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 || func == conv_sjistodisp)
 592                                 len = strlen(inbuf) * 2 + 1;
 593                         else
 594                                 len = strlen(inbuf) + 1;
 595                         buf = g_malloc(len);
 596                         if (!buf) return NULL;
 597                         func(buf, len, inbuf);
 598                         buf = g_realloc(buf, strlen(buf) + 1);
 599                         return buf;
 600                 }
 601         }
 602
 603         /* don't convert if src and dest codeset are identical */
 604         if (src_codeset && dest_codeset &&
 605             !strcasecmp(src_codeset, dest_codeset))
 606                 return g_strdup(inbuf);
 607
 608 #if HAVE_LIBJCONV
 609         if (src_codeset) {
 610                 codesets = &src_codeset;
 611                 n_codesets = 1;
 612         } else
 613                 codesets = jconv_info_get_pref_codesets(&n_codesets);
 614         if (!dest_codeset) {
 615                 dest_codeset = conv_get_current_charset_str();
 616                 /* don't convert if current codeset is US-ASCII */
 617                 if (!strcasecmp(dest_codeset, CS_US_ASCII))
 618                         return g_strdup(inbuf);
 619         }
 620
 621         if (jconv_alloc_conv(inbuf, strlen(inbuf), &buf, &len,
 622                              codesets, n_codesets,
 623                              &actual_codeset, dest_codeset)
 624             == 0)
 625                 return buf;
 626         else {
 627                 g_warning("code conversion from %s to %s failed\n",
 628                           codesets && codesets[0] ? codesets[0] : "(unknown)",
 629                           dest_codeset);
 630                 return NULL;
 631         }
 632 #else /* !HAVE_LIBJCONV */
 633         if (src_codeset) {
 634                 if (!strcasecmp(src_codeset, CS_EUC_JP) ||
 635                     !strcasecmp(src_codeset, CS_EUCJP))
 636                         src_charset = C_EUC_JP;
 637                 else if (!strcasecmp(src_codeset, CS_SHIFT_JIS) ||
 638                          !strcasecmp(src_codeset, "SHIFT-JIS") ||
 639                          !strcasecmp(src_codeset, "SJIS"))
 640                         src_charset = C_SHIFT_JIS;
 641                 if (dest_codeset && !strcasecmp(dest_codeset, CS_ISO_2022_JP))
 642                         dest_charset = C_ISO_2022_JP;
 643         }
 644
 645         if ((src_charset == C_EUC_JP || src_charset == C_SHIFT_JIS) &&
 646             dest_charset == C_ISO_2022_JP) {
 647                 len = (strlen(inbuf) + 1) * 3;
 648                 buf = g_malloc(len);
 649                 if (buf) {
 650                         if (src_charset == C_EUC_JP)
 651                                 conv_euctojis(buf, len, inbuf);
 652                         else
 653                                 conv_anytojis(buf, len, inbuf);
 654                         buf = g_realloc(buf, strlen(buf) + 1);
 655                 }
 656         } else
 657                 buf = g_strdup(inbuf);
 658
 659         return buf;
 660 #endif /* !HAVE_LIBJCONV */
 661 }
 662
 663 CodeConvFunc conv_get_code_conv_func(const gchar *charset)
 664 {
 665         CodeConvFunc code_conv;
 666
 667         if (!charset) {
 668                 if (conv_get_outgoing_charset() == C_ISO_2022_JP)
 669                         return conv_jistodisp;
 670                 else
 671                         return conv_noconv;
 672         }
 673
 674         if (!strcasecmp(charset, CS_ISO_2022_JP) ||
 675             !strcasecmp(charset, CS_ISO_2022_JP_2))
 676                 code_conv = conv_jistodisp;
 677         else if (!strcasecmp(charset, CS_US_ASCII))
 678                 code_conv = conv_ustodisp;
 679         else if (!strncasecmp(charset, CS_ISO_8859_1, 10))
 680                 code_conv = conv_latintodisp;
 681 #if !HAVE_LIBJCONV
 682         else if (!strncasecmp(charset, "ISO-8859-", 9))
 683                 code_conv = conv_latintodisp;
 684 #endif
 685         else if (!strcasecmp(charset, CS_SHIFT_JIS) ||
 686                  !strcasecmp(charset, "SHIFT-JIS")  ||
 687                  !strcasecmp(charset, "SJIS")       ||
 688                  !strcasecmp(charset, "X-SJIS"))
 689                 code_conv = conv_sjistodisp;
