src/codeconv.c

   1 /*
   2  * Sylpheed -- a GTK+ based, lightweight, and fast e-mail client
   3  * Copyright (C) 1999-2001 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         /* don't convert if src and dest codeset are identical */
 355         if (src_codeset && dest_codeset &&
 356             !strcasecmp(src_codeset, dest_codeset))
 357                 return g_strdup(inbuf);
 358
 359 #if HAVE_LIBJCONV
 360         if (src_codeset) {
 361                 codesets = &src_codeset;
 362                 n_codesets = 1;
 363         } else
 364                 codesets = jconv_info_get_pref_codesets(&n_codesets);
 365         if (!dest_codeset)
 366                 dest_codeset = conv_get_current_charset_str();
 367
 368         if (jconv_alloc_conv(inbuf, strlen(inbuf), &buf, &len,
 369                              codesets, n_codesets,
 370                              &actual_codeset, dest_codeset)
 371             == 0)
 372                 return buf;
 373         else
 374                 return NULL;
 375 #else /* !HAVE_LIBJCONV */
 376         if (src_codeset) {
 377                 if (!strcasecmp(src_codeset, CS_EUC_JP) ||
 378                     !strcasecmp(src_codeset, "EUCJP"))
 379                         src_charset = C_EUC_JP;
 380                 else if (!strcasecmp(src_codeset, CS_SHIFT_JIS) ||
 381                          !strcasecmp(src_codeset, "SHIFT-JIS") ||
 382                          !strcasecmp(src_codeset, "SJIS"))
 383                         src_charset = C_SHIFT_JIS;
 384                 if (dest_codeset && !strcasecmp(dest_codeset, CS_ISO_2022_JP))
 385                         dest_charset = C_ISO_2022_JP;
 386         }
 387
 388         if ((src_charset == C_EUC_JP || src_charset == C_SHIFT_JIS) &&
 389             dest_charset == C_ISO_2022_JP) {
 390                 len = (strlen(inbuf) + 1) * 3;
 391                 buf = g_malloc(len);
 392                 if (buf) {
 393                         if (src_charset == C_EUC_JP)
 394                                 conv_euctojis(buf, len, inbuf);
 395                         else
 396                                 conv_anytojis(buf, len, inbuf);
 397                         buf = g_realloc(buf, strlen(buf) + 1);
 398                 }
 399         } else
 400                 buf = g_strdup(inbuf);
 401
 402         return buf;
 403 #endif /* !HAVE_LIBJCONV */
 404 }
 405
 406 CodeConvFunc conv_get_code_conv_func(const gchar *charset)
 407 {
 408         CodeConvFunc code_conv;
 409
 410         if (!charset) {
 411                 if (conv_get_outgoing_charset() == C_ISO_2022_JP)
 412                         return conv_jistodisp;
 413                 else
 414                         return conv_noconv;
 415         }
 416
 417         if (!strcasecmp(charset, CS_ISO_2022_JP) ||
 418             !strcasecmp(charset, CS_ISO_2022_JP_2))
 419                 code_conv = conv_jistodisp;
 420         else if (!strcasecmp(charset, CS_US_ASCII))
 421                 code_conv = conv_ustodisp;
 422         else if (!strncasecmp(charset, CS_ISO_8859_1, 10))
 423                 code_conv = conv_latintodisp;
 424 #if !HAVE_LIBJCONV
 425         else if (!strncasecmp(charset, "ISO-8859-", 9))
 426                 code_conv = conv_latintodisp;
 427 #endif
 428         else if (!strcasecmp(charset, CS_SHIFT_JIS) ||
 429                  !strcasecmp(charset, "SHIFT-JIS")  ||
 430                  !strcasecmp(charset, "SJIS")       ||
 431                  !strcasecmp(charset, "X-SJIS"))
 432                 code_conv = conv_sjistodisp;
 433         else if (!strcasecmp(charset, CS_EUC_JP) ||
 434                  !strcasecmp(charset, "EUCJP"))
 435                 code_conv = conv_euctodisp;
 436         else
 437                 code_conv = conv_noconv;
 438
 439         return code_conv;
 440 }
 441
 442 static const struct {
 443         CharSet charset;
 444         gchar *const name;
 445 } charsets[] = {
 446         {C_US_ASCII,            CS_US_ASCII},
 447         {C_US_ASCII,            CS_ANSI_X3_4_1968},
 448         {C_UTF_8,               CS_UTF_8},
 449         {C_ISO_8859_1,          CS_ISO_8859_1},
 450         {C_ISO_8859_2,          CS_ISO_8859_2},
 451         {C_ISO_8859_4,          CS_ISO_8859_4},
 452         {C_ISO_8859_5,          CS_ISO_8859_5},
