src/codeconv.c

   1 /*
   2  * Sylpheed -- a GTK+ based, lightweight, and fast e-mail client
   3  * Copyright (C) 1999-2002 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 "intl.h"
  43 #include "codeconv.h"
  44 #include "unmime.h"
  45 #include "common/base64.h"
  46 #include "utils.h"
  47 #include "prefs_common.h"
  48
  49 typedef enum
  50 {
  51         JIS_ASCII,
  52         JIS_KANJI,
  53         JIS_HWKANA,
  54         JIS_AUXKANJI
  55 } JISState;
  56
  57 #define SUBST_CHAR      '_'
  58 #define ESC             '\033'
  59
  60 #define iseuckanji(c) \
  61         (((c) & 0xff) >= 0xa1 && ((c) & 0xff) <= 0xfe)
  62 #define iseuchwkana1(c) \
  63         (((c) & 0xff) == 0x8e)
  64 #define iseuchwkana2(c) \
  65         (((c) & 0xff) >= 0xa1 && ((c) & 0xff) <= 0xdf)
  66 #define iseucaux(c) \
  67         (((c) & 0xff) == 0x8f)
  68 #define isunprintableeuckanji(c) \
  69         (((c) & 0xff) >= 0xa9 && ((c) & 0xff) <= 0xaf)
  70 #define issjiskanji1(c) \
  71         ((((c) & 0xff) >= 0x81 && ((c) & 0xff) <= 0x9f) || \
  72          (((c) & 0xff) >= 0xe0 && ((c) & 0xff) <= 0xfc))
  73 #define issjiskanji2(c) \
  74         ((((c) & 0xff) >= 0x40 && ((c) & 0xff) <= 0x7e) || \
  75          (((c) & 0xff) >= 0x80 && ((c) & 0xff) <= 0xfc))
  76 #define issjishwkana(c) \
  77         (((c) & 0xff) >= 0xa1 && ((c) & 0xff) <= 0xdf)
  78
  79 #define K_IN()                          \
  80         if (state != JIS_KANJI) {       \
  81                 *out++ = ESC;           \
  82                 *out++ = '$';           \
  83                 *out++ = 'B';           \
  84                 state = JIS_KANJI;      \
  85         }
  86
  87 #define K_OUT()                         \
  88         if (state != JIS_ASCII) {       \
  89                 *out++ = ESC;           \
  90                 *out++ = '(';           \
  91                 *out++ = 'B';           \
  92                 state = JIS_ASCII;      \
  93         }
  94
  95 #define HW_IN()                         \
  96         if (state != JIS_HWKANA) {      \
  97                 *out++ = ESC;           \
  98                 *out++ = '(';           \
  99                 *out++ = 'I';           \
 100                 state = JIS_HWKANA;     \
 101         }
 102
 103 #define AUX_IN()                        \
 104         if (state != JIS_AUXKANJI) {    \
 105                 *out++ = ESC;           \
 106                 *out++ = '$';           \
 107                 *out++ = '(';           \
 108                 *out++ = 'D';           \
 109                 state = JIS_AUXKANJI;   \
 110         }
 111
 112 void conv_jistoeuc(gchar *outbuf, gint outlen, const gchar *inbuf)
 113 {
 114         const guchar *in = inbuf;
 115         guchar *out = outbuf;
 116         JISState state = JIS_ASCII;
 117
 118         while (*in != '\0') {
 119                 if (*in == ESC) {
 120                         in++;
 121                         if (*in == '$') {
 122                                 if (*(in + 1) == '@' || *(in + 1) == 'B') {
 123                                         state = JIS_KANJI;
 124                                         in += 2;
 125                                 } else if (*(in + 1) == '(' &&
 126                                            *(in + 2) == 'D') {
 127                                         state = JIS_AUXKANJI;
 128                                         in += 3;
 129                                 } else {
 130                                         /* unknown escape sequence */
 131                                         state = JIS_ASCII;
 132                                 }
 133                         } else if (*in == '(') {
 134                                 if (*(in + 1) == 'B' || *(in + 1) == 'J') {
 135                                         state = JIS_ASCII;
 136                                         in += 2;
 137                                 } else if (*(in + 1) == 'I') {
 138                                         state = JIS_HWKANA;
 139                                         in += 2;
 140                                 } else {
 141                                         /* unknown escape sequence */
 142                                         state = JIS_ASCII;
 143                                 }
 144                         } else {
 145                                 /* unknown escape sequence */
 146                                 state = JIS_ASCII;
 147                         }
 148                 } else if (*in == 0x0e) {
 149                         state = JIS_HWKANA;
 150                         in++;
 151                 } else if (*in == 0x0f) {
 152                         state = JIS_ASCII;
 153                         in++;
 154                 } else {
 155                         switch (state) {
 156                         case JIS_ASCII:
 157                                 *out++ = *in++;
 158                                 break;
 159                         case JIS_KANJI:
 160                                 *out++ = *in++ | 0x80;
 161                                 if (*in == '\0') break;
 162                                 *out++ = *in++ | 0x80;
 163                                 break;
 164                         case JIS_HWKANA:
 165                                 *out++ = 0x8e;
 166                                 *out++ = *in++ | 0x80;
 167                                 break;
 168                         case JIS_AUXKANJI:
 169                                 *out++ = 0x8f;
 170                                 *out++ = *in++ | 0x80;
 171                                 if (*in == '\0') break;
 172                                 *out++ = *in++ | 0x80;
 173                                 break;
 174                         }
 175                 }
 176         }
 177
 178         *out = '\0';
 179 }
 180
 181 void conv_euctojis(gchar *outbuf, gint outlen, const gchar *inbuf)
 182 {
 183         const guchar *in = inbuf;
 184         guchar *out = outbuf;
 185         JISState state = JIS_ASCII;
 186
 187         while (*in != '\0') {
 188                 if (isascii(*in)) {
 189                         K_OUT();
 190                         *out++ = *in++;
 191                 } else if (iseuckanji(*in)) {
 192                         if (iseuckanji(*(in + 1))) {
 193                                 K_IN();
 194                                 *out++ = *in++ & 0x7f;
 195                                 *out++ = *in++ & 0x7f;
 196                         } else {
 197                                 K_OUT();
 198                                 *out++ = SUBST_CHAR;
 199                                 in++;
 200                                 if (*in != '\0' && !isascii(*in)) {
 201                                         *out++ = SUBST_CHAR;
 202                                         in++;
 203                                 }
 204                         }
 205                 } else if (iseuchwkana1(*in)) {
 206                         in++;
 207                         if (iseuchwkana2(*in)) {
 208                                 HW_IN();
 209                                 *out++ = *in++ & 0x7f;
 210                         } else {
 211                                 K_OUT();
 212                                 if (*in != '\0' && !isascii(*in)) {
 213                                         *out++ = SUBST_CHAR;