 690         else if (!strcasecmp(charset, CS_EUC_JP) ||
 691                  !strcasecmp(charset, CS_EUCJP))
 692                 code_conv = conv_euctodisp;
 693         else
 694                 code_conv = conv_noconv;
 695
 696         return code_conv;
 697 }
 698
 699 static const struct {
 700         CharSet charset;
 701         gchar *const name;
 702 } charsets[] = {
 703         {C_US_ASCII,            CS_US_ASCII},
 704         {C_US_ASCII,            CS_ANSI_X3_4_1968},
 705         {C_UTF_8,               CS_UTF_8},
 706         {C_ISO_8859_1,          CS_ISO_8859_1},
 707         {C_ISO_8859_2,          CS_ISO_8859_2},
 708         {C_ISO_8859_4,          CS_ISO_8859_4},
 709         {C_ISO_8859_5,          CS_ISO_8859_5},
 710         {C_ISO_8859_7,          CS_ISO_8859_7},
 711         {C_ISO_8859_8,          CS_ISO_8859_8},
 712         {C_ISO_8859_9,          CS_ISO_8859_9},
 713         {C_ISO_8859_11,         CS_ISO_8859_11},
 714         {C_ISO_8859_13,         CS_ISO_8859_13},
 715         {C_ISO_8859_15,         CS_ISO_8859_15},
 716         {C_BALTIC,              CS_BALTIC},
 717         {C_CP1251,              CS_CP1251},
 718         {C_WINDOWS_1251,        CS_WINDOWS_1251},
 719         {C_KOI8_R,              CS_KOI8_R},
 720         {C_KOI8_U,              CS_KOI8_U},
 721         {C_ISO_2022_JP,         CS_ISO_2022_JP},
 722         {C_ISO_2022_JP_2,       CS_ISO_2022_JP_2},
 723         {C_EUC_JP,              CS_EUC_JP},
 724         {C_EUC_JP,              CS_EUCJP},
 725         {C_SHIFT_JIS,           CS_SHIFT_JIS},
 726         {C_ISO_2022_KR,         CS_ISO_2022_KR},
 727         {C_EUC_KR,              CS_EUC_KR},
 728         {C_ISO_2022_CN,         CS_ISO_2022_CN},
 729         {C_EUC_CN,              CS_EUC_CN},
 730         {C_GB2312,              CS_GB2312},
 731         {C_EUC_TW,              CS_EUC_TW},
 732         {C_BIG5,                CS_BIG5},
 733         {C_TIS_620,             CS_TIS_620},
 734         {C_WINDOWS_874,         CS_WINDOWS_874},
 735 };
 736
 737 #if !HAVE_LIBJCONV
 738 static const struct {
 739         gchar *const locale;
 740         CharSet charset;
 741 } locale_table[] = {
 742         {"ja_JP.eucJP"  , C_EUC_JP},
 743         {"ja_JP.ujis"   , C_EUC_JP},
 744         {"ja_JP.EUC"    , C_EUC_JP},
 745         {"ja_JP.SJIS"   , C_SHIFT_JIS},
 746         {"ja_JP.JIS"    , C_ISO_2022_JP},
 747         {"ja_JP"        , C_EUC_JP},
 748         {"ko_KR"        , C_EUC_KR},
 749         {"zh_CN.GB2312" , C_GB2312},
 750         {"zh_CN"        , C_GB2312},
 751         {"zh_TW.eucTW"  , C_EUC_TW},
 752         {"zh_TW.Big5"   , C_BIG5},
 753         {"zh_TW"        , C_BIG5},
 754
 755         {"ru_RU.KOI8-R" , C_KOI8_R},
 756         {"ru_RU.CP1251" , C_WINDOWS_1251},
 757
 758         {"en_US"        , C_ISO_8859_1},
 759         {"ca_ES"        , C_ISO_8859_1},
 760         {"da_DK"        , C_ISO_8859_1},
 761         {"de_DE"        , C_ISO_8859_1},
 762         {"nl_NL"        , C_ISO_8859_1},
 763         {"et_EE"        , C_ISO_8859_1},
 764         {"fi_FI"        , C_ISO_8859_1},
 765         {"fr_FR"        , C_ISO_8859_1},
 766         {"is_IS"        , C_ISO_8859_1},
 767         {"it_IT"        , C_ISO_8859_1},
 768         {"no_NO"        , C_ISO_8859_1},
 769         {"pt_PT"        , C_ISO_8859_1},
 770         {"pt_BR"        , C_ISO_8859_1},
 771         {"es_ES"        , C_ISO_8859_1},
 772         {"sv_SE"        , C_ISO_8859_1},
 773
 774         {"hr_HR"        , C_ISO_8859_2},
 775         {"hu_HU"        , C_ISO_8859_2},
 776         {"pl_PL"        , C_ISO_8859_2},
 777         {"ro_RO"        , C_ISO_8859_2},
 778         {"sk_SK"        , C_ISO_8859_2},
 779         {"sl_SI"        , C_ISO_8859_2},
 780         {"ru_RU"        , C_ISO_8859_5},
 781         {"el_GR"        , C_ISO_8859_7},
 782         {"iw_IL"        , C_ISO_8859_8},
 783         {"tr_TR"        , C_ISO_8859_9},
 784
 785         {"th_TH"        , C_TIS_620},
 786         /* {"th_TH"     , C_WINDOWS_874}, */
 787         /* {"th_TH"     , C_ISO_8859_11}, */
 788
 789         {"lt_LT.iso88594"       , C_ISO_8859_4},
 790         {"lt_LT.ISO8859-4"      , C_ISO_8859_4},
 791         {"lt_LT.ISO_8859-4"     , C_ISO_8859_4},
 792         {"lt_LT"                , C_ISO_8859_13},
 793         {"lv_LV"                , C_ISO_8859_13},
 794