 453         {C_ISO_8859_7,          CS_ISO_8859_7},
 454         {C_ISO_8859_8,          CS_ISO_8859_8},
 455         {C_ISO_8859_9,          CS_ISO_8859_9},
 456         {C_ISO_8859_13,         CS_ISO_8859_13},
 457         {C_ISO_8859_15,         CS_ISO_8859_15},
 458         {C_BALTIC,              CS_BALTIC},
 459         {C_CP1251,              CS_CP1251},
 460         {C_WINDOWS_1251,        CS_WINDOWS_1251},
 461         {C_KOI8_R,              CS_KOI8_R},
 462         {C_KOI8_U,              CS_KOI8_U},
 463         {C_ISO_2022_JP,         CS_ISO_2022_JP},
 464         {C_ISO_2022_JP_2,       CS_ISO_2022_JP_2},
 465         {C_EUC_JP,              CS_EUC_JP},
 466         {C_SHIFT_JIS,           CS_SHIFT_JIS},
 467         {C_ISO_2022_KR,         CS_ISO_2022_KR},
 468         {C_EUC_KR,              CS_EUC_KR},
 469         {C_ISO_2022_CN,         CS_ISO_2022_CN},
 470         {C_EUC_CN,              CS_EUC_CN},
 471         {C_GB2312,              CS_GB2312},
 472         {C_EUC_TW,              CS_EUC_TW},
 473         {C_BIG5,                CS_BIG5},
 474 };
 475
 476 #if !HAVE_LIBJCONV
 477 static const struct {
 478         gchar *const locale;
 479         CharSet charset;
 480 } locale_table[] = {
 481         {"ja_JP.eucJP"  , C_EUC_JP},
 482         {"ja_JP.ujis"   , C_EUC_JP},
 483         {"ja_JP.EUC"    , C_EUC_JP},
 484         {"ja_JP.SJIS"   , C_SHIFT_JIS},
 485         {"ja_JP.JIS"    , C_ISO_2022_JP},
 486         {"ja_JP"        , C_EUC_JP},
 487         {"ko_KR"        , C_EUC_KR},
 488         {"zh_CN.GB2312" , C_GB2312},
 489         {"zh_CN"        , C_GB2312},
 490         {"zh_TW.eucTW"  , C_EUC_TW},
 491         {"zh_TW.Big5"   , C_BIG5},
 492         {"zh_TW"        , C_BIG5},
 493
 494         {"ru_RU.KOI8-R" , C_KOI8_R},
 495         {"ru_RU.CP1251" , C_CP1251},
 496
 497         {"en_US"        , C_ISO_8859_1},
 498         {"ca_ES"        , C_ISO_8859_1},
 499         {"da_DK"        , C_ISO_8859_1},
 500         {"de_DE"        , C_ISO_8859_1},
 501         {"nl_NL"        , C_ISO_8859_1},
 502         {"et_EE"        , C_ISO_8859_1},
 503         {"fi_FI"        , C_ISO_8859_1},
 504         {"fr_FR"        , C_ISO_8859_1},
 505         {"is_IS"        , C_ISO_8859_1},
 506         {"it_IT"        , C_ISO_8859_1},
 507         {"no_NO"        , C_ISO_8859_1},
 508         {"pt_PT"        , C_ISO_8859_1},
 509         {"pt_BR"        , C_ISO_8859_1},
 510         {"es_ES"        , C_ISO_8859_1},
 511         {"sv_SE"        , C_ISO_8859_1},
 512
 513         {"hr_HR"        , C_ISO_8859_2},
 514         {"hu_HU"        , C_ISO_8859_2},
 515         {"pl_PL"        , C_ISO_8859_2},
 516         {"ro_RO"        , C_ISO_8859_2},
 517         {"sk_SK"        , C_ISO_8859_2},
 518         {"sl_SI"        , C_ISO_8859_2},
 519         {"ru_RU"        , C_ISO_8859_5},
 520         {"el_GR"        , C_ISO_8859_7},
 521         {"iw_IL"        , C_ISO_8859_8},
 522         {"tr_TR"        , C_ISO_8859_9},
 523
 524         {"lt_LT.iso88594"       , C_ISO_8859_4},
 525         {"lt_LT.ISO8859-4"      , C_ISO_8859_4},
 526         {"lt_LT.ISO_8859-4"     , C_ISO_8859_4},
 527         {"lt_LT"                , C_ISO_8859_13},
 528         {"lv_LV"                , C_ISO_8859_13},
 529
 530         {"C"            , C_US_ASCII},
 531         {"POSIX"        , C_US_ASCII},
 532 };
 533 #endif /* !HAVE_LIBJCONV */
 534
 535 const gchar *conv_get_charset_str(CharSet charset)
 536 {
 537         gint i;
 538
 539         for (i = 0; i < sizeof(charsets) / sizeof(charsets[0]); i++) {
 540                 if (charsets[i].charset == charset)
 541                         return charsets[i].name;
 542         }
 543
 544         return NULL;
 545 }
 546
 547 CharSet conv_get_charset_from_str(const gchar *charset)
 548 {
 549         gint i;
 550
 551         if (!charset) return C_AUTO;
 552
 553         for (i = 0; i < sizeof(charsets) / sizeof(charsets[0]); i++) {
 554                 if (!strcasecmp(charsets[i].name, charset))
 555                         return charsets[i].charset;
 556         }
 557
 558         return C_AUTO;
 559 }
 560
 561 CharSet conv_get_current_charset(void)
 562 {
 563         static CharSet cur_charset = -1;
 564         gint i;
 565
 566 #if HAVE_LIBJCONV
 567         const gchar *cur_codeset;
 568 #else
 569         const gchar *cur_locale;
