoing_charset_str();
 781
 782         /* convert to wide-character string */
 783         wsrcp = wsrc = strdup_mbstowcs(src);
 784
 785         mimehdr_len = strlen(mimehdr_init) + strlen(mimehdr_end) +
 786                 strlen(mimehdr_charset) + strlen(mimehdr_enctype);
 787         mimehdr_begin_len = strlen(mimehdr_init) +
 788                 strlen(mimehdr_charset) + strlen(mimehdr_enctype);
 789         /* line_len = 1; */
 790         line_len = header_len;
 791         destp = dest;
 792         *dest = '\0';
 793
 794         g_return_if_fail(wsrc != NULL);
 795
 796         while (*wsrcp) {
 797                 wchar_t *wp, *wtmp, *wtmpp;
 798                 gint nspc = 0;
 799
 800                 /* irresponsible buffer overrun check */
 801                 if ((len - (destp - dest)) < (MAX_LINELEN + 1) * 2) break;
 802
 803                 /* encode string including space
 804                    if non-ASCII string follows */
 805                 if (is_next_nonascii(wsrcp)) {
 806                         wp = wsrcp;
 807                         while ((wp = find_wspace(wp)) != NULL)
 808                                 if (!is_next_nonascii(wp)) break;
 809                 } else
 810                         wp = find_wspace(wsrcp);
 811
 812                 if (wp != NULL) {
 813                         wtmp = wcsndup(wsrcp, wp - wsrcp);
 814                         wsrcp = wp + 1;
 815                         while (iswspace(wsrcp[nspc])) nspc++;
 816                 } else {
 817                         wtmp = wcsdup(wsrcp);
 818                         wsrcp += wcslen(wsrcp);
 819                 }
 820
 821                 wtmpp = wtmp;
 822
 823                 do {
 824                         gint tlen = 0, str_ascii = 1;
 825                         gchar *tmp; /* internal codeset */
 826                         gchar *raw; /* converted, but not base64 encoded */
 827                         register gchar *tmpp;
 828                         gint raw_len;
 829
 830                         tmpp = tmp = g_malloc(wcslen(wtmpp) * MB_CUR_MAX + 1);
 831                         *tmp = '\0';
 832                         raw = g_strdup("");
 833                         raw_len = 0;
 834
 835                         while (*wtmpp != (wchar_t)0) {
 836                                 gint mbl;
 837                                 gint dummy;
 838                                 gchar *raw_new = NULL;
 839                                 int raw_new_len = 0;
 840                                 const gchar *src_codeset;
 841
 842                                 if (*wtmpp < 32 || *wtmpp >= 127)
 843                                         str_ascii = 0;
 844                                 mbl = wctomb(tmpp, *wtmpp);
 845                                 if (mbl == -1) {
 846                                         g_warning("invalid wide character\n");
 847                                         wtmpp++;
 848                                         continue;
 849                                 }
 850                                 /* g_free(raw); */
 851                                 src_codeset = conv_get_current_charset_str();
 852                                 /* printf ("tmp = %s, tlen = %d, mbl\n",
 853                                         tmp, tlen, mbl); */
 854                                 if (jconv_alloc_conv(tmp, tlen + mbl,
 855                                                      &raw_new, &raw_new_len,
 856                                                      &src_codeset, 1,
 857                                                      &dummy, mimehdr_charset)
 858                                     != 0) {
 859                                         g_warning("can't convert\n");
 860                                         tmpp[0] = '\0';
 861                                         continue;
 862                                 }
 863                                 if (!str_ascii) {
 864                                         gint dlen = mimehdr_len +
 865                                                 B64LEN(raw_len);
 866                                         if ((line_len + dlen +
 867                                              (*(wtmpp + 1) ? 0 : nspc) +
 868                                              (line_len > 1 ? 1 : 0))
 869                                             > MAX_LINELEN) {
 870                                                 g_free(raw_new);
 871                                                 if (tlen == 0) {
 872                                                         *destp++ = '\n';
 873                                                         *destp++ = ' ';
 874                                                         line_len = 1;
 875                                                         str_ascii = 1;
 876                                                         continue;
 877                                                 } else {
 878                                                         *tmpp = '\0';
 879                                                         break;