 795         {"C"                    , C_US_ASCII},
 796         {"POSIX"                , C_US_ASCII},
 797         {"ANSI_X3.4-1968"       , C_US_ASCII},
 798 };
 799 #endif /* !HAVE_LIBJCONV */
 800
 801 const gchar *conv_get_charset_str(CharSet charset)
 802 {
 803         gint i;
 804
 805         for (i = 0; i < sizeof(charsets) / sizeof(charsets[0]); i++) {
 806                 if (charsets[i].charset == charset)
 807                         return charsets[i].name;
 808         }
 809
 810         return NULL;
 811 }
 812
 813 CharSet conv_get_charset_from_str(const gchar *charset)
 814 {
 815         gint i;
 816
 817         if (!charset) return C_AUTO;
 818
 819         for (i = 0; i < sizeof(charsets) / sizeof(charsets[0]); i++) {
 820                 if (!strcasecmp(charsets[i].name, charset))
 821                         return charsets[i].charset;
 822         }
 823
 824         return C_AUTO;
 825 }
 826
 827 CharSet conv_get_current_charset(void)
 828 {
 829         static CharSet cur_charset = -1;
 830         gint i;
 831
 832 #if HAVE_LIBJCONV
 833         const gchar *cur_codeset;
 834 #else
 835         const gchar *cur_locale;
 836 #endif
 837
 838         if (cur_charset != -1)
 839                 return cur_charset;
 840
 841 #if HAVE_LIBJCONV
 842         cur_codeset = jconv_info_get_current_codeset();
 843         for (i = 0; i < sizeof(charsets) / sizeof(charsets[0]); i++) {
 844                 if (!strcasecmp(cur_codeset, charsets[i].name)) {
 845                         cur_charset = charsets[i].charset;
 846                         return cur_charset;
 847                 }
 848         }
 849 #else
 850         cur_locale = g_getenv("LC_ALL");
 851         if (!cur_locale) cur_locale = g_getenv("LC_CTYPE");
 852         if (!cur_locale) cur_locale = g_getenv("LANG");
 853         if (!cur_locale) cur_locale = setlocale(LC_CTYPE, NULL);
 854
 855         debug_print("current locale: %s\n",
 856                     cur_locale ? cur_locale : "(none)");
 857
 858         if (!cur_locale) {
 859                 cur_charset = C_US_ASCII;
 860                 return cur_charset;
 861         }
 862
 863         if (strcasestr(cur_locale, "UTF-8")) {
 864                 cur_charset = C_UTF_8;
 865                 return cur_charset;
 866         }
 867
 868         for (i = 0; i < sizeof(locale_table) / sizeof(locale_table[0]); i++) {
 869                 const gchar *p;
 870
 871                 /* "ja_JP.EUC" matches with "ja_JP.eucJP" and "ja_JP.EUC" */
 872                 /* "ja_JP" matches with "ja_JP.xxxx" and "ja" */
 873                 if (!strncasecmp(cur_locale, locale_table[i].locale,
 874                                  strlen(locale_table[i].locale))) {
 875                         cur_charset = locale_table[i].charset;
 876                         return cur_charset;
 877                 } else if ((p = strchr(locale_table[i].locale, '_')) &&
 878                          !strchr(p + 1, '.')) {
 879                         if (strlen(cur_locale) == 2 &&
 880                             !strncasecmp(cur_locale, locale_table[i].locale, 2)) {
 881                                 cur_charset = locale_table[i].charset;
 882                                 return cur_charset;
 883                         }
 884                 }
 885         }
 886 #endif
 887
 888         cur_charset = C_AUTO;
 889         return cur_charset;
 890 }
 891
 892 const gchar *conv_get_current_charset_str(void)
 893 {
 894         static const gchar *codeset = NULL;
 895
 896         if (!codeset)
 897                 codeset = conv_get_charset_str(conv_get_current_charset());
 898
 899         return codeset ? codeset : "US-ASCII";
 900 }
 901
 902 CharSet conv_get_outgoing_charset(void)
 903 {
 904         static CharSet out_charset = -1;
 905
 906 #if HAVE_LIBJCONV
 907         gint i, j, n_pref_codesets;
 908         const gchar *const *pref_codesets;
 909 #else
 910         CharSet cur_charset;
 911 #endif
 912
 913         if (out_charset != -1)
 914                 return out_charset;
 915
 916 #if HAVE_LIBJCONV
 917         /* skip US-ASCII and UTF-8 */
 918         pref_codesets = jconv_info_get_pref_codesets(&n_pref_codesets);
 919         for (i = 0; i < n_pref_codesets; i++) {
 920                 for (j = 3; j < sizeof(charsets) / sizeof(charsets[0]); j++) {