 570 #endif
 571
 572         if (cur_charset != -1)
 573                 return cur_charset;
 574
 575 #if HAVE_LIBJCONV
 576         cur_codeset = jconv_info_get_current_codeset();
 577         for (i = 0; i < sizeof(charsets) / sizeof(charsets[0]); i++) {
 578                 if (!strcasecmp(cur_codeset, charsets[i].name)) {
 579                         cur_charset = charsets[i].charset;
 580                         return cur_charset;
 581                 }
 582         }
 583 #else
 584         cur_locale = g_getenv("LC_ALL");
 585         if (!cur_locale) cur_locale = g_getenv("LC_CTYPE");
 586         if (!cur_locale) cur_locale = g_getenv("LANG");
 587         if (!cur_locale) cur_locale = setlocale(LC_CTYPE, NULL);
 588
 589         debug_print("current locale: %s\n",
 590                     cur_locale ? cur_locale : "(none)");
 591
 592         if (!cur_locale) {
 593                 cur_charset = C_US_ASCII;
 594                 return cur_charset;
 595         }
 596
 597         if (strcasestr(cur_locale, "UTF-8")) {
 598                 cur_charset = C_UTF_8;
 599                 return cur_charset;
 600         }
 601
 602         for (i = 0; i < sizeof(locale_table) / sizeof(locale_table[0]); i++) {
 603                 const gchar *p;
 604
 605                 /* "ja_JP.EUC" matches with "ja_JP.eucJP" and "ja_JP.EUC" */
 606                 /* "ja_JP" matches with "ja_JP.xxxx" and "ja" */
 607                 if (!strncasecmp(cur_locale, locale_table[i].locale,
 608                                  strlen(locale_table[i].locale))) {
 609                         cur_charset = locale_table[i].charset;
 610                         return cur_charset;
 611                 } else if ((p = strchr(locale_table[i].locale, '_')) &&
 612                          !strchr(p + 1, '.')) {
 613                         if (strlen(cur_locale) == 2 &&
 614                             !strncasecmp(cur_locale, locale_table[i].locale, 2)) {
 615                                 cur_charset = locale_table[i].charset;
 616                                 return cur_charset;
 617                         }
 618                 }
 619         }
 620 #endif
 621
 622         cur_charset = C_AUTO;
 623         return cur_charset;
 624 }
 625
 626 const gchar *conv_get_current_charset_str(void)
 627 {
 628         static const gchar *codeset = NULL;
 629
 630         if (!codeset)
 631                 codeset = conv_get_charset_str(conv_get_current_charset());
 632
 633         return codeset ? codeset : "US-ASCII";
 634 }
 635
 636 CharSet conv_get_outgoing_charset(void)
 637 {
 638         static CharSet out_charset = -1;
 639
 640 #if HAVE_LIBJCONV
 641         gint i, j, n_pref_codesets;
 642         const gchar *const *pref_codesets;
 643 #else
 644         CharSet cur_charset;
 645 #endif
 646
 647         if (out_charset != -1)
 648                 return out_charset;
 649
 650 #if HAVE_LIBJCONV
 651         /* skip US-ASCII and UTF-8 */
 652         pref_codesets = jconv_info_get_pref_codesets(&n_pref_codesets);
 653         for (i = 0; i < n_pref_codesets; i++) {
 654                 for (j = 2; j < sizeof(charsets) / sizeof(charsets[0]); j++) {
 655                         if (!strcasecmp(pref_codesets[i], charsets[j].name)) {
 656                                 out_charset = charsets[j].charset;
 657                                 return out_charset;
 658                         }
 659                 }
 660         }
 661
 662         for (i = 0; i < n_pref_codesets; i++) {
 663                 if (!strcasecmp(pref_codesets[i], "UTF-8")) {
 664                         out_charset = C_UTF_8;
 665                         return out_charset;
 666                 }
 667         }
 668
 669         out_charset = C_AUTO;
 670 #else
 671         cur_charset = conv_get_current_charset();
 672         switch (cur_charset) {
 673         case C_EUC_JP:
 674         case C_SHIFT_JIS:
 675                 out_charset = C_ISO_2022_JP;
 676                 break;
 677         default:
 678                 out_charset = cur_charset;
 679         }
 680 #endif
 681
 682         return out_charset;
 683 }
 684
 685 const gchar *conv_get_outgoing_charset_str(void)
 686 {
 687         CharSet out_charset;
 688         const gchar *str;
 689
 690         if (prefs_common.outgoing_charset) {