 214                                         in++;
 215                                 }
 216                         }
 217                 } else if (iseucaux(*in)) {
 218                         in++;
 219                         if (iseuckanji(*in) && iseuckanji(*(in + 1))) {
 220                                 AUX_IN();
 221                                 *out++ = *in++ & 0x7f;
 222                                 *out++ = *in++ & 0x7f;
 223                         } else {
 224                                 K_OUT();
 225                                 if (*in != '\0' && !isascii(*in)) {
 226                                         *out++ = SUBST_CHAR;
 227                                         in++;
 228                                         if (*in != '\0' && !isascii(*in)) {
 229                                                 *out++ = SUBST_CHAR;
 230                                                 in++;
 231                                         }
 232                                 }
 233                         }
 234                 } else {
 235                         K_OUT();
 236                         *out++ = SUBST_CHAR;
 237                         in++;
 238                 }
 239         }
 240
 241         K_OUT();
 242         *out = '\0';
 243 }
 244
 245 void conv_sjistoeuc(gchar *outbuf, gint outlen, const gchar *inbuf)
 246 {
 247         const guchar *in = inbuf;
 248         guchar *out = outbuf;
 249
 250         while (*in != '\0') {
 251                 if (isascii(*in)) {
 252                         *out++ = *in++;
 253                 } else if (issjiskanji1(*in)) {
 254                         if (issjiskanji2(*(in + 1))) {
 255                                 guchar out1 = *in;
 256                                 guchar out2 = *(in + 1);
 257                                 guchar row;
 258
 259                                 row = out1 < 0xa0 ? 0x70 : 0xb0;
 260                                 if (out2 < 0x9f) {
 261                                         out1 = (out1 - row) * 2 - 1;
 262                                         out2 -= out2 > 0x7f ? 0x20 : 0x1f;
 263                                 } else {
 264                                         out1 = (out1 - row) * 2;
 265                                         out2 -= 0x7e;
 266                                 }
 267
 268                                 *out++ = out1 | 0x80;
 269                                 *out++ = out2 | 0x80;
 270                                 in += 2;
 271                         } else {
 272                                 *out++ = SUBST_CHAR;
 273                                 in++;
 274                                 if (*in != '\0' && !isascii(*in)) {
 275                                         *out++ = SUBST_CHAR;
 276                                         in++;
 277                                 }
 278                         }
 279                 } else if (issjishwkana(*in)) {
 280                         *out++ = 0x8e;
 281                         *out++ = *in++;
 282                 } else {
 283                         *out++ = SUBST_CHAR;
 284                         in++;
 285                 }
 286         }
 287
 288         *out = '\0';
 289 }
 290
 291 void conv_anytoeuc(gchar *outbuf, gint outlen, const gchar *inbuf)
 292 {
 293         switch (conv_guess_encoding(inbuf)) {
 294         case C_ISO_2022_JP:
 295                 conv_jistoeuc(outbuf, outlen, inbuf);
 296                 break;
 297         case C_SHIFT_JIS:
 298                 conv_sjistoeuc(outbuf, outlen, inbuf);
 299                 break;
 300         default:
 301                 strncpy2(outbuf, inbuf, outlen);
 302                 break;
 303         }
 304 }
 305
 306 void conv_anytojis(gchar *outbuf, gint outlen, const gchar *inbuf)
 307 {
 308         switch (conv_guess_encoding(inbuf)) {
 309         case C_EUC_JP:
 310                 conv_euctojis(outbuf, outlen, inbuf);
 311                 break;
 312         default:
 313                 strncpy2(outbuf, inbuf, outlen);
 314                 break;
 315         }
 316 }
 317
 318 void conv_unreadable_eucjp(gchar *str)
 319 {
 320         register guchar *p = str;
 321
 322         while (*p != '\0') {
 323                 if (isascii(*p)) {
 324                         /* convert CR+LF -> LF */
 325                         if (*p == '\r' && *(p + 1) == '\n')
 326                                 memmove(p, p + 1, strlen(p));
 327                         /* printable 7 bit code */
 328                         p++;
 329                 } else if (iseuckanji(*p)) {
 330                         if (iseuckanji(*(p + 1)) && !isunprintableeuckanji(*p))
 331                                 /* printable euc-jp code */
 332                                 p += 2;
 333                         else {
 334                                 /* substitute unprintable code */
 335                                 *p++ = SUBST_CHAR;
 336                                 if (*p != '\0') {
 337                                         if (isascii(*p))
 338                                                 p++;
 339                                         else
 340                                                 *p++ = SUBST_CHAR;
 341                                 }
 342                         }
 343                 } else if (iseuchwkana1(*p)) {
 344                         if (iseuchwkana2(*(p + 1)))
 345                                 /* euc-jp hankaku kana */
 346                                 p += 2;
 347                         else
 348                                 *p++ = SUBST_CHAR;
 349                 } else if (iseucaux(*p)) {
 350                         if (iseuckanji(*(p + 1)) && iseuckanji(*(p + 2))) {
 351                                 /* auxiliary kanji */
 352                                 p += 3;
 353                         } else
 354                                 *p++ = SUBST_CHAR;
 355                 } else
 356                         /* substitute unprintable 1 byte code */
 357                         *p++ = SUBST_CHAR;
 358         }
 359 }
 360
 361 void conv_unreadable_8bit(gchar *str)
 362 {
 363         register guchar *p = str;
 364
 365         while (*p != '\0') {
 366                 /* convert CR+LF -> LF */
 367                 if (*p == '\r' && *(p + 1) == '\n')
 368                         memmove(p, p + 1, strlen(p));
 369                 else if (!isascii(*p)) *p = SUBST_CHAR;
 370                 p++;
 371         }
 372 }
 373
 374 void conv_unreadable_latin(gchar *str)
 375 {
 376         register guchar *p = str;
 377
 378         while (*p != '\0') {
 379                 /* convert CR+LF -> LF */
 380                 if (*p == '\r' && *(p + 1) == '\n')
 381                         memmove(p, p + 1, strlen(p));
 382                 else if ((*p & 0xff) >= 0x80 && (*p & 0xff) <= 0x9f)
 383                         *p = SUBST_CHAR;
 384                 p++;
 385         }
 386 }
 387
 388 #define NCV     '\0'
 389
 390 void conv_mb_alnum(gchar *str)
 391 {
 392         static guchar char_tbl[] = {
 393                 /* 0xa0 - 0xaf */
 394                 NCV, ' ', NCV, NCV, ',', '.', NCV, ':',
 395                 ';', '?', '!', NCV, NCV, NCV, NCV, NCV,
 396                 /* 0xb0 - 0xbf */
 397                 NCV, NCV, NCV, NCV, NCV, NCV, NCV, NCV,
 398                 NCV, NCV, NCV, NCV, NCV, NCV, NCV, NCV,
