src/codeconv.c

   1 /*
   2  * Sylpheed -- a GTK+ based, lightweight, and fast e-mail client
   3  * Copyright (C) 1999-2012 Hiroyuki Yamamoto and the Claws Mail team
   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 3 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, see <http://www.gnu.org/licenses/>.
  17  *
  18  */
  19
  20 #ifdef HAVE_CONFIG_H
  21 #  include "config.h"
  22 #include "claws-features.h"
  23 #endif
  24
  25 #include "defs.h"
  26
  27 #include <glib.h>
  28 #include <glib/gi18n.h>
  29 #include <string.h>
  30 #include <ctype.h>
  31 #include <stdlib.h>
  32 #include <errno.h>
  33
  34 #if HAVE_LOCALE_H
  35 #  include <locale.h>
  36 #endif
  37
  38 #include "codeconv.h"
  39 #include "unmime.h"
  40 #include "quoted-printable.h"
  41 #include "utils.h"
  42 #include "prefs_common.h"
  43
  44 /* For unknown reasons the inconv.m4 macro undefs that macro if no
  45    const is needed.  This would break the code below so we define it. */
  46 #ifndef ICONV_CONST
  47 #define ICONV_CONST
  48 #endif
  49
  50 typedef enum
  51 {
  52         JIS_ASCII,
  53         JIS_KANJI,
  54         JIS_HWKANA,
  55         JIS_AUXKANJI
  56 } JISState;
  57
  58 #define SUBST_CHAR      0x5f;
  59 #define ESC             '\033'
  60
  61 #define iseuckanji(c) \
  62         (((c) & 0xff) >= 0xa1 && ((c) & 0xff) <= 0xfe)
  63 #define iseuchwkana1(c) \
  64         (((c) & 0xff) == 0x8e)
  65 #define iseuchwkana2(c) \
  66         (((c) & 0xff) >= 0xa1 && ((c) & 0xff) <= 0xdf)
  67 #define iseucaux(c) \
  68         (((c) & 0xff) == 0x8f)
  69 #define issjiskanji1(c) \
  70         ((((c) & 0xff) >= 0x81 && ((c) & 0xff) <= 0x9f) || \
  71          (((c) & 0xff) >= 0xe0 && ((c) & 0xff) <= 0xfc))
  72 #define issjiskanji2(c) \
  73         ((((c) & 0xff) >= 0x40 && ((c) & 0xff) <= 0x7e) || \
  74          (((c) & 0xff) >= 0x80 && ((c) & 0xff) <= 0xfc))
  75 #define issjishwkana(c) \
  76         (((c) & 0xff) >= 0xa1 && ((c) & 0xff) <= 0xdf)
  77
  78 #define K_IN()                          \
  79         if (state != JIS_KANJI) {       \
  80                 *out++ = ESC;           \
  81                 *out++ = '$';           \
  82                 *out++ = 'B';           \
  83                 state = JIS_KANJI;      \
  84         }
  85
  86 #define K_OUT()                         \
  87         if (state != JIS_ASCII) {       \
  88                 *out++ = ESC;           \
  89                 *out++ = '(';           \
  90                 *out++ = 'B';           \
  91                 state = JIS_ASCII;      \
  92         }
  93
  94 #define HW_IN()                         \
  95         if (state != JIS_HWKANA) {      \
  96                 *out++ = ESC;           \
  97                 *out++ = '(';           \
  98                 *out++ = 'I';           \
  99                 state = JIS_HWKANA;     \
 100         }
 101
 102 #define AUX_IN()                        \
 103         if (state != JIS_AUXKANJI) {    \
 104                 *out++ = ESC;           \
 105                 *out++ = '$';           \
 106                 *out++ = '(';           \
 107                 *out++ = 'D';           \
 108                 state = JIS_AUXKANJI;   \
 109         }
 110
 111 static CodeConvFunc conv_get_code_conv_func     (const gchar    *src_charset_str,
 112                                          const gchar    *dest_charset_str);
 113
 114 static gchar *conv_iconv_strdup_with_cd (const gchar    *inbuf,
 115                                          iconv_t         cd);
 116
 117 static gchar *conv_iconv_strdup         (const gchar    *inbuf,
 118                                          const gchar    *src_code,
 119                                          const gchar    *dest_code);
 120
 121 static CharSet conv_get_locale_charset                  (void);
 122 static CharSet conv_get_outgoing_charset                (void);
 123 static CharSet conv_guess_ja_encoding(const gchar *str);
 124 static gboolean conv_is_ja_locale                       (void);
 125
 126 static gint conv_jistoeuc(gchar *outbuf, gint outlen, const gchar *inbuf);
 127 static gint conv_euctojis(gchar *outbuf, gint outlen, const gchar *inbuf);
 128 static gint conv_sjistoeuc(gchar *outbuf, gint outlen, const gchar *inbuf);
 129
 130 static gint conv_jistoutf8(gchar *outbuf, gint outlen, const gchar *inbuf);
 131 static gint conv_sjistoutf8(gchar *outbuf, gint outlen, const gchar *inbuf);
 132 static gint conv_euctoutf8(gchar *outbuf, gint outlen, const gchar *inbuf);
 133 static gint conv_anytoutf8(gchar *outbuf, gint outlen, const gchar *inbuf);
 134
 135 static gint conv_utf8toeuc(gchar *outbuf, gint outlen, const gchar *inbuf);
 136 static gint conv_utf8tojis(gchar *outbuf, gint outlen, const gchar *inbuf);
 137
 138 static void conv_unreadable_8bit(gchar *str);
 139
 140 static gint conv_jistodisp(gchar *outbuf, gint outlen, const gchar *inbuf);
 141 static gint conv_sjistodisp(gchar *outbuf, gint outlen, const gchar *inbuf);
 142 static gint conv_euctodisp(gchar *outbuf, gint outlen, const gchar *inbuf);
 143
 144 static gint conv_anytodisp(gchar *outbuf, gint outlen, const gchar *inbuf);
 145 static gint conv_ustodisp(gchar *outbuf, gint outlen, const gchar *inbuf);
 146 static gint conv_noconv(gchar *outbuf, gint outlen, const gchar *inbuf);
 147
 148 static gboolean strict_mode = FALSE;
 149
 150 void codeconv_set_strict(gboolean mode)
 151 {
 152         strict_mode = mode;
 153 }
 154
 155 static gint conv_jistoeuc(gchar *outbuf, gint outlen, const gchar *inbuf)