 691                 if (!isalpha(prefs_common.outgoing_charset[0])) {
 692                         g_free(prefs_common.outgoing_charset);
 693                         prefs_common.outgoing_charset = g_strdup(CS_AUTO);
 694                 } else if (strcmp(prefs_common.outgoing_charset, CS_AUTO) != 0)
 695                         return prefs_common.outgoing_charset;
 696         }
 697
 698         out_charset = conv_get_outgoing_charset();
 699         str = conv_get_charset_str(out_charset);
 700
 701         return str ? str : "US-ASCII";
 702 }
 703
 704 const gchar *conv_get_current_locale(void)
 705 {
 706         gchar *cur_locale;
 707
 708         cur_locale = g_getenv("LC_ALL");
 709         if (!cur_locale) cur_locale = g_getenv("LANG");
 710         if (!cur_locale) cur_locale = setlocale(LC_CTYPE, NULL);
 711
 712         debug_print("current locale: %s\n",
 713                     cur_locale ? cur_locale : "(none)");
 714
 715         return cur_locale;
 716 }
 717
 718 void conv_unmime_header_overwrite(gchar *str)
 719 {
 720         gchar *buf;
 721         gint outlen;
 722         CharSet cur_charset;
 723
 724         cur_charset = conv_get_current_charset();
 725
 726 #if HAVE_LIBJCONV
 727         Xstrdup_a(buf, str, return);
 728         outlen = strlen(str) + 1;
 729         UnMimeHeaderConv(buf, str, outlen);
 730         if (cur_charset == C_EUC_JP) {
 731                 gchar *tmp;
 732                 gint len;
 733
 734                 len = strlen(str) * 2 + 1;
 735                 Xalloca(tmp, len, return);
 736                 conv_jistodisp(tmp, len, str);
 737                 strncpy2(str, tmp, outlen);
 738         }
 739 #else
 740         if (cur_charset == C_EUC_JP) {
 741                 gchar *tmp;
 742                 gint len;
 743
 744                 Xstrdup_a(buf, str, return);
 745                 outlen = strlen(str) + 1;
 746                 UnMimeHeader(buf);
 747                 len = strlen(buf) * 2 + 1;
 748                 Xalloca(tmp, len, {strncpy2(str, buf, outlen); return;});
 749                 conv_jistodisp(tmp, len, buf);
 750                 strncpy2(str, tmp, outlen);
 751         } else
 752                 UnMimeHeader(str);
 753 #endif
 754 }
 755
 756 void conv_unmime_header(gchar *outbuf, gint outlen, const gchar *str,
 757                         const gchar *charset)
 758 {
 759         gchar *buf;
 760         CharSet cur_charset;
 761
 762         cur_charset = conv_get_current_charset();
 763         Xstrdup_a(buf, str, return);
 764
 765 #if HAVE_LIBJCONV
 766         UnMimeHeaderConv(buf, outbuf, outlen);
 767 #else
 768         UnMimeHeader(buf);
 769         strncpy2(outbuf, buf, outlen);
 770 #endif
 771         if (cur_charset == C_EUC_JP) {
 772                 gint len;
 773
 774                 len = strlen(outbuf) * 2 + 1;
 775                 Xalloca(buf, len, return);
 776                 conv_jistodisp(buf, len, outbuf);
 777                 strncpy2(outbuf, buf, outlen);
 778         }
 779 }
 780
 781 #define MAX_ENCLEN      75
 782 #define MAX_LINELEN     76
 783
 784 #define B64LEN(len)     ((len) / 3 * 4 + ((len) % 3 ? 4 : 0))
 785
 786 #if HAVE_LIBJCONV
 787 void conv_encode_header(gchar *dest, gint len, const gchar *src,
 788                         gint header_len)
 789 {
 790         wchar_t *wsrc;
 791         wchar_t *wsrcp;
 792         gchar *destp;
 793         size_t line_len, mimehdr_len, mimehdr_begin_len;
 794         gchar *mimehdr_init = "=?";
 795         gchar *mimehdr_end = "?=";
 796         gchar *mimehdr_enctype = "?B?";
 797         const gchar *mimehdr_charset;
 798
 799         /* g_print("src = %s\n", src); */
 800         mimehdr_charset = conv_get_outgoing_charset_str();
 801
 802         /* convert to wide-character string */
 803         wsrcp = wsrc = strdup_mbstowcs(src);
 804
 805         mimehdr_len = strlen(mimehdr_init) + strlen(mimehdr_end) +
 806                 strlen(mimehdr_charset) + strlen(mimehdr_enctype);
 807         mimehdr_begin_len = strlen(mimehdr_init) +
 808                 strlen(mimehdr_charset) + strlen(mimehdr_enctype);
 809         /* line_len = 1; */
 810         line_len = header_len;
 811         destp = dest;
 812         *dest = '\0';
 813
 814         g_return_if_fail(wsrc != NULL);
 815
 816         while (*wsrcp) {