 921                         if (!strcasecmp(pref_codesets[i], charsets[j].name)) {
 922                                 out_charset = charsets[j].charset;
 923                                 return out_charset;
 924                         }
 925                 }
 926         }
 927
 928         for (i = 0; i < n_pref_codesets; i++) {
 929                 if (!strcasecmp(pref_codesets[i], "UTF-8")) {
 930                         out_charset = C_UTF_8;
 931                         return out_charset;
 932                 }
 933         }
 934
 935         out_charset = C_AUTO;
 936 #else
 937         cur_charset = conv_get_current_charset();
 938         switch (cur_charset) {
 939         case C_EUC_JP:
 940         case C_SHIFT_JIS:
 941                 out_charset = C_ISO_2022_JP;
 942                 break;
 943         default:
 944                 out_charset = cur_charset;
 945         }
 946 #endif
 947
 948         return out_charset;
 949 }
 950
 951 const gchar *conv_get_outgoing_charset_str(void)
 952 {
 953         CharSet out_charset;
 954         const gchar *str;
 955
 956         if (prefs_common.outgoing_charset) {
 957                 if (!isalpha(prefs_common.outgoing_charset[0])) {
 958                         g_free(prefs_common.outgoing_charset);
 959                         prefs_common.outgoing_charset = g_strdup(CS_AUTO);
 960                 } else if (strcmp(prefs_common.outgoing_charset, CS_AUTO) != 0)
 961                         return prefs_common.outgoing_charset;
 962         }
 963
 964         out_charset = conv_get_outgoing_charset();
 965         str = conv_get_charset_str(out_charset);
 966
 967         return str ? str : "US-ASCII";
 968 }
 969
 970 const gchar *conv_get_current_locale(void)
 971 {
 972         gchar *cur_locale;
 973
 974         cur_locale = g_getenv("LC_ALL");
 975         if (!cur_locale) cur_locale = g_getenv("LANG");
 976         if (!cur_locale) cur_locale = setlocale(LC_CTYPE, NULL);
 977
 978         debug_print("current locale: %s\n",
 979                     cur_locale ? cur_locale : "(none)");
 980
 981         return cur_locale;
 982 }
 983
 984 void conv_unmime_header_overwrite(gchar *str)
 985 {
 986         gchar *buf;
 987         gint outlen;
 988         CharSet cur_charset;
 989
 990         cur_charset = conv_get_current_charset();
 991
 992         outlen = strlen(str) + 1;
 993         Xalloca(buf, outlen, return);
 994         unmime_header(buf, str);
 995         if (cur_charset == C_EUC_JP)
 996                 conv_jistodisp(str, outlen, buf);
 997         else
 998                 strncpy2(str, buf, outlen);
 999 }
1000
1001 void conv_unmime_header(gchar *outbuf, gint outlen, const gchar *str,
1002                         const gchar *charset)
1003 {
1004         gchar *buf;
1005         CharSet cur_charset;
1006
1007         cur_charset = conv_get_current_charset();
1008
1009         if (cur_charset == C_EUC_JP) {
1010                 Xalloca(buf, outlen, return);
1011                 unmime_header(buf, str);
1012                 conv_jistodisp(outbuf, outlen, buf);
1013         } else
1014                 unmime_header(outbuf, str);
1015 }
1016
1017 #define MAX_ENCLEN      75
1018 #define MAX_LINELEN     76
1019
1020 #define B64LEN(len)     ((len) / 3 * 4 + ((len) % 3 ? 4 : 0))
1021
1022 #if HAVE_LIBJCONV
1023 void conv_encode_header(gchar *dest, gint len, const gchar *src,
1024                         gint header_len)
1025 {
1026         wchar_t *wsrc;
1027         wchar_t *wsrcp;
1028         gchar *destp;
1029         size_t line_len, mimehdr_len, mimehdr_begin_len;
1030         gchar *mimehdr_init = "=?";
1031         gchar *mimehdr_end = "?=";
1032         gchar *mimehdr_enctype = "?B?";
1033         const gchar *mimehdr_charset;
1034
1035         /* g_print("src = %s\n", src); */
1036         mimehdr_charset = conv_get_outgoing_charset_str();
1037
1038         /* convert to wide-character string */
1039         wsrcp = wsrc = strdup_mbstowcs(src);
1040
1041         mimehdr_len = strlen(mimehdr_init) + strlen(mimehdr_end) +
1042                 strlen(mimehdr_charset) + strlen(mimehdr_enctype);
1043         mimehdr_begin_len = strlen(mimehdr_init) +
1044                 strlen(mimehdr_charset) + strlen(mimehdr_enctype);
1045         line_len = header_len;
1046         destp = dest;
1047         *dest = '\0';
1048
1049         g_return_if_fail(wsrc != NULL);