 156 {
 157         const guchar *in = inbuf;
 158         gchar *out = outbuf;
 159         JISState state = JIS_ASCII;
 160
 161         /*
 162          * Loop outputs up to 3 bytes in each pass (aux kanji) and we
 163          * need 1 byte to terminate the output
 164          */
 165         while (*in != '\0' && (out - outbuf) < outlen - 4) {
 166                 if (*in == ESC) {
 167                         in++;
 168                         if (*in == '$') {
 169                                 if (*(in + 1) == '@' || *(in + 1) == 'B') {
 170                                         state = JIS_KANJI;
 171                                         in += 2;
 172                                 } else if (*(in + 1) == '(' &&
 173                                            *(in + 2) == 'D') {
 174                                         state = JIS_AUXKANJI;
 175                                         in += 3;
 176                                 } else {
 177                                         /* unknown escape sequence */
 178                                         state = JIS_ASCII;
 179                                 }
 180                         } else if (*in == '(') {
 181                                 if (*(in + 1) == 'B' || *(in + 1) == 'J') {
 182                                         state = JIS_ASCII;
 183                                         in += 2;
 184                                 } else if (*(in + 1) == 'I') {
 185                                         state = JIS_HWKANA;
 186                                         in += 2;
 187                                 } else {
 188                                         /* unknown escape sequence */
 189                                         state = JIS_ASCII;
 190                                 }
 191                         } else {
 192                                 /* unknown escape sequence */
 193                                 state = JIS_ASCII;
 194                         }
 195                 } else if (*in == 0x0e) {
 196                         state = JIS_HWKANA;
 197                         in++;
 198                 } else if (*in == 0x0f) {
 199                         state = JIS_ASCII;
 200                         in++;
 201                 } else {
 202                         switch (state) {
 203                         case JIS_ASCII:
 204                                 *out++ = *in++;
 205                                 break;
 206                         case JIS_KANJI:
 207                                 *out++ = *in++ | 0x80;
 208                                 if (*in == '\0') break;
 209                                 *out++ = *in++ | 0x80;
 210                                 break;
 211                         case JIS_HWKANA:
 212                                 *out++ = 0x8e;
 213                                 *out++ = *in++ | 0x80;
 214                                 break;
 215                         case JIS_AUXKANJI:
 216                                 *out++ = 0x8f;
 217                                 *out++ = *in++ | 0x80;
 218                                 if (*in == '\0') break;
 219                                 *out++ = *in++ | 0x80;
 220                                 break;
 221                         }
 222                 }
 223         }
 224
 225         *out = '\0';
 226         return 0;
 227 }
 228
 229 #define JIS_HWDAKUTEN           0x5e
 230 #define JIS_HWHANDAKUTEN        0x5f
 231
 232 static gint conv_jis_hantozen(guchar *outbuf, guchar jis_code, guchar sound_sym)
 233 {
 234         static guint16 h2z_tbl[] = {
 235                 /* 0x20 - 0x2f */
 236                 0x0000, 0x2123, 0x2156, 0x2157, 0x2122, 0x2126, 0x2572, 0x2521,
 237                 0x2523, 0x2525, 0x2527, 0x2529, 0x2563, 0x2565, 0x2567, 0x2543,
 238                 /* 0x30 - 0x3f */
 239                 0x213c, 0x2522, 0x2524, 0x2526, 0x2528, 0x252a, 0x252b, 0x252d,
 240                 0x252f, 0x2531, 0x2533, 0x2535, 0x2537, 0x2539, 0x253b, 0x253d,
 241                 /* 0x40 - 0x4f */
 242                 0x253f, 0x2541, 0x2544, 0x2546, 0x2548, 0x254a, 0x254b, 0x254c,
 243                 0x254d, 0x254e, 0x254f, 0x2552, 0x2555, 0x2558, 0x255b, 0x255e,
 244                 /* 0x50 - 0x5f */
 245                 0x255f, 0x2560, 0x2561, 0x2562, 0x2564, 0x2566, 0x2568, 0x2569,
 246                 0x256a, 0x256b, 0x256c, 0x256d, 0x256f, 0x2573, 0x212b, 0x212c
 247         };
 248
 249         static guint16 dakuten_tbl[] = {
 250                 /* 0x30 - 0x3f */
 251                 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x252c, 0x252e,
 252                 0x2530, 0x2532, 0x2534, 0x2536, 0x2538, 0x253a, 0x253c, 0x253e,
 253                 /* 0x40 - 0x4f */
 254                 0x2540, 0x2542, 0x2545, 0x2547, 0x2549, 0x0000, 0x0000, 0x0000,
 255                 0x0000, 0x0000, 0x2550, 0x2553, 0x2556, 0x2559, 0x255c, 0x0000
 256         };
 257
 258         static guint16 handakuten_tbl[] = {
 259                 /* 0x4a - 0x4e */
 260                 0x2551, 0x2554, 0x2557, 0x255a, 0x255d
 261         };
 262
 263         guint16 out_code;
 264
 265         jis_code &= 0x7f;
 266         sound_sym &= 0x7f;
 267
 268         if (jis_code < 0x21 || jis_code > 0x5f)
 269                 return 0;
 270
 271         if (sound_sym == JIS_HWDAKUTEN &&
 272             jis_code >= 0x36 && jis_code <= 0x4e) {
 273                 out_code = dakuten_tbl[jis_code - 0x30];
 274                 if (out_code != 0) {