 817                 wchar_t *wp, *wtmp, *wtmpp;
 818                 gint nspc = 0;
 819
 820                 /* irresponsible buffer overrun check */
 821                 if ((len - (destp - dest)) < (MAX_LINELEN + 1) * 2) break;
 822
 823                 /* encode string including space
 824                    if non-ASCII string follows */
 825                 if (is_next_nonascii(wsrcp)) {
 826                         wp = wsrcp;
 827                         while ((wp = find_wspace(wp)) != NULL)
 828                                 if (!is_next_nonascii(wp)) break;
 829                 } else
 830                         wp = find_wspace(wsrcp);
 831
 832                 if (wp != NULL) {
 833                         wtmp = wcsndup(wsrcp, wp - wsrcp);
 834                         wsrcp = wp + 1;
 835                         while (iswspace(wsrcp[nspc])) nspc++;
 836                 } else {
 837                         wtmp = wcsdup(wsrcp);
 838                         wsrcp += wcslen(wsrcp);
 839                 }
 840
 841                 wtmpp = wtmp;
 842
 843                 do {
 844                         gint tlen = 0, str_ascii = 1;
 845                         gchar *tmp; /* internal codeset */
 846                         gchar *raw; /* converted, but not base64 encoded */
 847                         register gchar *tmpp;
 848                         gint raw_len;
 849
 850                         tmpp = tmp = g_malloc(wcslen(wtmpp) * MB_CUR_MAX + 1);
 851                         *tmp = '\0';
 852                         raw = g_strdup("");
 853                         raw_len = 0;
 854
 855                         while (*wtmpp != (wchar_t)0) {
 856                                 gint mbl;
 857                                 gint dummy;
 858                                 gchar *raw_new = NULL;
 859                                 int raw_new_len = 0;
 860                                 const gchar *src_codeset;
 861
 862                                 if (*wtmpp < 32 || *wtmpp >= 127)
 863                                         str_ascii = 0;
 864                                 mbl = wctomb(tmpp, *wtmpp);
 865                                 if (mbl == -1) {
 866                                         g_warning("invalid wide character\n");
 867                                         wtmpp++;
 868                                         continue;
 869                                 }
 870                                 /* g_free(raw); */
 871                                 src_codeset = conv_get_current_charset_str();
 872                                 /* printf ("tmp = %s, tlen = %d, mbl\n",
 873                                         tmp, tlen, mbl); */
 874                                 if (jconv_alloc_conv(tmp, tlen + mbl,
 875                                                      &raw_new, &raw_new_len,
 876                                                      &src_codeset, 1,
 877                                                      &dummy, mimehdr_charset)
 878                                     != 0) {
 879                                         g_warning("can't convert\n");
 880                                         tmpp[0] = '\0';
 881                                         continue;
 882                                 }
 883                                 if (!str_ascii) {
 884                                         gint dlen = mimehdr_len +
 885                                                 B64LEN(raw_len);
 886                                         if ((line_len + dlen +
 887                                              (*(wtmpp + 1) ? 0 : nspc) +
 888                                              (line_len > 1 ? 1 : 0))
 889                                             > MAX_LINELEN) {
 890                                                 g_free(raw_new);
 891                                                 if (tlen == 0) {
 892                                                         *destp++ = '\n';
 893                                                         *destp++ = ' ';
 894                                                         line_len = 1;
 895                                                         str_ascii = 1;
 896                                                         continue;