1050
1051         while (*wsrcp) {
1052                 wchar_t *wp, *wtmp, *wtmpp;
1053                 gint nspc = 0;
1054
1055                 /* irresponsible buffer overrun check */
1056                 if ((len - (destp - dest)) < (MAX_LINELEN + 1) * 2) break;
1057
1058                 /* encode string including space
1059                    if non-ASCII string follows */
1060                 if (is_next_nonascii(wsrcp)) {
1061                         wp = wsrcp;
1062                         while ((wp = find_wspace(wp)) != NULL)
1063                                 if (!is_next_nonascii(wp)) break;
1064                 } else
1065                         wp = find_wspace(wsrcp);
1066
1067                 if (wp != NULL) {
1068                         wtmp = wcsndup(wsrcp, wp - wsrcp);
1069                         wsrcp = wp + 1;
1070                         while (iswspace(wsrcp[nspc])) nspc++;
1071                 } else {
1072                         wtmp = wcsdup(wsrcp);
1073                         wsrcp += wcslen(wsrcp);
1074                 }
1075
1076                 wtmpp = wtmp;
1077
1078                 do {
1079                         gint tlen = 0, str_ascii = 1;
1080                         gchar *tmp; /* internal codeset */
1081                         gchar *raw; /* converted, but not base64 encoded */
1082                         register gchar *tmpp;
1083                         gint raw_len;
1084
1085                         tmpp = tmp = g_malloc(wcslen(wtmpp) * MB_CUR_MAX + 1);
1086                         *tmp = '\0';
1087                         raw = g_strdup("");
1088                         raw_len = 0;
1089
1090                         while (*wtmpp != (wchar_t)0) {
1091                                 gint mbl;
1092                                 gint dummy;
1093                                 gchar *raw_new = NULL;
1094                                 int raw_new_len = 0;
1095                                 const gchar *src_codeset;
1096
1097                                 if (*wtmpp < 32 || *wtmpp >= 127)
1098                                         str_ascii = 0;
1099                                 mbl = wctomb(tmpp, *wtmpp);
1100                                 if (mbl == -1) {
1101                                         g_warning("invalid wide character\n");
1102                                         wtmpp++;
1103                                         continue;
1104                                 }
1105                                 /* g_free(raw); */
1106                                 src_codeset = conv_get_current_charset_str();
1107                                 /* printf ("tmp = %s, tlen = %d, mbl\n",
1108                                         tmp, tlen, mbl); */
1109                                 if (jconv_alloc_conv(tmp, tlen + mbl,
1110                                                      &raw_new, &raw_new_len,
1111                                                      &src_codeset, 1,
1112                                                      &dummy, mimehdr_charset)
1113                                     != 0) {
1114                                         g_warning("can't convert\n");
1115                                         tmpp[0] = '\0';
1116                                         wtmpp++;
1117                                         continue;
1118                                 }
1119                                 if (!str_ascii) {
1120                                         gint dlen = mimehdr_len +
1121                                                 B64LEN(raw_len);
1122                                         if ((line_len + dlen +
1123                                              (*(wtmpp + 1) ? 0 : nspc) +
1124                                              (line_len > 1 ? 1 : 0))
1125                                             > MAX_LINELEN) {
1126                                                 g_free(raw_new);
1127                                                 if (tlen == 0) {
1128                                                         *destp++ = '\n';
1129                                                         *destp++ = ' ';
1130                                                         line_len = 1;
1131                                                         str_ascii = 1;
1132                                                         continue;