 275                         *outbuf = out_code >> 8;
 276                         *(outbuf + 1) = out_code & 0xff;
 277                         return 2;
 278                 }
 279         }
 280
 281         if (sound_sym == JIS_HWHANDAKUTEN &&
 282             jis_code >= 0x4a && jis_code <= 0x4e) {
 283                 out_code = handakuten_tbl[jis_code - 0x4a];
 284                 *outbuf = out_code >> 8;
 285                 *(outbuf + 1) = out_code & 0xff;
 286                 return 2;
 287         }
 288
 289         out_code = h2z_tbl[jis_code - 0x20];
 290         *outbuf = out_code >> 8;
 291         *(outbuf + 1) = out_code & 0xff;
 292         return 1;
 293 }
 294
 295 static gint conv_euctojis(gchar *outbuf, gint outlen, const gchar *inbuf)
 296 {
 297         const guchar *in = inbuf;
 298         gchar *out = outbuf;
 299         JISState state = JIS_ASCII;
 300
 301         /*
 302          * Loop outputs up to 6 bytes in each pass (aux shift + aux
 303          * kanji) and we need up to 4 bytes to terminate the output
 304          * (ASCII shift + null)
 305          */
 306         while (*in != '\0' && (out - outbuf) < outlen - 10) {
 307                 if (IS_ASCII(*in)) {
 308                         K_OUT();
 309                         *out++ = *in++;
 310                 } else if (iseuckanji(*in)) {
 311                         if (iseuckanji(*(in + 1))) {
 312                                 K_IN();
 313                                 *out++ = *in++ & 0x7f;
 314                                 *out++ = *in++ & 0x7f;
 315                         } else {
 316                                 K_OUT();
 317                                 *out++ = SUBST_CHAR;
 318                                 in++;
 319                                 if (*in != '\0' && !IS_ASCII(*in)) {
 320                                         *out++ = SUBST_CHAR;
 321                                         in++;
 322                                 }
 323                         }
 324                 } else if (iseuchwkana1(*in)) {
 325                         if (iseuchwkana2(*(in + 1))) {
 326                                 if (prefs_common.allow_jisx0201_kana) {
 327                                         HW_IN();
 328                                         in++;
 329                                         *out++ = *in++ & 0x7f;
 330                                 } else {
 331                                         guchar jis_ch[2];
 332                                         gint len;
 333
 334                                         if (iseuchwkana1(*(in + 2)) &&
 335                                             iseuchwkana2(*(in + 3)))
 336                                                 len = conv_jis_hantozen
 337                                                         (jis_ch,
 338                                                          *(in + 1), *(in + 3));
 339                                         else
 340                                                 len = conv_jis_hantozen
 341                                                         (jis_ch,
 342                                                          *(in + 1), '\0');
 343                                         if (len == 0)
 344                                                 in += 2;
 345                                         else {
 346                                                 K_IN();
 347                                                 in += len * 2;
 348                                                 *out++ = jis_ch[0];
 349                                                 *out++ = jis_ch[1];
 350                                         }
 351                                 }
 352                         } else {
 353                                 K_OUT();
 354                                 in++;
 355                                 if (*in != '\0' && !IS_ASCII(*in)) {
 356                                         *out++ = SUBST_CHAR;
 357                                         in++;
 358                                 }
 359                         }
 360                 } else if (iseucaux(*in)) {
 361                         in++;
 362                         if (iseuckanji(*in) && iseuckanji(*(in + 1))) {
 363                                 AUX_IN();
 364                                 *out++ = *in++ & 0x7f;
 365                                 *out++ = *in++ & 0x7f;
 366                         } else {
 367                                 K_OUT();
 368                                 if (*in != '\0' && !IS_ASCII(*in)) {
 369                                         *out++ = SUBST_CHAR;
 370                                         in++;
 371                                         if (*in != '\0' && !IS_ASCII(*in)) {
 372                                                 *out++ = SUBST_CHAR;
 373                                                 in++;
 374                                         }
 375                                 }
 376                         }
 377                 } else {
 378                         K_OUT();
 379                         *out++ = SUBST_CHAR;
 380                         in++;
 381                 }
 382         }
 383
 384         K_OUT();
 385         *out = '\0';
 386         return 0;
 387 }
 388
 389 static gint conv_sjistoeuc(gchar *outbuf, gint outlen, const gchar *inbuf)
 390 {
 391         const guchar *in = inbuf;
 392         gchar *out = outbuf;
 393
 394         /*