 897                                                 } else {
 898                                                         *tmpp = '\0';
 899                                                         break;
 900                                                 }
 901                                         }
 902                                 } else if ((line_len + tlen + mbl +
 903                                             (*(wtmpp + 1) ? 0 : nspc) +
 904                                             (line_len > 1 ? 1 : 0))
 905                                            > MAX_LINELEN) {
 906                                         g_free(raw_new);
 907                                         if (1 + tlen + mbl +
 908                                             (*(wtmpp + 1) ? 0 : nspc)
 909                                             >= MAX_LINELEN) {
 910                                                 *tmpp = '\0';
 911                                                 break;
 912                                         }
 913                                         *destp++ = '\n';
 914                                         *destp++ = ' ';
 915                                         line_len = 1;
 916                                         continue;
 917                                 }
 918
 919                                 tmpp += mbl;
 920                                 *tmpp = '\0';
 921
 922                                 tlen += mbl;
 923
 924                                 g_free(raw);
 925                                 raw = raw_new;
 926                                 raw_len = raw_new_len;
 927
 928                                 wtmpp++;
 929                         }
 930                         /*
 931                           g_print("tmp = %s, tlen = %d, mb_seqlen = %d\n",
 932                                 tmp, tlen, mb_seqlen);
 933                         */
 934
 935                         if (tlen == 0 || raw_len == 0) {
 936                                 g_free(tmp);
 937                                 g_free(raw);
 938                                 continue;
 939                         }
 940
 941                         if (line_len > 1 && destp > dest) {
 942                                 *destp++ = ' ';
 943                                 *destp = '\0';
 944                                 line_len++;
 945                         }
 946
 947                         if (!str_ascii) {
 948                                 g_snprintf(destp, len - strlen(dest), "%s%s%s",
 949                                            mimehdr_init, mimehdr_charset,
 950                                            mimehdr_enctype);
 951                                 destp += mimehdr_begin_len;
 952                                 line_len += mimehdr_begin_len;
 953
 954                                 to64frombits(destp, raw, raw_len);
 955                                 line_len += strlen(destp);
 956                                 destp += strlen(destp);
 957
 958                                 strcpy(destp, mimehdr_end);
 959                                 destp += strlen(mimehdr_end);
 960                                 line_len += strlen(mimehdr_end);
 961                         } else {
 962                                 strcpy(destp, tmp);
 963                                 line_len += strlen(destp);
 964                                 destp += strlen(destp);
 965                         }
 966
 967                         g_free(tmp);
 968                         g_free(raw);
 969                         /* g_print("line_len = %d\n\n", line_len); */
 970                 } while (*wtmpp != (wchar_t)0);
 971
 972                 while (iswspace(*wsrcp)) {
 973                         gint mbl;
 974
 975                         mbl = wctomb(destp, *wsrcp++);
 976                         if (mbl != -1) {
 977                                 destp += mbl;
 978                                 line_len += mbl;
 979                         }
 980                 }
 981                 *destp = '\0';
 982
 983                 g_free(wtmp);
 984         }
 985
 986         g_free(wsrc);
 987
 988         /* g_print("dest = %s\n", dest);  */
 989 }
 990 #else /* !HAVE_LIBJCONV */
 991
 992 #define JIS_SEQLEN      3
 993
 994 void conv_encode_header(gchar *dest, gint len, const gchar *src,