 399                 /* 0xc0 - 0xcf */
 400                 NCV, NCV, NCV, NCV, NCV, NCV, NCV, NCV,
 401                 NCV, NCV, '(', ')', NCV, NCV, '[', ']',
 402                 /* 0xd0 - 0xdf */
 403                 '{', '}', NCV, NCV, NCV, NCV, NCV, NCV,
 404                 NCV, NCV, NCV, NCV, '+', '-', NCV, NCV,
 405                 /* 0xe0 - 0xef */
 406                 NCV, '=', NCV, '<', '>', NCV, NCV, NCV,
 407                 NCV, NCV, NCV, NCV, NCV, NCV, NCV, NCV
 408         };
 409
 410         register guchar *p = str;
 411         register gint len;
 412
 413         len = strlen(str);
 414
 415         while (len > 1) {
 416                 if (*p == 0xa3) {
 417                         register guchar ch = *(p + 1);
 418
 419                         if (ch >= 0xb0 && ch <= 0xfa) {
 420                                 /* [a-zA-Z] */
 421                                 *p = ch & 0x7f;
 422                                 p++;
 423                                 len--;
 424                                 memmove(p, p + 1, len);
 425                                 len--;
 426                         } else  {
 427                                 p += 2;
 428                                 len -= 2;
 429                         }
 430                 } else if (*p == 0xa1) {
 431                         register guchar ch = *(p + 1);
 432
 433                         if (ch >= 0xa0 && ch <= 0xef &&
 434                             NCV != char_tbl[ch - 0xa0]) {
 435                                 *p = char_tbl[ch - 0xa0];
 436                                 p++;
 437                                 len--;
 438                                 memmove(p, p + 1, len);
 439                                 len--;
 440                         } else {
 441                                 p += 2;
 442                                 len -= 2;
 443                         }
 444                 } else if (iseuckanji(*p)) {
 445                         p += 2;
 446                         len -= 2;
 447                 } else {
 448                         p++;
 449                         len--;
 450                 }
 451         }
 452 }
 453
 454 CharSet conv_guess_encoding(const gchar *str)
 455 {
 456         const guchar *p = str;
 457         CharSet guessed = C_US_ASCII;
 458
 459         while (*p != '\0') {
 460                 if (*p == ESC && (*(p + 1) == '$' || *(p + 1) == '(')) {
 461                         if (guessed == C_US_ASCII)
 462                                 return C_ISO_2022_JP;
 463                         p += 2;
 464                 } else if (isascii(*p)) {
 465                         p++;
 466                 } else if (iseuckanji(*p) && iseuckanji(*(p + 1))) {
 467                         if (*p >= 0xfd && *p <= 0xfe)
 468                                 return C_EUC_JP;
 469                         else if (guessed == C_SHIFT_JIS) {
 470                                 if ((issjiskanji1(*p) &&
 471                                      issjiskanji2(*(p + 1))) ||
 472                                     issjishwkana(*p))
 473                                         guessed = C_SHIFT_JIS;
 474                                 else
 475                                         guessed = C_EUC_JP;
 476                         } else
 477                                 guessed = C_EUC_JP;
 478                         p += 2;
 479                 } else if (issjiskanji1(*p) && issjiskanji2(*(p + 1))) {
 480                         if (iseuchwkana1(*p) && iseuchwkana2(*(p + 1)))
 481                                 guessed = C_SHIFT_JIS;
 482                         else
 483                                 return C_SHIFT_JIS;
 484                         p += 2;
 485                 } else if (issjishwkana(*p)) {
 486                         guessed = C_SHIFT_JIS;
 487                         p++;
 488                 } else {
 489                         p++;
 490                 }
 491         }
 492
 493         return guessed;
 494 }
 495
 496 void conv_jistodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
 497 {
 498         conv_jistoeuc(outbuf, outlen, inbuf);
 499         conv_unreadable_eucjp(outbuf);
 500 }
 501
 502 void conv_sjistodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
 503 {
 504         conv_sjistoeuc(outbuf, outlen, inbuf);
 505         conv_unreadable_eucjp(outbuf);
 506 }
 507
 508 void conv_euctodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
 509 {
 510         strncpy2(outbuf, inbuf, outlen);
 511         conv_unreadable_eucjp(outbuf);
 512 }
 513
 514 void conv_anytodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
 515 {
 516         conv_anytoeuc(outbuf, outlen, inbuf);
 517         conv_unreadable_eucjp(outbuf);
 518 }
 519
 520 void conv_ustodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
 521 {
 522         strncpy2(outbuf, inbuf, outlen);
 523         conv_unreadable_8bit(outbuf);
 524 }
 525
 526 void conv_latintodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
 527 {
 528         strncpy2(outbuf, inbuf, outlen);
 529         conv_unreadable_latin(outbuf);
 530 }
 531
 532 void conv_noconv(gchar *outbuf, gint outlen, const gchar *inbuf)
 533 {
 534         strncpy2(outbuf, inbuf, outlen);
 535 }
 536
 537 CodeConverter *conv_code_converter_new(const gchar *charset)
 538 {
 539         CodeConverter *conv;
 540
 541         conv = g_new0(CodeConverter, 1);
 542 #if !HAVE_LIBJCONV
 543         conv->code_conv_func = conv_get_code_conv_func(charset);
 544 #endif
 545         conv->charset_str = g_strdup(charset);
 546         conv->charset = conv_get_charset_from_str(charset);
 547
 548         return conv;
 549 }
 550
 551 void conv_code_converter_destroy(CodeConverter *conv)
 552 {
 553         g_free(conv->charset_str);
 554         g_free(conv);
 555 }
 556
 557 gint conv_convert(CodeConverter *conv, gchar *outbuf, gint outlen,
 558                   const gchar *inbuf)
 559 {
 560 #if HAVE_LIBJCONV
 561         gchar *str;
 562
 563         str = conv_codeset_strdup(inbuf, conv->charset_str, NULL);
 564         if (!str)
 565                 return -1;
 566         else {
 567                 strncpy2(outbuf, str, outlen);
 568                 g_free(str);
 569         }
 570 #else /* !HAVE_LIBJCONV */
 571         conv->code_conv_func(outbuf, outlen, inbuf);
 572 #endif
 573
 574         return 0;
 575 }
 576
 577 gchar *conv_codeset_strdup(const gchar *inbuf,
 578                            const gchar *src_codeset, const gchar *dest_codeset)
 579 {
 580         gchar *buf;
 581         size_t len;
 582 #if HAVE_LIBJCONV
 583         gint actual_codeset;
 584         const gchar *const *codesets;
 585         gint n_codesets;
 586 #else /* !HAVE_LIBJCONV */
 587         CharSet src_charset = C_AUTO, dest_charset = C_AUTO;
 588 #endif
 589
 590         if (!dest_codeset) {
 591                 CodeConvFunc func;
 592
 593                 func = conv_get_code_conv_func(src_codeset);
 594                 if (func != conv_noconv) {
 595                         if (func == conv_jistodisp ||
 596                             func == conv_sjistodisp ||
 597                             func == conv_anytodisp)
 598                                 len = strlen(inbuf) * 2 + 1;
 599                         else
 600                                 len = strlen(inbuf) + 1;
 601                         buf = g_malloc(len);
 602                         if (!buf) return NULL;