 395          * Loop outputs up to 2 bytes in each pass and we need 1 byte
 396          * to terminate the output
 397          */
 398         while (*in != '\0' && (out - outbuf) < outlen - 3) {
 399                 if (IS_ASCII(*in)) {
 400                         *out++ = *in++;
 401                 } else if (issjiskanji1(*in)) {
 402                         if (issjiskanji2(*(in + 1))) {
 403                                 guchar out1 = *in;
 404                                 guchar out2 = *(in + 1);
 405                                 guchar row;
 406
 407                                 row = out1 < 0xa0 ? 0x70 : 0xb0;
 408                                 if (out2 < 0x9f) {
 409                                         out1 = (out1 - row) * 2 - 1;
 410                                         out2 -= out2 > 0x7f ? 0x20 : 0x1f;
 411                                 } else {
 412                                         out1 = (out1 - row) * 2;
 413                                         out2 -= 0x7e;
 414                                 }
 415
 416                                 *out++ = out1 | 0x80;
 417                                 *out++ = out2 | 0x80;
 418                                 in += 2;
 419                         } else {
 420                                 *out++ = SUBST_CHAR;
 421                                 in++;
 422                                 if (*in != '\0' && !IS_ASCII(*in)) {
 423                                         *out++ = SUBST_CHAR;
 424                                         in++;
 425                                 }
 426                         }
 427                 } else if (issjishwkana(*in)) {
 428                         *out++ = 0x8e;
 429                         *out++ = *in++;
 430                 } else {
 431                         *out++ = SUBST_CHAR;
 432                         in++;
 433                 }
 434         }
 435
 436         *out = '\0';
 437         return 0;
 438 }
 439
 440 static gint conv_jistoutf8(gchar *outbuf, gint outlen, const gchar *inbuf)
 441 {
 442         gchar *eucstr;
 443
 444         Xalloca(eucstr, outlen, return -1);
 445
 446         if (conv_jistoeuc(eucstr, outlen, inbuf) <0)
 447                 return -1;
 448         if (conv_euctoutf8(outbuf, outlen, eucstr) < 0)
 449                 return -1;
 450         return 0;
 451 }
 452
 453 static gint conv_sjistoutf8(gchar *outbuf, gint outlen, const gchar *inbuf)
 454 {
 455         gchar *tmpstr;
 456
 457         tmpstr = conv_iconv_strdup(inbuf, CS_SHIFT_JIS, CS_UTF_8);
 458         if (tmpstr) {
 459                 strncpy2(outbuf, tmpstr, outlen);
 460                 g_free(tmpstr);
 461                 return 0;
 462         } else {
 463                 strncpy2(outbuf, inbuf, outlen);
 464                 return -1;
 465         }
 466 }
 467
 468 static gint conv_euctoutf8(gchar *outbuf, gint outlen, const gchar *inbuf)
 469 {
 470         static iconv_t cd = (iconv_t)-1;
 471         static gboolean iconv_ok = TRUE;
 472         gchar *tmpstr;
 473
 474         if (cd == (iconv_t)-1) {
 475                 if (!iconv_ok) {
 476                         strncpy2(outbuf, inbuf, outlen);
 477                         return -1;
 478                 }
 479                 cd = iconv_open(CS_UTF_8, CS_EUC_JP_MS);
 480                 if (cd == (iconv_t)-1) {
 481                         cd = iconv_open(CS_UTF_8, CS_EUC_JP);
 482                         if (cd == (iconv_t)-1) {
 483                                 g_warning("conv_euctoutf8(): %s",
 484                                           g_strerror(errno));
 485                                 iconv_ok = FALSE;
 486                                 strncpy2(outbuf, inbuf, outlen);
 487                                 return -1;
 488                         }
 489                 }
 490         }
 491
 492         tmpstr = conv_iconv_strdup_with_cd(inbuf, cd);
 493         if (tmpstr) {
 494                 strncpy2(outbuf, tmpstr, outlen);
 495                 g_free(tmpstr);
 496                 return 0;
 497         } else {
 498                 strncpy2(outbuf, inbuf, outlen);
 499                 return -1;
 500         }
 501 }
 502
 503 static gint conv_anytoutf8(gchar *outbuf, gint outlen, const gchar *inbuf)
 504 {
 505         gint r = -1;
 506         switch (conv_guess_ja_encoding(inbuf)) {
 507         case C_ISO_2022_JP:
 508                 r = conv_jistoutf8(outbuf, outlen, inbuf);
 509                 break;
 510         case C_SHIFT_JIS:
 511                 r = conv_sjistoutf8(outbuf, outlen, inbuf);
 512                 break;
 513         case C_EUC_JP:
 514                 r = conv_euctoutf8(outbuf, outlen, inbuf);
 515                 break;
 516         default:
 517                 r = 0;
 518                 strncpy2(outbuf, inbuf, outlen);
 519                 break;
 520         }
 521
 522         return r;
 523 }
 524
 525 static gint conv_utf8toeuc(gchar *outbuf, gint outlen, const gchar *inbuf)
 526 {
 527         static iconv_t cd = (iconv_t)-1;
 528         static gboolean iconv_ok = TRUE;
 529         gchar *tmpstr;
 530
 531         if (cd == (iconv_t)-1) {
 532                 if (!iconv_ok) {
 533                         strncpy2(outbuf, inbuf, outlen);
 534                         return -1;
 535                 }
 536                 cd = iconv_open(CS_EUC_JP_MS, CS_UTF_8);
 537                 if (cd == (iconv_t)-1) {
 538                         cd = iconv_open(CS_EUC_JP, CS_UTF_8);
 539                         if (cd == (iconv_t)-1) {