1133                                                 } else {
1134                                                         *tmpp = '\0';
1135                                                         break;
1136                                                 }
1137                                         }
1138                                 } else if ((line_len + tlen + mbl +
1139                                             (*(wtmpp + 1) ? 0 : nspc) +
1140                                             (line_len > 1 ? 1 : 0))
1141                                            > MAX_LINELEN) {
1142                                         g_free(raw_new);
1143                                         if (1 + tlen + mbl +
1144                                             (*(wtmpp + 1) ? 0 : nspc)
1145                                             >= MAX_LINELEN) {
1146                                                 *tmpp = '\0';
1147                                                 break;
1148                                         }
1149                                         *destp++ = '\n';
1150                                         *destp++ = ' ';
1151                                         line_len = 1;
1152                                         continue;
1153                                 }
1154
1155                                 tmpp += mbl;
1156                                 *tmpp = '\0';
1157
1158                                 tlen += mbl;
1159
1160                                 g_free(raw);
1161                                 raw = raw_new;
1162                                 raw_len = raw_new_len;
1163
1164                                 wtmpp++;
1165                         }
1166                         /* g_print("tmp = %s, tlen = %d, mb_seqlen = %d\n",
1167                                 tmp, tlen, mb_seqlen); */
1168
1169                         if (tlen == 0 || raw_len == 0) {
1170                                 g_free(tmp);
1171                                 g_free(raw);
1172                                 continue;
1173                         }
1174
1175                         if (line_len > 1 && destp > dest) {
1176                                 *destp++ = ' ';
1177                                 *destp = '\0';
1178                                 line_len++;
1179                         }
1180
1181                         if (!str_ascii) {
1182                                 g_snprintf(destp, len - strlen(dest), "%s%s%s",
1183                                            mimehdr_init, mimehdr_charset,
1184                                            mimehdr_enctype);
1185                                 destp += mimehdr_begin_len;
1186                                 line_len += mimehdr_begin_len;
1187
1188                                 base64_encode(destp, raw, raw_len);
1189                                 line_len += strlen(destp);
1190                                 destp += strlen(destp);
1191
1192                                 strcpy(destp, mimehdr_end);
1193                                 destp += strlen(mimehdr_end);
1194                                 line_len += strlen(mimehdr_end);
1195                         } else {
1196                                 strcpy(destp, tmp);
1197                                 line_len += strlen(destp);
1198                                 destp += strlen(destp);
1199                         }
1200
1201                         g_free(tmp);
1202                         g_free(raw);
1203                         /* g_print("line_len = %d\n\n", line_len); */
1204                 } while (*wtmpp != (wchar_t)0);
1205
1206                 while (iswspace(*wsrcp)) {
1207                         gint mbl;
1208
1209                         mbl = wctomb(destp, *wsrcp++);
1210                         if (mbl != -1) {
1211                                 destp += mbl;
1212                                 line_len += mbl;
1213                         }
1214                 }
1215                 *destp = '\0';
1216
1217                 g_free(wtmp);
1218         }
1219
1220         g_free(wsrc);
1221
1222         /* g_print("dest = %s\n", dest); */
1223 }
1224 #else /* !HAVE_LIBJCONV */
1225
1226 #define JIS_SEQLEN      3
1227
1228 void conv_encode_header(gchar *dest, gint len, const gchar *src,