 603                         func(buf, len, inbuf);
 604                         buf = g_realloc(buf, strlen(buf) + 1);
 605                         return buf;
 606                 }
 607         }
 608
 609         /* don't convert if src and dest codeset are identical */
 610         if (src_codeset && dest_codeset &&
 611             !strcasecmp(src_codeset, dest_codeset))
 612                 return g_strdup(inbuf);
 613
 614 #if HAVE_LIBJCONV
 615         if (src_codeset) {
 616                 codesets = &src_codeset;
 617                 n_codesets = 1;
 618         } else
 619                 codesets = jconv_info_get_pref_codesets(&n_codesets);
 620         if (!dest_codeset) {
 621                 dest_codeset = conv_get_current_charset_str();
 622                 /* don't convert if current codeset is US-ASCII */
 623                 if (!strcasecmp(dest_codeset, CS_US_ASCII))
 624                         return g_strdup(inbuf);
 625         }
 626
 627         if (jconv_alloc_conv(inbuf, strlen(inbuf), &buf, &len,
 628                              codesets, n_codesets,
 629                              &actual_codeset, dest_codeset)
 630             == 0)
 631                 return buf;
 632         else {
 633 #if 0
 634                 g_warning("code conversion from %s to %s failed\n",
 635                           codesets && codesets[0] ? codesets[0] : "(unknown)",
 636                           dest_codeset);
 637 #endif /* 0 */
 638                 return NULL;
 639         }
 640 #else /* !HAVE_LIBJCONV */
 641         if (src_codeset) {
 642                 if (!strcasecmp(src_codeset, CS_EUC_JP) ||
 643                     !strcasecmp(src_codeset, CS_EUCJP))
 644                         src_charset = C_EUC_JP;
 645                 else if (!strcasecmp(src_codeset, CS_SHIFT_JIS) ||
 646                          !strcasecmp(src_codeset, "SHIFT-JIS") ||
 647                          !strcasecmp(src_codeset, "SJIS"))
 648                         src_charset = C_SHIFT_JIS;
 649                 if (dest_codeset && !strcasecmp(dest_codeset, CS_ISO_2022_JP))
 650                         dest_charset = C_ISO_2022_JP;
 651         }
 652
 653         if ((src_charset == C_EUC_JP || src_charset == C_SHIFT_JIS) &&
 654             dest_charset == C_ISO_2022_JP) {
 655                 len = (strlen(inbuf) + 1) * 3;
 656                 buf = g_malloc(len);
 657                 if (buf) {
 658                         if (src_charset == C_EUC_JP)
 659                                 conv_euctojis(buf, len, inbuf);
 660                         else
 661                                 conv_anytojis(buf, len, inbuf);
 662                         buf = g_realloc(buf, strlen(buf) + 1);
 663                 }
 664         } else
 665                 buf = g_strdup(inbuf);
 666
 667         return buf;
 668 #endif /* !HAVE_LIBJCONV */
 669 }
 670
 671 CodeConvFunc conv_get_code_conv_func(const gchar *charset)
 672 {
 673         CodeConvFunc code_conv;
 674         CharSet cur_charset;
 675
 676         if (!charset) {
 677                 cur_charset = conv_get_current_charset();
 678                 if (cur_charset == C_EUC_JP || cur_charset == C_SHIFT_JIS)
 679                         return conv_anytodisp;
 680                 else
 681                         return conv_noconv;
 682         }
 683
 684         if (!strcasecmp(charset, CS_ISO_2022_JP) ||
 685             !strcasecmp(charset, CS_ISO_2022_JP_2))
 686                 code_conv = conv_jistodisp;
 687         else if (!strcasecmp(charset, CS_US_ASCII))
 688                 code_conv = conv_ustodisp;
 689         else if (!strncasecmp(charset, CS_ISO_8859_1, 10))
 690                 code_conv = conv_latintodisp;
 691 #if !HAVE_LIBJCONV
 692         else if (!strncasecmp(charset, "ISO-8859-", 9))
 693                 code_conv = conv_latintodisp;
 694 #endif
 695         else if (!strcasecmp(charset, CS_SHIFT_JIS) ||
 696                  !strcasecmp(charset, "SHIFT-JIS")  ||
 697                  !strcasecmp(charset, "SJIS")       ||
 698                  !strcasecmp(charset, "X-SJIS"))
 699                 code_conv = conv_sjistodisp;
 700         else if (!strcasecmp(charset, CS_EUC_JP) ||
 701                  !strcasecmp(charset, CS_EUCJP))
 702                 code_conv = conv_euctodisp;
 703         else
 704                 code_conv = conv_noconv;
 705
 706         return code_conv;
 707 }
 708
 709 static const struct {
 710         CharSet charset;
 711         gchar *const name;
 712 } charsets[] = {
 713         {C_US_ASCII,            CS_US_ASCII},
 714         {C_US_ASCII,            CS_ANSI_X3_4_1968},
 715         {C_UTF_8,               CS_UTF_8},
 716         {C_ISO_8859_1,          CS_ISO_8859_1},
 717         {C_ISO_8859_2,          CS_ISO_8859_2},
 718         {C_ISO_8859_4,          CS_ISO_8859_4},
 719         {C_ISO_8859_5,          CS_ISO_8859_5},
 720         {C_ISO_8859_7,          CS_ISO_8859_7},
 721         {C_ISO_8859_8,          CS_ISO_8859_8},
 722         {C_ISO_8859_9,          CS_ISO_8859_9},
 723         {C_ISO_8859_11,         CS_ISO_8859_11},
 724         {C_ISO_8859_13,         CS_ISO_8859_13},
 725         {C_ISO_8859_15,         CS_ISO_8859_15},
 726         {C_BALTIC,              CS_BALTIC},
 727         {C_CP1251,              CS_CP1251},
 728         {C_WINDOWS_1251,        CS_WINDOWS_1251},
 729         {C_KOI8_R,              CS_KOI8_R},
 730         {C_KOI8_U,              CS_KOI8_U},
 731         {C_ISO_2022_JP,         CS_ISO_2022_JP},
 732         {C_ISO_2022_JP_2,       CS_ISO_2022_JP_2},
 733         {C_EUC_JP,              CS_EUC_JP},
 734         {C_EUC_JP,              CS_EUCJP},
 735         {C_SHIFT_JIS,           CS_SHIFT_JIS},
 736         {C_ISO_2022_KR,         CS_ISO_2022_KR},
 737         {C_EUC_KR,              CS_EUC_KR},
 738         {C_ISO_2022_CN,         CS_ISO_2022_CN},
 739         {C_EUC_CN,              CS_EUC_CN},
 740         {C_GB2312,              CS_GB2312},
 741         {C_EUC_TW,              CS_EUC_TW},
 742         {C_BIG5,                CS_BIG5},
 743         {C_TIS_620,             CS_TIS_620},
 744         {C_WINDOWS_874,         CS_WINDOWS_874},
 745 };
 746
 747 #if !HAVE_LIBJCONV
 748 static const struct {
 749         gchar *const locale;
 750         CharSet charset;
 751         CharSet out_charset;
 752 } locale_table[] = {
 753         {"ja_JP.eucJP"  , C_EUC_JP      , C_ISO_2022_JP},
 754         {"ja_JP.ujis"   , C_EUC_JP      , C_ISO_2022_JP},
 755         {"ja_JP.EUC"    , C_EUC_JP      , C_ISO_2022_JP},
 756         {"ja_JP.SJIS"   , C_SHIFT_JIS   , C_ISO_2022_JP},
 757         {"ja_JP.JIS"    , C_ISO_2022_JP , C_ISO_2022_JP},
 758         {"ja_JP"        , C_EUC_JP      , C_ISO_2022_JP},
 759         {"ko_KR"        , C_EUC_KR      , C_EUC_KR},
 760         {"zh_CN.GB2312" , C_GB2312      , C_GB2312},
 761         {"zh_CN"        , C_GB2312      , C_GB2312},
 762         {"zh_TW.eucTW"  , C_EUC_TW      , C_BIG5},
 763         {"zh_TW.Big5"   , C_BIG5        , C_BIG5},
 764         {"zh_TW"        , C_BIG5        , C_BIG5},
 765
 766         {"ru_RU.KOI8-R" , C_KOI8_R      , C_ISO_8859_5},
 767         {"ru_RU.CP1251" , C_WINDOWS_1251, C_ISO_8859_5},
 768
 769         {"bg_BG"        , C_WINDOWS_1251, C_WINDOWS_1251},
 770
 771         {"en_US"        , C_ISO_8859_1  , C_ISO_8859_1},
 772         {"ca_ES"        , C_ISO_8859_1  , C_ISO_8859_1},