 540                                 g_warning("conv_utf8toeuc(): %s",
 541                                           g_strerror(errno));
 542                                 iconv_ok = FALSE;
 543                                 strncpy2(outbuf, inbuf, outlen);
 544                                 return -1;
 545                         }
 546                 }
 547         }
 548
 549         tmpstr = conv_iconv_strdup_with_cd(inbuf, cd);
 550         if (tmpstr) {
 551                 strncpy2(outbuf, tmpstr, outlen);
 552                 g_free(tmpstr);
 553                 return 0;
 554         } else {
 555                 strncpy2(outbuf, inbuf, outlen);
 556                 return -1;
 557         }
 558 }
 559
 560 static gint conv_utf8tojis(gchar *outbuf, gint outlen, const gchar *inbuf)
 561 {
 562         gchar *eucstr;
 563
 564         Xalloca(eucstr, outlen, return -1);
 565
 566         if (conv_utf8toeuc(eucstr, outlen, inbuf) < 0)
 567                 return -1;
 568         if (conv_euctojis(outbuf, outlen, eucstr) < 0)
 569                 return -1;
 570
 571         return 0;
 572 }
 573
 574 static void conv_unreadable_8bit(gchar *str)
 575 {
 576         register guchar *p = str;
 577
 578         while (*p != '\0') {
 579                 /* convert CR+LF -> LF */
 580                 if (*p == '\r' && *(p + 1) == '\n')
 581                         memmove(p, p + 1, strlen(p));
 582                 else if (!IS_ASCII(*p)) *p = SUBST_CHAR;
 583                 p++;
 584         }
 585 }
 586
 587 static CharSet conv_guess_ja_encoding(const gchar *str)
 588 {
 589         const guchar *p = str;
 590         CharSet guessed = C_US_ASCII;
 591
 592         while (*p != '\0') {
 593                 if (*p == ESC && (*(p + 1) == '$' || *(p + 1) == '(')) {
 594                         if (guessed == C_US_ASCII)
 595                                 return C_ISO_2022_JP;
 596                         p += 2;
 597                 } else if (IS_ASCII(*p)) {
 598                         p++;
 599                 } else if (iseuckanji(*p) && iseuckanji(*(p + 1))) {
 600                         if (*p >= 0xfd && *p <= 0xfe)
 601                                 return C_EUC_JP;
 602                         else if (guessed == C_SHIFT_JIS) {
 603                                 if ((issjiskanji1(*p) &&
 604                                      issjiskanji2(*(p + 1))) ||
 605                                     issjishwkana(*p))
 606                                         guessed = C_SHIFT_JIS;
 607                                 else
 608                                         guessed = C_EUC_JP;
 609                         } else
 610                                 guessed = C_EUC_JP;
 611                         p += 2;
 612                 } else if (issjiskanji1(*p) && issjiskanji2(*(p + 1))) {
 613                         if (iseuchwkana1(*p) && iseuchwkana2(*(p + 1)))
 614                                 guessed = C_SHIFT_JIS;
 615                         else
 616                                 return C_SHIFT_JIS;
 617                         p += 2;
 618                 } else if (issjishwkana(*p)) {
 619                         guessed = C_SHIFT_JIS;
 620                         p++;
 621                 } else {
 622                         p++;
 623                 }
 624         }
 625
 626         return guessed;
 627 }
 628
 629 static gint conv_jistodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
 630 {
 631         return conv_jistoutf8(outbuf, outlen, inbuf);
 632 }
 633
 634 static gint conv_sjistodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
 635 {
 636         return conv_sjistoutf8(outbuf, outlen, inbuf);
 637 }
 638
 639 static gint conv_euctodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
 640 {
 641         return conv_euctoutf8(outbuf, outlen, inbuf);
 642 }
 643
 644 void conv_utf8todisp(gchar *outbuf, gint outlen, const gchar *inbuf)
 645 {
 646         if (g_utf8_validate(inbuf, -1, NULL) == TRUE)
 647                 strncpy2(outbuf, inbuf, outlen);
 648         else
 649                 conv_ustodisp(outbuf, outlen, inbuf);
 650 }
 651
 652 static gint conv_anytodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
 653 {
 654         gint r = 0;
 655         if (conv_anytoutf8(outbuf, outlen, inbuf) < 0)
 656                 r = -1;
 657         if (g_utf8_validate(outbuf, -1, NULL) != TRUE)
 658                 conv_unreadable_8bit(outbuf);
 659         return r;
 660 }
 661
 662 static gint conv_ustodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
 663 {
 664         strncpy2(outbuf, inbuf, outlen);
 665         conv_unreadable_8bit(outbuf);
 666
 667         return 0;
 668 }
 669
 670 void conv_localetodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
 671 {
 672         gchar *tmpstr;
 673
 674         codeconv_set_strict(TRUE);
 675         tmpstr = conv_iconv_strdup(inbuf, conv_get_locale_charset_str(),
 676                                    CS_INTERNAL);
 677         codeconv_set_strict(FALSE);
 678         if (tmpstr && g_utf8_validate(tmpstr, -1, NULL)) {
 679                 strncpy2(outbuf, tmpstr, outlen);
 680                 g_free(tmpstr);
 681                 return;
 682         } else if (tmpstr && !g_utf8_validate(tmpstr, -1, NULL)) {
 683                 g_free(tmpstr);
 684                 codeconv_set_strict(TRUE);
 685                 tmpstr = conv_iconv_strdup(inbuf,
 686                                 conv_get_locale_charset_str_no_utf8(),