1229                         gint header_len)
1230 {
1231         wchar_t *wsrc;
1232         wchar_t *wsrcp;
1233         gchar *destp;
1234         size_t line_len, mimehdr_len, mimehdr_begin_len;
1235         gchar *mimehdr_init = "=?";
1236         gchar *mimehdr_end = "?=";
1237         gchar *mimehdr_enctype = "?B?";
1238         const gchar *mimehdr_charset;
1239
1240         /* g_print("src = %s\n", src); */
1241         mimehdr_charset = conv_get_outgoing_charset_str();
1242         if (strcmp(mimehdr_charset, "ISO-2022-JP") != 0) {
1243                 /* currently only supports Japanese */
1244                 strncpy2(dest, src, len);
1245                 return;
1246         }
1247
1248         /* convert to wide-character string */
1249         wsrcp = wsrc = strdup_mbstowcs(src);
1250
1251         mimehdr_len = strlen(mimehdr_init) + strlen(mimehdr_end) +
1252                       strlen(mimehdr_charset) + strlen(mimehdr_enctype);
1253         mimehdr_begin_len = strlen(mimehdr_init) +
1254                             strlen(mimehdr_charset) + strlen(mimehdr_enctype);
1255         line_len = header_len;
1256         destp = dest;
1257         *dest = '\0';
1258
1259         g_return_if_fail(wsrc != NULL);
1260
1261         while (*wsrcp) {
1262                 wchar_t *wp, *wtmp, *wtmpp;
1263                 gint nspc = 0;
1264                 gboolean str_is_non_ascii;
1265
1266                 /* irresponsible buffer overrun check */
1267                 if ((len - (destp - dest)) < (MAX_LINELEN + 1) * 2) break;
1268
1269                 /* encode string including space
1270                    if non-ASCII string follows */
1271                 if (is_next_nonascii(wsrcp)) {
1272                         wp = wsrcp;
1273                         while ((wp = find_wspace(wp)) != NULL)
1274                                 if (!is_next_nonascii(wp)) break;
1275                         str_is_non_ascii = TRUE;
1276                 } else {
1277                         wp = find_wspace(wsrcp);
1278                         str_is_non_ascii = FALSE;
1279                 }
1280
1281                 if (wp != NULL) {
1282                         wtmp = wcsndup(wsrcp, wp - wsrcp);
1283                         wsrcp = wp + 1;
1284                         while (iswspace(wsrcp[nspc])) nspc++;
1285                 } else {
1286                         wtmp = wcsdup(wsrcp);
1287                         wsrcp += wcslen(wsrcp);
1288                 }
1289
1290                 wtmpp = wtmp;
1291
1292                 do {
1293                         gint prev_mbl = 1, tlen = 0, mb_seqlen = 0;
1294                         gchar *tmp;
1295                         register gchar *tmpp;
1296
1297                         tmpp = tmp = g_malloc(wcslen(wtmpp) * MB_CUR_MAX + 1);
1298                         *tmp = '\0';
1299
1300                         while (*wtmpp != (wchar_t)0) {
1301                                 gint mbl;
1302
1303                                 mbl = wctomb(tmpp, *wtmpp);
1304                                 if (mbl == -1) {
1305                                         g_warning("invalid wide character\n");
1306                                         wtmpp++;
1307                                         continue;
1308                                 }
1309
1310                                 /* length of KI + KO */
1311                                 if (prev_mbl == 1 && mbl == 2)
1312                                         mb_seqlen += JIS_SEQLEN * 2;
1313
1314                                 if (str_is_non_ascii) {
1315                                         gint dlen = mimehdr_len +
1316                                                 B64LEN(tlen + mb_seqlen + mbl);
1317
1318                                         if ((line_len + dlen +
1319                                              (*(wtmpp + 1) ? 0 : nspc) +
1320                                              (line_len > 1 ? 1 : 0))
1321                                             > MAX_LINELEN) {
1322                                                 if (tlen == 0) {
1323                                                         *destp++ = '\n';
1324                                                         *destp++ = ' ';
1325                                                         line_len = 1;
1326                                                         mb_seqlen = 0;