 773         {"da_DK"        , C_ISO_8859_1  , C_ISO_8859_1},
 774         {"de_DE"        , C_ISO_8859_1  , C_ISO_8859_1},
 775         {"nl_NL"        , C_ISO_8859_1  , C_ISO_8859_1},
 776         {"et_EE"        , C_ISO_8859_1  , C_ISO_8859_1},
 777         {"fi_FI"        , C_ISO_8859_1  , C_ISO_8859_1},
 778         {"fr_FR"        , C_ISO_8859_1  , C_ISO_8859_1},
 779         {"is_IS"        , C_ISO_8859_1  , C_ISO_8859_1},
 780         {"it_IT"        , C_ISO_8859_1  , C_ISO_8859_1},
 781         {"no_NO"        , C_ISO_8859_1  , C_ISO_8859_1},
 782         {"pt_PT"        , C_ISO_8859_1  , C_ISO_8859_1},
 783         {"pt_BR"        , C_ISO_8859_1  , C_ISO_8859_1},
 784         {"es_ES"        , C_ISO_8859_1  , C_ISO_8859_1},
 785         {"sv_SE"        , C_ISO_8859_1  , C_ISO_8859_1},
 786
 787         {"hr_HR"        , C_ISO_8859_2  , C_ISO_8859_2},
 788         {"hu_HU"        , C_ISO_8859_2  , C_ISO_8859_2},
 789         {"pl_PL"        , C_ISO_8859_2  , C_ISO_8859_2},
 790         {"ro_RO"        , C_ISO_8859_2  , C_ISO_8859_2},
 791         {"sk_SK"        , C_ISO_8859_2  , C_ISO_8859_2},
 792         {"sl_SI"        , C_ISO_8859_2  , C_ISO_8859_2},
 793         {"ru_RU"        , C_ISO_8859_5  , C_ISO_8859_5},
 794         {"el_GR"        , C_ISO_8859_7  , C_ISO_8859_7},
 795         {"iw_IL"        , C_ISO_8859_8  , C_ISO_8859_8},
 796         {"tr_TR"        , C_ISO_8859_9  , C_ISO_8859_9},
 797
 798         {"th_TH"        , C_TIS_620     , C_TIS_620},
 799         /* {"th_TH"     , C_WINDOWS_874}, */
 800         /* {"th_TH"     , C_ISO_8859_11}, */
 801
 802         {"lt_LT.iso88594"       , C_ISO_8859_4  , C_ISO_8859_4},
 803         {"lt_LT.ISO8859-4"      , C_ISO_8859_4  , C_ISO_8859_4},
 804         {"lt_LT.ISO_8859-4"     , C_ISO_8859_4  , C_ISO_8859_4},
 805         {"lt_LT"                , C_ISO_8859_13 , C_ISO_8859_13},
 806         {"lv_LV"                , C_ISO_8859_13 , C_ISO_8859_13},
 807
 808         {"C"                    , C_US_ASCII    , C_US_ASCII},
 809         {"POSIX"                , C_US_ASCII    , C_US_ASCII},
 810         {"ANSI_X3.4-1968"       , C_US_ASCII    , C_US_ASCII},
 811 };
 812 #endif /* !HAVE_LIBJCONV */
 813
 814 const gchar *conv_get_charset_str(CharSet charset)
 815 {
 816         gint i;
 817
 818         for (i = 0; i < sizeof(charsets) / sizeof(charsets[0]); i++) {
 819                 if (charsets[i].charset == charset)
 820                         return charsets[i].name;
 821         }
 822
 823         return NULL;
 824 }
 825
 826 CharSet conv_get_charset_from_str(const gchar *charset)
 827 {
 828         gint i;
 829
 830         if (!charset) return C_AUTO;
 831
 832         for (i = 0; i < sizeof(charsets) / sizeof(charsets[0]); i++) {
 833                 if (!strcasecmp(charsets[i].name, charset))
 834                         return charsets[i].charset;
 835         }
 836
 837         return C_AUTO;
 838 }
 839
 840 CharSet conv_get_current_charset(void)
 841 {
 842         static CharSet cur_charset = -1;
 843         gint i;
 844
 845 #if HAVE_LIBJCONV
 846         const gchar *cur_codeset;
 847 #else
 848         const gchar *cur_locale;
 849 #endif
 850
 851         if (cur_charset != -1)
 852                 return cur_charset;
 853
 854 #if HAVE_LIBJCONV
 855         cur_codeset = jconv_info_get_current_codeset();
 856         for (i = 0; i < sizeof(charsets) / sizeof(charsets[0]); i++) {
 857                 if (!strcasecmp(cur_codeset, charsets[i].name)) {
 858                         cur_charset = charsets[i].charset;
 859                         return cur_charset;
 860                 }
 861         }
 862 #else
 863         cur_locale = conv_get_current_locale();
 864         if (!cur_locale) {
 865                 cur_charset = C_US_ASCII;
 866                 return cur_charset;
 867         }
 868
 869         if (strcasestr(cur_locale, "UTF-8")) {
 870                 cur_charset = C_UTF_8;
 871                 return cur_charset;
 872         }
 873
 874         for (i = 0; i < sizeof(locale_table) / sizeof(locale_table[0]); i++) {
 875                 const gchar *p;
 876
 877                 /* "ja_JP.EUC" matches with "ja_JP.eucJP" and "ja_JP.EUC" */
 878                 /* "ja_JP" matches with "ja_JP.xxxx" and "ja" */
 879                 if (!strncasecmp(cur_locale, locale_table[i].locale,
 880                                  strlen(locale_table[i].locale))) {
 881                         cur_charset = locale_table[i].charset;
 882                         return cur_charset;
 883                 } else if ((p = strchr(locale_table[i].locale, '_')) &&
 884                          !strchr(p + 1, '.')) {
 885                         if (strlen(cur_locale) == 2 &&
 886                             !strncasecmp(cur_locale, locale_table[i].locale, 2)) {
 887                                 cur_charset = locale_table[i].charset;
 888                                 return cur_charset;
 889                         }
 890                 }
 891         }
 892 #endif
 893
 894         cur_charset = C_AUTO;
 895         return cur_charset;
 896 }
 897
 898 const gchar *conv_get_current_charset_str(void)
 899 {
 900         static const gchar *codeset = NULL;
 901
 902         if (!codeset)
 903                 codeset = conv_get_charset_str(conv_get_current_charset());
 904
 905         return codeset ? codeset : "US-ASCII";
 906 }
 907
 908 CharSet conv_get_outgoing_charset(void)
 909 {
 910         static CharSet out_charset = -1;
 911         gint i;
 912
 913 #if HAVE_LIBJCONV
 914         gint j, n_pref_codesets;
 915         const gchar *const *pref_codesets;
 916 #else
 917         const gchar *cur_locale;
 918 #endif
 919
 920         if (out_charset != -1)
 921                 return out_charset;
 922
 923 #if HAVE_LIBJCONV
 924         /* skip US-ASCII and UTF-8 */
 925         pref_codesets = jconv_info_get_pref_codesets(&n_pref_codesets);
 926         for (i = 0; i < n_pref_codesets; i++) {
 927                 for (j = 3; j < sizeof(charsets) / sizeof(charsets[0]); j++) {
 928                         if (!strcasecmp(pref_codesets[i], charsets[j].name)) {
 929                                 out_charset = charsets[j].charset;
 930                                 return out_charset;
 931                         }
 932                 }
 933         }
 934
 935         for (i = 0; i < n_pref_codesets; i++) {
 936                 if (!strcasecmp(pref_codesets[i], "UTF-8")) {
 937                         out_charset = C_UTF_8;
 938                         return out_charset;
 939                 }
 940         }
 941
 942         out_charset = C_AUTO;
 943 #else
 944         cur_locale = conv_get_current_locale();
 945         if (!cur_locale) {
 946                 out_charset = C_AUTO;
 947                 return out_charset;
 948         }
 949
 950         for (i = 0; i < sizeof(locale_table) / sizeof(locale_table[0]); i++) {
 951                 const gchar *p;
 952
 953                 if (!strncasecmp(cur_locale, locale_table[i].locale,
 954                                  strlen(locale_table[i].locale))) {
 955                         out_charset = locale_table[i].out_charset;
 956                         break;
 957                 } else if ((p = strchr(locale_table[i].locale, '_')) &&
 958                          !strchr(p + 1, '.')) {
 959                         if (strlen(cur_locale) == 2 &&
