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