 995                         gint header_len)
 996 {
 997         wchar_t *wsrc;
 998         wchar_t *wsrcp;
 999         gchar *destp;
1000         size_t line_len, mimehdr_len, mimehdr_begin_len;
1001         gchar *mimehdr_init = "=?";
1002         gchar *mimehdr_end = "?=";
1003         gchar *mimehdr_enctype = "?B?";
1004         const gchar *mimehdr_charset;
1005
1006         /* g_print("src = %s\n", src); */
1007         mimehdr_charset = conv_get_outgoing_charset_str();
1008         if (strcmp(mimehdr_charset, "ISO-2022-JP") != 0) {
1009                 /* currently only supports Japanese */
1010                 strncpy2(dest, src, len);
1011                 return;
1012         }
1013
1014         /* convert to wide-character string */
1015         wsrcp = wsrc = strdup_mbstowcs(src);
1016
1017         mimehdr_len = strlen(mimehdr_init) + strlen(mimehdr_end) +
1018                       strlen(mimehdr_charset) + strlen(mimehdr_enctype);
1019         mimehdr_begin_len = strlen(mimehdr_init) +
1020                             strlen(mimehdr_charset) + strlen(mimehdr_enctype);
1021         /* line_len = 1; */
1022         line_len = header_len;
1023         destp = dest;
1024         *dest = '\0';
1025
1026         g_return_if_fail(wsrc != NULL);
1027
1028         while (*wsrcp) {
1029                 wchar_t *wp, *wtmp, *wtmpp;
1030                 gint nspc = 0;
1031
1032                 /* irresponsible buffer overrun check */
1033                 if ((len - (destp - dest)) < (MAX_LINELEN + 1) * 2) break;
1034
1035                 /* encode string including space
1036                    if non-ASCII string follows */
1037                 if (is_next_nonascii(wsrcp)) {
1038                         wp = wsrcp;
1039                         while ((wp = find_wspace(wp)) != NULL)
1040                                 if (!is_next_nonascii(wp)) break;
1041                 } else
1042                         wp = find_wspace(wsrcp);
1043
1044                 if (wp != NULL) {
1045                         wtmp = wcsndup(wsrcp, wp - wsrcp);
1046                         wsrcp = wp + 1;
1047                         while (iswspace(wsrcp[nspc])) nspc++;
1048                 } else {
1049                         wtmp = wcsdup(wsrcp);
1050                         wsrcp += wcslen(wsrcp);
1051                 }
1052
1053                 wtmpp = wtmp;
1054
1055                 do {
1056                         gint prev_mbl = 1, tlen = 0, mb_seqlen = 0;
1057                         gchar *tmp;
1058                         register gchar *tmpp;
1059
1060                         tmpp = tmp = g_malloc(wcslen(wtmpp) * MB_CUR_MAX + 1);
1061                         *tmp = '\0';
1062
1063                         while (*wtmpp != (wchar_t)0) {
1064                                 gint mbl;
1065
1066                                 mbl = wctomb(tmpp, *wtmpp);
1067                                 if (mbl == -1) {
1068                                         g_warning("invalid wide character\n");
1069                                         wtmpp++;
1070                                         continue;
1071                                 }
1072
1073                                 /* length of KI + KO */
1074                                 if (prev_mbl == 1 && mbl == 2)
1075                                         mb_seqlen += JIS_SEQLEN * 2;
1076
1077                                 if (mb_seqlen) {
1078                                         gint dlen = mimehdr_len +
1079                                                 B64LEN(tlen + mb_seqlen + mbl);
1080
1081                                         if ((line_len + dlen +
1082                                              (*(wtmpp + 1) ? 0 : nspc) +
1083                                              (line_len > 1 ? 1 : 0))
1084                                             > MAX_LINELEN) {
1085                                                 if (tlen == 0) {
1086                                                         *destp++ = '\n';
1087                                                         *destp++ = ' ';
1088                                                         line_len = 1;
1089                                                         mb_seqlen = 0;
1090                                                         continue;