 687                                 CS_INTERNAL);
 688                 codeconv_set_strict(FALSE);
 689         }
 690         if (tmpstr && g_utf8_validate(tmpstr, -1, NULL)) {
 691                 strncpy2(outbuf, tmpstr, outlen);
 692                 g_free(tmpstr);
 693                 return;
 694         } else {
 695                 g_free(tmpstr);
 696                 conv_utf8todisp(outbuf, outlen, inbuf);
 697         }
 698 }
 699
 700 static gint conv_noconv(gchar *outbuf, gint outlen, const gchar *inbuf)
 701 {
 702         strncpy2(outbuf, inbuf, outlen);
 703         return 0;
 704 }
 705
 706 static const gchar *
 707 conv_get_fallback_for_private_encoding(const gchar *encoding)
 708 {
 709         if (encoding) {
 710                 if ((encoding[0] == 'X' || encoding[0] == 'x') &&
 711                     encoding[1] == '-') {
 712                         if (!g_ascii_strcasecmp(encoding, CS_X_MACCYR))
 713                                 return CS_MACCYR;
 714                         if (!g_ascii_strcasecmp(encoding, CS_X_GBK))
 715                                 return CS_GBK;
 716                 }
 717                 else if(!g_ascii_strcasecmp(encoding, CS_ISO_8859_8_I)) {
 718                         /*
 719                          * ISO-8859-8-I is a variant which fully
 720                          * agrees with ISO-8859-8 on character
 721                          * codings, and differs only in directionality
 722                          * implications, which are ignored here
 723                          * anyway; and is not recognized by iconv
 724                          */
 725                         return CS_ISO_8859_8;
 726                 }
 727         }
 728
 729         return encoding;
 730 }
 731
 732 CodeConverter *conv_code_converter_new(const gchar *src_charset)
 733 {
 734         CodeConverter *conv;
 735
 736         src_charset = conv_get_fallback_for_private_encoding(src_charset);
 737
 738         conv = g_new0(CodeConverter, 1);
 739         conv->code_conv_func = conv_get_code_conv_func(src_charset, NULL);
 740         conv->charset_str = g_strdup(src_charset);
 741         conv->charset = conv_get_charset_from_str(src_charset);
 742
 743         return conv;
 744 }
 745
 746 void conv_code_converter_destroy(CodeConverter *conv)
 747 {
 748         g_free(conv->charset_str);
 749         g_free(conv);
 750 }
 751
 752 gint conv_convert(CodeConverter *conv, gchar *outbuf, gint outlen,
 753                   const gchar *inbuf)
 754 {
 755         if (conv->code_conv_func != conv_noconv)
 756                 return conv->code_conv_func(outbuf, outlen, inbuf);
 757         else {
 758                 gchar *str;
 759
 760                 str = conv_iconv_strdup(inbuf, conv->charset_str, NULL);
 761                 if (!str)
 762                         return -1;
 763                 else {
 764                         strncpy2(outbuf, str, outlen);
 765                         g_free(str);
 766                 }
 767         }
 768
 769         return 0;
 770 }
 771
 772 gchar *conv_codeset_strdup(const gchar *inbuf,
 773                            const gchar *src_code, const gchar *dest_code)
 774 {
 775         gchar *buf;
 776         size_t len;
 777         CodeConvFunc conv_func;
 778
 779         if (!strcmp2(src_code, dest_code)) {
 780                 CharSet dest_charset = conv_get_charset_from_str(dest_code);
 781                 if (strict_mode && dest_charset == C_UTF_8) {
 782                         /* ensure valid UTF-8 if target is UTF-8 */
 783                         if (!g_utf8_validate(inbuf, -1, NULL)) {
 784                                 return NULL;
 785                         }
 786                 }
 787                 /* otherwise, try for a lucky day */
 788                 return g_strdup(inbuf);
 789         }
 790
 791         src_code = conv_get_fallback_for_private_encoding(src_code);
 792         conv_func = conv_get_code_conv_func(src_code, dest_code);
 793         if (conv_func == conv_ustodisp && strict_mode && !is_ascii_str(inbuf))
 794                 return NULL;
 795
 796         if (conv_func != conv_noconv) {
 797                 len = (strlen(inbuf) + 1) * 3;
 798                 buf = g_malloc(len);
 799
 800                 if (conv_func(buf, len, inbuf) == 0 || !strict_mode)
 801                         return g_realloc(buf, strlen(buf) + 1);
 802                 else {
 803                         g_free(buf);
 804                         return NULL;
 805                 }
 806         }
 807
 808         return conv_iconv_strdup(inbuf, src_code, dest_code);
 809 }
 810
 811 static CodeConvFunc conv_get_code_conv_func(const gchar *src_charset_str,
 812                                      const gchar *dest_charset_str)
 813 {
 814         CodeConvFunc code_conv = conv_noconv;
 815         CharSet src_charset;
 816         CharSet dest_charset;
 817
 818         if (!src_charset_str)
 819                 src_charset = conv_get_locale_charset();
 820         else
 821                 src_charset = conv_get_charset_from_str(src_charset_str);
 822
 823         /* auto detection mode */
 824         if (!src_charset_str && !dest_charset_str) {
 825                 if (conv_is_ja_locale())
 826                         return conv_anytodisp;
 827                 else
 828                         return conv_noconv;
 829         }
 830
 831         dest_charset = conv_get_charset_from_str(dest_charset_str);
 832
 833         if (dest_charset == C_US_ASCII)