 880                                                 }
 881                                         }
 882                                 } else if ((line_len + tlen + mbl +
 883                                             (*(wtmpp + 1) ? 0 : nspc) +
 884                                             (line_len > 1 ? 1 : 0))
 885                                            > MAX_LINELEN) {
 886                                         g_free(raw_new);
 887                                         if (1 + tlen + mbl +
 888                                             (*(wtmpp + 1) ? 0 : nspc)
 889                                             >= MAX_LINELEN) {
 890                                                 *tmpp = '\0';
 891                                                 break;
 892                                         }
 893                                         *destp++ = '\n';
 894                                         *destp++ = ' ';
 895                                         line_len = 1;
 896                                         continue;
 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 if ((line_len + tlen + mbl +
1077                                             (*(wtmpp + 1) ? 0 : nspc) +
1078                                             (line_len > 1 ? 1 : 0))
1079                                            > MAX_LINELEN) {
1080                                         if (1 + tlen + mbl +
1081                                             (*(wtmpp + 1) ? 0 : nspc)
1082                                             >= MAX_LINELEN) {
1083                                                 *tmpp = '\0';
1084                                                 break;
1085                                         }
1086                                         *destp++ = '\n';
1087                                         *destp++ = ' ';
1088                                         line_len = 1;
1089                                         continue;
1090                                 }
1091
1092                                 tmpp += mbl;
1093                                 *tmpp = '\0';
1094
1095                                 tlen += mbl;
1096                                 prev_mbl = mbl;
1097
1098                                 wtmpp++;
1099                         }
1100                         /* g_print("tmp = %s, tlen = %d, mb_seqlen = %d\n",
1101                                 tmp, tlen, mb_seqlen);
1102                         */
1103
1104                         if (tlen == 0) {
1105                                 g_free(tmp);
1106                                 continue;
1107                         }
1108
1109                         if (line_len > 1 && destp > dest) {
1110                                 *destp++ = ' ';
1111                                 *destp = '\0';
1112                                 line_len++;
1113                         }
1114
1115                         if (mb_seqlen) {
1116                                 gchar *tmp_jis;
1117
1118                                 tmp_jis = g_new(gchar, tlen + mb_seqlen + 1);
1119                                 conv_euctojis(tmp_jis,
1120                                               tlen + mb_seqlen + 1, tmp);
1121                                 g_snprintf(destp, len - strlen(dest), "%s%s%s",
1122                                            mimehdr_init, mimehdr_charset,
1123                                            mimehdr_enctype);
1124                                 destp += mimehdr_begin_len;
1125                                 line_len += mimehdr_begin_len;
1126
1127                                 to64frombits(destp, tmp_jis, strlen(tmp_jis));
1128                                 line_len += strlen(destp);
1129                                 destp += strlen(destp);
1130
1131                                 strcpy(destp, mimehdr_end);
1132                                 destp += strlen(mimehdr_end);
1133                                 line_len += strlen(mimehdr_end);
1134
1135                                 g_free(tmp_jis);
1136                         } else {
1137                                 strcpy(destp, tmp);
1138                                 line_len += strlen(destp);
1139                                 destp += strlen(destp);
1140                         }
1141
1142                         g_free(tmp);
1143                         /* g_print("line_len = %d\n\n", line_len); */
1144                 } while (*wtmpp != (wchar_t)0);
1145
1146                 while (iswspace(*wsrcp)) {
1147                         gint mbl;
1148
1149                         mbl = wctomb(destp, *wsrcp++);
1150                         if (mbl != -1) {
1151                                 destp += mbl;
1152                                 line_len += mbl;
1153                         }
1154                 }
1155                 *destp = '\0';
1156
1157                 g_free(wtmp);
1158         }
1159
1160         g_free(wsrc);
1161
1162         /* g_print("dest = %s\n", dest); */
1163 }
1164 #endif /* HAVE_LIBJCONV */