 960                             !strncasecmp(cur_locale, locale_table[i].locale, 2)) {
 961                                 out_charset = locale_table[i].out_charset;
 962                                 break;
 963                         }
 964                 }
 965         }
 966
 967         /* encoding conversion without libjconv is only supported
 968            on Japanese locale for now */
 969         if (out_charset == C_ISO_2022_JP)
 970                 return out_charset;
 971
 972         out_charset = conv_get_current_charset();
 973 #endif
 974
 975         return out_charset;
 976 }
 977
 978 const gchar *conv_get_outgoing_charset_str(void)
 979 {
 980         CharSet out_charset;
 981         const gchar *str;
 982
 983         if (prefs_common.outgoing_charset) {
 984                 if (!isalpha(prefs_common.outgoing_charset[0])) {
 985                         g_free(prefs_common.outgoing_charset);
 986                         prefs_common.outgoing_charset = g_strdup(CS_AUTO);
 987                 } else if (strcmp(prefs_common.outgoing_charset, CS_AUTO) != 0)
 988                         return prefs_common.outgoing_charset;
 989         }
 990
 991         out_charset = conv_get_outgoing_charset();
 992         str = conv_get_charset_str(out_charset);
 993
 994         return str ? str : "US-ASCII";
 995 }
 996
 997 const gchar *conv_get_current_locale(void)
 998 {
 999         gchar *cur_locale;
1000
1001         cur_locale = g_getenv("LC_ALL");
1002         if (!cur_locale) cur_locale = g_getenv("LC_CTYPE");
1003         if (!cur_locale) cur_locale = g_getenv("LANG");
1004         if (!cur_locale) cur_locale = setlocale(LC_CTYPE, NULL);
1005
1006         debug_print("current locale: %s\n",
1007                     cur_locale ? cur_locale : "(none)");
1008
1009         return cur_locale;
1010 }
1011
1012 void conv_unmime_header_overwrite(gchar *str)
1013 {
1014         gchar *buf;
1015         gint buflen;
1016         CharSet cur_charset;
1017
1018         cur_charset = conv_get_current_charset();
1019
1020         if (cur_charset == C_EUC_JP) {
1021                 buflen = strlen(str) * 2 + 1;
1022                 Xalloca(buf, buflen, return);
1023                 conv_anytodisp(buf, buflen, str);
1024                 unmime_header(str, buf);
1025         } else {
1026                 buflen = strlen(str) + 1;
1027                 Xalloca(buf, buflen, return);
1028                 unmime_header(buf, str);
1029                 strncpy2(str, buf, buflen);
1030         }
1031 }
1032
1033 void conv_unmime_header(gchar *outbuf, gint outlen, const gchar *str,
1034                         const gchar *charset)
1035 {
1036         CharSet cur_charset;
1037
1038         cur_charset = conv_get_current_charset();
1039
1040         if (cur_charset == C_EUC_JP) {
1041                 gchar *buf;
1042                 gint buflen;
1043
1044                 buflen = strlen(str) * 2 + 1;
1045                 Xalloca(buf, buflen, return);
1046                 conv_anytodisp(buf, buflen, str);
1047                 unmime_header(outbuf, buf);
1048         } else
1049                 unmime_header(outbuf, str);
1050 }
1051
1052 #define MAX_ENCLEN      75
1053 #define MAX_LINELEN     76
1054
1055 #define B64LEN(len)     ((len) / 3 * 4 + ((len) % 3 ? 4 : 0))
1056
1057 #if HAVE_LIBJCONV
1058 void conv_encode_header(gchar *dest, gint len, const gchar *src,
1059                         gint header_len)
1060 {
1061         wchar_t *wsrc;
1062         wchar_t *wsrcp;
1063         gchar *destp;
1064         size_t line_len, mimehdr_len, mimehdr_begin_len;
1065         gchar *mimehdr_init = "=?";
1066         gchar *mimehdr_end = "?=";
1067         gchar *mimehdr_enctype = "?B?";
1068         const gchar *mimehdr_charset;
1069
1070         /* g_print("src = %s\n", src); */
1071         mimehdr_charset = conv_get_outgoing_charset_str();
1072
1073         /* convert to wide-character string */
1074         wsrcp = wsrc = strdup_mbstowcs(src);
1075         if (!wsrc) {
1076                 g_warning("Can't convert string to wide characters.\n");
1077                 strncpy2(dest, src, len);
1078                 return;
1079         }
1080
1081         mimehdr_len = strlen(mimehdr_init) + strlen(mimehdr_end) +
1082                 strlen(mimehdr_charset) + strlen(mimehdr_enctype);
1083         mimehdr_begin_len = strlen(mimehdr_init) +
1084                 strlen(mimehdr_charset) + strlen(mimehdr_enctype);
1085         line_len = header_len;
1086         destp = dest;
1087         *dest = '\0';
1088
1089         while (*wsrcp) {
1090                 wchar_t *wp, *wtmp, *wtmpp;
1091                 gint nspc = 0;
1092                 gboolean str_is_non_ascii;
1093
1094                 /* irresponsible buffer overrun check */
1095                 if ((len - (destp - dest)) < (MAX_LINELEN + 1) * 2) break;
1096
1097                 /* encode string including space
1098                    if non-ASCII string follows */
1099                 if (is_next_nonascii(wsrcp)) {
1100                         wp = wsrcp;
1101                         while ((wp = find_wspace(wp)) != NULL)
1102                                 if (!is_next_nonascii(wp)) break;
1103                         str_is_non_ascii = TRUE;
1104                 } else {
1105                         wp = find_wspace(wsrcp);
1106                         str_is_non_ascii = FALSE;
1107                 }
1108
1109                 if (wp != NULL) {
1110                         wtmp = wcsndup(wsrcp, wp - wsrcp);
1111                         wsrcp = wp + 1;
1112                         while (iswspace(wsrcp[nspc])) nspc++;
1113                 } else {
1114                         wtmp = wcsdup(wsrcp);
1115                         wsrcp += wcslen(wsrcp);
1116                 }
1117
1118                 wtmpp = wtmp;
1119
1120                 do {
1121                         gint tlen = 0;
1122                         gchar *tmp; /* internal codeset */
1123                         gchar *raw; /* converted, but not base64 encoded */
1124                         register gchar *tmpp;
1125                         gint raw_len;
1126
1127                         tmpp = tmp = g_malloc(wcslen(wtmpp) * MB_CUR_MAX + 1);
1128                         *tmp = '\0';
1129                         raw = g_strdup("");
1130                         raw_len = 0;
1131
1132                         while (*wtmpp != (wchar_t)0) {
1133                                 gint mbl;
1134                                 gint dummy;
1135                                 gchar *raw_new = NULL;
1136                                 int raw_new_len = 0;
1137                                 const gchar *src_codeset;
1138
1139                                 mbl = wctomb(tmpp, *wtmpp);
1140                                 if (mbl == -1) {
1141                                         g_warning("invalid wide character\n");
1142                                         wtmpp++;
1143                                         continue;
1144                                 }
1145                                 /* g_free(raw); */
1146                                 src_codeset = conv_get_current_charset_str();
1147                                 /* printf ("tmp = %s, tlen = %d, mbl\n",
1148                                         tmp, tlen, mbl); */
1149                                 if (jconv_alloc_conv(tmp, tlen + mbl,
1150                                                      &raw_new, &raw_new_len,