1091                                                 } else {
1092                                                         *tmpp = '\0';
1093                                                         break;
1094                                                 }
1095                                         }
1096                                 } else if ((line_len + tlen + mbl +
1097                                             (*(wtmpp + 1) ? 0 : nspc) +
1098                                             (line_len > 1 ? 1 : 0))
1099                                            > MAX_LINELEN) {
1100                                         if (1 + tlen + mbl +
1101                                             (*(wtmpp + 1) ? 0 : nspc)
1102                                             >= MAX_LINELEN) {
1103                                                 *tmpp = '\0';
1104                                                 break;
1105                                         }
1106                                         *destp++ = '\n';
1107                                         *destp++ = ' ';
1108                                         line_len = 1;
1109                                         continue;
1110                                 }
1111
1112                                 tmpp += mbl;
1113                                 *tmpp = '\0';
1114
1115                                 tlen += mbl;
1116                                 prev_mbl = mbl;
1117
1118                                 wtmpp++;
1119                         }
1120                         /* g_print("tmp = %s, tlen = %d, mb_seqlen = %d\n",
1121                                 tmp, tlen, mb_seqlen);
1122                         */
1123
1124                         if (tlen == 0) {
1125                                 g_free(tmp);
1126                                 continue;
1127                         }
1128
1129                         if (line_len > 1 && destp > dest) {
1130                                 *destp++ = ' ';
1131                                 *destp = '\0';
1132                                 line_len++;
1133                         }
1134
1135                         if (mb_seqlen) {
1136                                 gchar *tmp_jis;
1137
1138                                 tmp_jis = g_new(gchar, tlen + mb_seqlen + 1);
1139                                 conv_euctojis(tmp_jis,
1140                                               tlen + mb_seqlen + 1, tmp);
1141                                 g_snprintf(destp, len - strlen(dest), "%s%s%s",
1142                                            mimehdr_init, mimehdr_charset,
1143                                            mimehdr_enctype);
1144                                 destp += mimehdr_begin_len;
1145                                 line_len += mimehdr_begin_len;
1146
1147                                 to64frombits(destp, tmp_jis, strlen(tmp_jis));
1148                                 line_len += strlen(destp);
1149                                 destp += strlen(destp);
1150
1151                                 strcpy(destp, mimehdr_end);
1152                                 destp += strlen(mimehdr_end);
1153                                 line_len += strlen(mimehdr_end);
1154
1155                                 g_free(tmp_jis);
1156                         } else {
1157                                 strcpy(destp, tmp);
1158                                 line_len += strlen(destp);
1159                                 destp += strlen(destp);
1160                         }
1161
1162                         g_free(tmp);
1163                         /* g_print("line_len = %d\n\n", line_len); */
1164                 } while (*wtmpp != (wchar_t)0);
1165
1166                 while (iswspace(*wsrcp)) {
1167                         gint mbl;
1168
1169                         mbl = wctomb(destp, *wsrcp++);
1170                         if (mbl != -1) {
1171                                 destp += mbl;
1172                                 line_len += mbl;
1173                         }
1174                 }
1175                 *destp = '\0';
1176
1177                 g_free(wtmp);
1178         }
1179
1180         g_free(wsrc);
1181
1182         /* g_print("dest = %s\n", dest); */
1183 }
1184 #endif /* HAVE_LIBJCONV */