1327                                                         continue;
1328                                                 } else {
1329                                                         *tmpp = '\0';
1330                                                         break;
1331                                                 }
1332                                         }
1333                                 } else if ((line_len + tlen + mbl +
1334                                             (*(wtmpp + 1) ? 0 : nspc) +
1335                                             (line_len > 1 ? 1 : 0))
1336                                            > MAX_LINELEN) {
1337                                         if (1 + tlen + mbl +
1338                                             (*(wtmpp + 1) ? 0 : nspc)
1339                                             >= MAX_LINELEN) {
1340                                                 *tmpp = '\0';
1341                                                 break;
1342                                         }
1343                                         *destp++ = '\n';
1344                                         *destp++ = ' ';
1345                                         line_len = 1;
1346                                         continue;
1347                                 }
1348
1349                                 tmpp += mbl;
1350                                 *tmpp = '\0';
1351
1352                                 tlen += mbl;
1353                                 prev_mbl = mbl;
1354
1355                                 wtmpp++;
1356                         }
1357                         /* g_print("tmp = %s, tlen = %d, mb_seqlen = %d\n",
1358                                 tmp, tlen, mb_seqlen); */
1359
1360                         if (tlen == 0) {
1361                                 g_free(tmp);
1362                                 continue;
1363                         }
1364
1365                         if (line_len > 1 && destp > dest) {
1366                                 *destp++ = ' ';
1367                                 *destp = '\0';
1368                                 line_len++;
1369                         }
1370
1371                         if (str_is_non_ascii) {
1372                                 gchar *tmp_jis;
1373
1374                                 tmp_jis = g_new(gchar, tlen + mb_seqlen + 1);
1375                                 conv_euctojis(tmp_jis,
1376                                               tlen + mb_seqlen + 1, tmp);
1377                                 g_snprintf(destp, len - strlen(dest), "%s%s%s",
1378                                            mimehdr_init, mimehdr_charset,
1379                                            mimehdr_enctype);
1380                                 destp += mimehdr_begin_len;
1381                                 line_len += mimehdr_begin_len;
1382
1383                                 base64_encode(destp, tmp_jis, strlen(tmp_jis));
1384                                 line_len += strlen(destp);
1385                                 destp += strlen(destp);
1386
1387                                 strcpy(destp, mimehdr_end);
1388                                 destp += strlen(mimehdr_end);
1389                                 line_len += strlen(mimehdr_end);
1390
1391                                 g_free(tmp_jis);
1392                         } else {
1393                                 strcpy(destp, tmp);
1394                                 line_len += strlen(destp);
1395                                 destp += strlen(destp);
1396                         }
1397
1398                         g_free(tmp);
1399                         /* g_print("line_len = %d\n\n", line_len); */
1400                 } while (*wtmpp != (wchar_t)0);
1401
1402                 while (iswspace(*wsrcp)) {
1403                         gint mbl;
1404
1405                         mbl = wctomb(destp, *wsrcp++);
1406                         if (mbl != -1) {
1407                                 destp += mbl;
1408                                 line_len += mbl;
1409                         }
1410                 }
1411                 *destp = '\0';
1412
1413                 g_free(wtmp);
1414         }
1415
1416         g_free(wsrc);
1417
1418         /* g_print("dest = %s\n", dest); */
1419 }
1420 #endif /* HAVE_LIBJCONV */