1151                                                      &src_codeset, 1,
1152                                                      &dummy, mimehdr_charset)
1153                                     != 0) {
1154                                         g_warning("can't convert\n");
1155                                         tmpp[0] = '\0';
1156                                         wtmpp++;
1157                                         continue;
1158                                 }
1159                                 if (str_is_non_ascii) {
1160                                         gint dlen = mimehdr_len +
1161                                                 B64LEN(raw_len);
1162                                         if ((line_len + dlen +
1163                                              (*(wtmpp + 1) ? 0 : nspc) +
1164                                              (line_len > 1 ? 1 : 0))
1165                                             > MAX_LINELEN) {
1166                                                 g_free(raw_new);
1167                                                 if (tlen == 0) {
1168                                                         *destp++ = '\n';
1169                                                         *destp++ = ' ';
1170                                                         line_len = 1;
1171                                                         continue;
1172                                                 } else {
1173                                                         *tmpp = '\0';
1174                                                         break;
1175                                                 }
1176                                         }
1177                                 } else if ((line_len + tlen + mbl +
1178                                             (*(wtmpp + 1) ? 0 : nspc) +
1179                                             (line_len > 1 ? 1 : 0))
1180                                            > MAX_LINELEN) {
1181                                         g_free(raw_new);
1182                                         if (1 + tlen + mbl +
1183                                             (*(wtmpp + 1) ? 0 : nspc)
1184                                             >= MAX_LINELEN) {
1185                                                 *tmpp = '\0';
1186                                                 break;
1187                                         }
1188                                         *destp++ = '\n';
1189                                         *destp++ = ' ';
1190                                         line_len = 1;
1191                                         continue;
1192                                 }
1193
1194                                 tmpp += mbl;
1195                                 *tmpp = '\0';
1196
1197                                 tlen += mbl;
1198
1199                                 g_free(raw);
1200                                 raw = raw_new;
1201                                 raw_len = raw_new_len;
1202
1203                                 wtmpp++;
1204                         }
1205                         /* g_print("tmp = %s, tlen = %d, mb_seqlen = %d\n",
1206                                 tmp, tlen, mb_seqlen); */
1207
1208                         if (tlen == 0 || raw_len == 0) {
1209                                 g_free(tmp);
1210                                 g_free(raw);
1211                                 continue;
1212                         }
1213
1214                         if (line_len > 1 && destp > dest) {
1215                                 *destp++ = ' ';
1216                                 *destp = '\0';
1217                                 line_len++;
1218                         }
1219
1220                         if (str_is_non_ascii) {
1221                                 g_snprintf(destp, len - strlen(dest), "%s%s%s",
1222                                            mimehdr_init, mimehdr_charset,
1223                                            mimehdr_enctype);
1224                                 destp += mimehdr_begin_len;
1225                                 line_len += mimehdr_begin_len;
1226
1227                                 base64_encode(destp, raw, raw_len);
1228                                 line_len += strlen(destp);
1229                                 destp += strlen(destp);
1230
1231                                 strcpy(destp, mimehdr_end);
1232                                 destp += strlen(mimehdr_end);
1233                                 line_len += strlen(mimehdr_end);
1234                         } else {
1235                                 strcpy(destp, tmp);
1236                                 line_len += strlen(destp);
1237                                 destp += strlen(destp);
1238                         }
1239
1240                         g_free(tmp);
1241                         g_free(raw);
1242                         /* g_print("line_len = %d\n\n", line_len); */
1243                 } while (*wtmpp != (wchar_t)0);
1244
1245                 while (iswspace(*wsrcp)) {
1246                         gint mbl;
1247
1248                         mbl = wctomb(destp, *wsrcp++);
1249                         if (mbl != -1) {
1250                                 destp += mbl;
1251                                 line_len += mbl;
1252                         }
1253                 }
1254                 *destp = '\0';
1255
1256                 g_free(wtmp);
1257         }
1258
1259         g_free(wsrc);
1260
1261         /* g_print("dest = %s\n", dest); */
1262 }
1263 #else /* !HAVE_LIBJCONV */
1264
1265 #define JIS_SEQLEN      3
1266
1267 void conv_encode_header(gchar *dest, gint len, const gchar *src,
1268                         gint header_len)
1269 {
1270         wchar_t *wsrc;
1271         wchar_t *wsrcp;
1272         gchar *destp;
1273         size_t line_len, mimehdr_len, mimehdr_begin_len;
1274         gchar *mimehdr_init = "=?";
1275         gchar *mimehdr_end = "?=";
1276         gchar *mimehdr_enctype = "?B?";
1277         const gchar *mimehdr_charset;
1278         gboolean do_conv = FALSE;
1279
1280         /* g_print("src = %s\n", src); */
1281         mimehdr_charset = conv_get_outgoing_charset_str();
1282         if (strcmp(mimehdr_charset, "ISO-2022-JP") == 0)
1283                 do_conv = TRUE;
1284         else if (strcmp(mimehdr_charset, "US-ASCII") == 0)
1285                 mimehdr_charset = "ISO-8859-1";
1286
1287         /* convert to wide-character string */
1288         wsrcp = wsrc = strdup_mbstowcs(src);
1289         if (!wsrc) {
1290                 g_warning("Can't convert string to wide characters.\n");
1291                 strncpy2(dest, src, len);
1292                 return;
1293         }
1294
1295         mimehdr_len = strlen(mimehdr_init) + strlen(mimehdr_end) +
1296                       strlen(mimehdr_charset) + strlen(mimehdr_enctype);
1297         mimehdr_begin_len = strlen(mimehdr_init) +
1298                             strlen(mimehdr_charset) + strlen(mimehdr_enctype);
1299         line_len = header_len;
1300         destp = dest;
1301         *dest = '\0';
1302
1303         while (*wsrcp) {
1304                 wchar_t *wp, *wtmp, *wtmpp;
1305                 gint nspc = 0;
1306                 gboolean str_is_non_ascii;
1307
1308                 /* irresponsible buffer overrun check */