 834                 return conv_ustodisp;
 835
 836         switch (src_charset) {
 837         case C_US_ASCII:
 838         case C_ISO_8859_1:
 839         case C_ISO_8859_2:
 840         case C_ISO_8859_3:
 841         case C_ISO_8859_4:
 842         case C_ISO_8859_5:
 843         case C_ISO_8859_6:
 844         case C_ISO_8859_7:
 845         case C_ISO_8859_8:
 846         case C_ISO_8859_9:
 847         case C_ISO_8859_10:
 848         case C_ISO_8859_11:
 849         case C_ISO_8859_13:
 850         case C_ISO_8859_14:
 851         case C_ISO_8859_15:
 852                 break;
 853         case C_ISO_2022_JP:
 854         case C_ISO_2022_JP_2:
 855         case C_ISO_2022_JP_3:
 856                 if (dest_charset == C_AUTO)
 857                         code_conv = conv_jistodisp;
 858                 else if (dest_charset == C_EUC_JP)
 859                         code_conv = conv_jistoeuc;
 860                 else if (dest_charset == C_UTF_8)
 861                         code_conv = conv_jistoutf8;
 862                 break;
 863         case C_SHIFT_JIS:
 864                 if (dest_charset == C_AUTO)
 865                         code_conv = conv_sjistodisp;
 866                 else if (dest_charset == C_EUC_JP)
 867                         code_conv = conv_sjistoeuc;
 868                 else if (dest_charset == C_UTF_8)
 869                         code_conv = conv_sjistoutf8;
 870                 break;
 871         case C_EUC_JP:
 872                 if (dest_charset == C_AUTO)
 873                         code_conv = conv_euctodisp;
 874                 else if (dest_charset == C_ISO_2022_JP   ||
 875                          dest_charset == C_ISO_2022_JP_2 ||
 876                          dest_charset == C_ISO_2022_JP_3)
 877                         code_conv = conv_euctojis;
 878                 else if (dest_charset == C_UTF_8)
 879                         code_conv = conv_euctoutf8;
 880                 break;
 881         case C_UTF_8:
 882                 if (dest_charset == C_EUC_JP)
 883                         code_conv = conv_utf8toeuc;
 884                 else if (dest_charset == C_ISO_2022_JP   ||
 885                          dest_charset == C_ISO_2022_JP_2 ||
 886                          dest_charset == C_ISO_2022_JP_3)
 887                         code_conv = conv_utf8tojis;
 888                 break;
 889         default:
 890                 break;
 891         }
 892
 893         return code_conv;
 894 }
 895
 896 static gchar *conv_iconv_strdup(const gchar *inbuf,
 897                          const gchar *src_code, const gchar *dest_code)
 898 {
 899         iconv_t cd;
 900         gchar *outbuf;
 901
 902         if (!src_code && !dest_code &&
 903             g_utf8_validate(inbuf, -1, NULL))
 904                 return g_strdup(inbuf);
 905
 906         if (!src_code)
 907                 src_code = conv_get_outgoing_charset_str();
 908         if (!dest_code)
 909                 dest_code = CS_INTERNAL;
 910
 911         /* don't convert if src and dest codeset are identical */
 912         if (!strcasecmp(src_code, dest_code))
 913                 return g_strdup(inbuf);
 914
 915         /* don't convert if dest codeset is US-ASCII */
 916         if (!strcasecmp(src_code, CS_US_ASCII))
 917                 return g_strdup(inbuf);
 918
 919         /* don't convert if dest codeset is US-ASCII */
 920         if (!strcasecmp(dest_code, CS_US_ASCII))
 921                 return g_strdup(inbuf);
 922
 923         cd = iconv_open(dest_code, src_code);
 924         if (cd == (iconv_t)-1)
 925                 return NULL;
 926
 927         outbuf = conv_iconv_strdup_with_cd(inbuf, cd);
 928
 929         iconv_close(cd);
 930
 931         return outbuf;
 932 }
 933
 934 gchar *conv_iconv_strdup_with_cd(const gchar *inbuf, iconv_t cd)
 935 {
 936         const gchar *inbuf_p;
 937         gchar *outbuf;
 938         gchar *outbuf_p;
 939         size_t in_size;
 940         size_t in_left;
 941         size_t out_size;
 942         size_t out_left;
 943         size_t n_conv;
 944         size_t len;
 945
 946         inbuf_p = inbuf;
 947         in_size = strlen(inbuf);
 948         in_left = in_size;
 949         out_size = (in_size + 1) * 2;
 950         outbuf = g_malloc(out_size);
 951         outbuf_p = outbuf;
 952         out_left = out_size;
 953
 954 #define EXPAND_BUF()                            \
 955 {                                               \
 956         len = outbuf_p - outbuf;                \
 957         out_size *= 2;                          \
 958         outbuf = g_realloc(outbuf, out_size);   \
 959         outbuf_p = outbuf + len;                \
 960         out_left = out_size - len;              \
 961 }
 962
 963         while ((n_conv = iconv(cd, (ICONV_CONST gchar **)&inbuf_p, &in_left,
 964                                &outbuf_p, &out_left)) == (size_t)-1) {
 965                 if (EILSEQ == errno) {
 966                         if (strict_mode) {
 967                                 g_free(outbuf);
 968                                 return NULL;
 969                         }
 970                         //g_print("iconv(): at %d: %s\n", in_size - in_left, g_strerror(errno));
 971                         inbuf_p++;
 972                         in_left--;
 973                         if (out_left == 0) {
 974                                 EXPAND_BUF();
 975                         }
 976                         *outbuf_p++ = SUBST_CHAR;
 977                         out_left--;