1309                 if ((len - (destp - dest)) < (MAX_LINELEN + 1) * 2) break;
1310
1311                 /* encode string including space
1312                    if non-ASCII string follows */
1313                 if (is_next_nonascii(wsrcp)) {
1314                         wp = wsrcp;
1315                         while ((wp = find_wspace(wp)) != NULL)
1316                                 if (!is_next_nonascii(wp)) break;
1317                         str_is_non_ascii = TRUE;
1318                 } else {
1319                         wp = find_wspace(wsrcp);
1320                         str_is_non_ascii = FALSE;
1321                 }
1322
1323                 if (wp != NULL) {
1324                         wtmp = wcsndup(wsrcp, wp - wsrcp);
1325                         wsrcp = wp + 1;
1326                         while (iswspace(wsrcp[nspc])) nspc++;
1327                 } else {
1328                         wtmp = wcsdup(wsrcp);
1329                         wsrcp += wcslen(wsrcp);
1330                 }
1331
1332                 wtmpp = wtmp;
1333
1334                 do {
1335                         gint prev_mbl = 1, tlen = 0, mb_seqlen = 0;
1336                         gchar *tmp;
1337                         register gchar *tmpp;
1338
1339                         tmpp = tmp = g_malloc(wcslen(wtmpp) * MB_CUR_MAX + 1);
1340                         *tmp = '\0';
1341
1342                         while (*wtmpp != (wchar_t)0) {
1343                                 gint mbl;
1344
1345                                 mbl = wctomb(tmpp, *wtmpp);
1346                                 if (mbl == -1) {
1347                                         g_warning("invalid wide character\n");
1348                                         wtmpp++;
1349                                         continue;
1350                                 }
1351
1352                                 /* length of KI + KO */
1353                                 if (do_conv && prev_mbl == 1 && mbl == 2)
1354                                         mb_seqlen += JIS_SEQLEN * 2;
1355
1356                                 if (str_is_non_ascii) {
1357                                         gint dlen = mimehdr_len +
1358                                                 B64LEN(tlen + mb_seqlen + mbl);
1359
1360                                         if ((line_len + dlen +
1361                                              (*(wtmpp + 1) ? 0 : nspc) +
1362                                              (line_len > 1 ? 1 : 0))
1363                                             > MAX_LINELEN) {
1364                                                 if (tlen == 0) {
1365                                                         *destp++ = '\n';
1366                                                         *destp++ = ' ';
1367                                                         line_len = 1;
1368                                                         mb_seqlen = 0;
1369                                                         continue;
1370                                                 } else {
1371                                                         *tmpp = '\0';
1372                                                         break;
1373                                                 }
1374                                         }
1375                                 } else if ((line_len + tlen + mbl +
1376                                             (*(wtmpp + 1) ? 0 : nspc) +
1377                                             (line_len > 1 ? 1 : 0))
1378                                            > MAX_LINELEN) {
1379                                         if (1 + tlen + mbl +
1380                                             (*(wtmpp + 1) ? 0 : nspc)
1381                                             >= MAX_LINELEN) {
1382                                                 *tmpp = '\0';
1383                                                 break;
1384                                         }
1385                                         *destp++ = '\n';
1386                                         *destp++ = ' ';
1387                                         line_len = 1;
1388                                         continue;
1389                                 }
1390
1391                                 tmpp += mbl;
1392                                 *tmpp = '\0';
1393
1394                                 tlen += mbl;
1395                                 prev_mbl = mbl;
1396
1397                                 wtmpp++;
1398                         }
1399                         /* g_print("tmp = %s, tlen = %d, mb_seqlen = %d\n",
1400                                 tmp, tlen, mb_seqlen); */
1401
1402                         if (tlen == 0) {
1403                                 g_free(tmp);
1404                                 continue;
1405                         }
1406
1407                         if (line_len > 1 && destp > dest) {
1408                                 *destp++ = ' ';
1409                                 *destp = '\0';
1410                                 line_len++;
1411                         }
1412
1413                         if (str_is_non_ascii) {
1414                                 gchar *raw;
1415
1416                                 raw = g_new(gchar, tlen + mb_seqlen + 1);
1417                                 if (do_conv)
1418                                         conv_euctojis(raw, tlen + mb_seqlen + 1,
1419                                                       tmp);
1420                                 else
1421                                         strcpy(raw, tmp);
1422                                 g_snprintf(destp, len - strlen(dest), "%s%s%s",
1423                                            mimehdr_init, mimehdr_charset,
1424                                            mimehdr_enctype);
1425                                 destp += mimehdr_begin_len;
1426                                 line_len += mimehdr_begin_len;
1427
1428                                 base64_encode(destp, raw, strlen(raw));
1429                                 line_len += strlen(destp);
1430                                 destp += strlen(destp);
1431
1432                                 strcpy(destp, mimehdr_end);
1433                                 destp += strlen(mimehdr_end);
1434                                 line_len += strlen(mimehdr_end);
1435
1436                                 g_free(raw);
1437                         } else {
1438                                 strcpy(destp, tmp);
1439                                 line_len += strlen(destp);
1440                                 destp += strlen(destp);
1441                         }
1442
1443                         g_free(tmp);
1444                         /* g_print("line_len = %d\n\n", line_len); */
1445                 } while (*wtmpp != (wchar_t)0);
1446
1447                 while (iswspace(*wsrcp)) {
1448                         gint mbl;
1449
1450                         mbl = wctomb(destp, *wsrcp++);
1451                         if (mbl != -1) {
1452                                 destp += mbl;
1453                                 line_len += mbl;
1454                         }
1455                 }
1456                 *destp = '\0';
1457
1458                 g_free(wtmp);
1459         }
1460
1461         g_free(wsrc);
1462
1463         /* g_print("dest = %s\n", dest); */
1464 }
1465 #endif /* HAVE_LIBJCONV */