 978                 } else if (EINVAL == errno) {
 979                         break;
 980                 } else if (E2BIG == errno) {
 981                         EXPAND_BUF();
 982                 } else {
 983                         g_warning("conv_iconv_strdup(): %s",
 984                                   g_strerror(errno));
 985                         break;
 986                 }
 987         }
 988
 989         while ((n_conv = iconv(cd, NULL, NULL, &outbuf_p, &out_left)) ==
 990                (size_t)-1) {
 991                 if (E2BIG == errno) {
 992                         EXPAND_BUF();
 993                 } else {
 994                         g_warning("conv_iconv_strdup(): %s",
 995                                   g_strerror(errno));
 996                         break;
 997                 }
 998         }
 999
1000 #undef EXPAND_BUF
1001
1002         len = outbuf_p - outbuf;
1003         outbuf = g_realloc(outbuf, len + 1);
1004         outbuf[len] = '\0';
1005
1006         return outbuf;
1007 }
1008
1009 static const struct {
1010         CharSet charset;
1011         gchar *const name;
1012 } charsets[] = {
1013         {C_US_ASCII,            CS_US_ASCII},
1014         {C_US_ASCII,            CS_ANSI_X3_4_1968},
1015         {C_UTF_8,               CS_UTF_8},
1016         {C_UTF_7,               CS_UTF_7},
1017         {C_ISO_8859_1,          CS_ISO_8859_1},
1018         {C_ISO_8859_2,          CS_ISO_8859_2},
1019         {C_ISO_8859_3,          CS_ISO_8859_3},
1020         {C_ISO_8859_4,          CS_ISO_8859_4},
1021         {C_ISO_8859_5,          CS_ISO_8859_5},
1022         {C_ISO_8859_6,          CS_ISO_8859_6},
1023         {C_ISO_8859_7,          CS_ISO_8859_7},
1024         {C_ISO_8859_8,          CS_ISO_8859_8},
1025         {C_ISO_8859_9,          CS_ISO_8859_9},
1026         {C_ISO_8859_10,         CS_ISO_8859_10},
1027         {C_ISO_8859_11,         CS_ISO_8859_11},
1028         {C_ISO_8859_13,         CS_ISO_8859_13},
1029         {C_ISO_8859_14,         CS_ISO_8859_14},
1030         {C_ISO_8859_15,         CS_ISO_8859_15},
1031         {C_BALTIC,              CS_BALTIC},
1032         {C_CP1250,              CS_CP1250},
1033         {C_CP1251,              CS_CP1251},
1034         {C_CP1252,              CS_CP1252},
1035         {C_CP1253,              CS_CP1253},
1036         {C_CP1254,              CS_CP1254},
1037         {C_CP1255,              CS_CP1255},
1038         {C_CP1256,              CS_CP1256},
1039         {C_CP1257,              CS_CP1257},
1040         {C_CP1258,              CS_CP1258},
1041         {C_WINDOWS_1250,        CS_WINDOWS_1250},
1042         {C_WINDOWS_1251,        CS_WINDOWS_1251},
1043         {C_WINDOWS_1252,        CS_WINDOWS_1252},
1044         {C_WINDOWS_1253,        CS_WINDOWS_1253},
1045         {C_WINDOWS_1254,        CS_WINDOWS_1254},
1046         {C_WINDOWS_1255,        CS_WINDOWS_1255},
1047         {C_WINDOWS_1256,        CS_WINDOWS_1256},
1048         {C_WINDOWS_1257,        CS_WINDOWS_1257},
1049         {C_WINDOWS_1258,        CS_WINDOWS_1258},
1050         {C_KOI8_R,              CS_KOI8_R},
1051         {C_MACCYR,              CS_MACCYR},
1052         {C_KOI8_T,              CS_KOI8_T},
1053         {C_KOI8_U,              CS_KOI8_U},
1054         {C_ISO_2022_JP,         CS_ISO_2022_JP},
1055         {C_ISO_2022_JP_2,       CS_ISO_2022_JP_2},
1056         {C_ISO_2022_JP_3,       CS_ISO_2022_JP_3},
1057         {C_EUC_JP,              CS_EUC_JP},
1058         {C_EUC_JP,              CS_EUCJP},
1059         {C_EUC_JP_MS,           CS_EUC_JP_MS},
1060         {C_SHIFT_JIS,           CS_SHIFT_JIS},
1061         {C_SHIFT_JIS,           CS_SHIFT__JIS},
1062         {C_SHIFT_JIS,           CS_SJIS},
1063         {C_ISO_2022_KR,         CS_ISO_2022_KR},
1064         {C_EUC_KR,              CS_EUC_KR},
1065         {C_ISO_2022_CN,         CS_ISO_2022_CN},
1066         {C_EUC_CN,              CS_EUC_CN},
1067         {C_GB18030,             CS_GB18030},
1068         {C_GB2312,              CS_GB2312},
1069         {C_GBK,                 CS_GBK},
1070         {C_EUC_TW,              CS_EUC_TW},
1071         {C_BIG5,                CS_BIG5},
1072         {C_BIG5_HKSCS,          CS_BIG5_HKSCS},
1073         {C_TIS_620,             CS_TIS_620},
1074         {C_WINDOWS_874,         CS_WINDOWS_874},
1075         {C_GEORGIAN_PS,         CS_GEORGIAN_PS},
1076         {C_TCVN5712_1,          CS_TCVN5712_1},
1077 };
1078
1079 static const struct {
1080         gchar *const locale;
1081         CharSet charset;
1082         CharSet out_charset;
1083 } locale_table[] = {
1084         {"ja_JP.eucJP"          , C_EUC_JP      , C_ISO_2022_JP},
1085         {"ja_JP.EUC-JP"         , C_EUC_JP      , C_ISO_2022_JP},
1086         {"ja_JP.EUC"            , C_EUC_JP      , C_ISO_2022_JP},
1087         {"ja_JP.ujis"           , C_EUC_JP      , C_ISO_2022_JP},
1088         {"ja_JP.SJIS"           , C_SHIFT_JIS   , C_ISO_2022_JP},
1089         {"ja_JP.JIS"            , C_ISO_2022_JP , C_ISO_2022_JP},
1090 #ifdef G_OS_WIN32
1091         {"ja_JP"                , C_SHIFT_JIS   , C_ISO_2022_JP},
1092 #else
1093         {"ja_JP"                , C_EUC_JP      , C_ISO_2022_JP},
1094 #endif
1095         {"ko_KR.EUC-KR"         , C_EUC_KR      , C_EUC_KR},
1096         {"ko_KR"                , C_EUC_KR      , C_EUC_KR},
1097         {"zh_CN.GB18030"        , C_GB18030     , C_GB18030},
1098         {"zh_CN.GB2312"         , C_GB2312      , C_GB2312},
1099         {"zh_CN.GBK"            , C_GBK         , C_GBK},
1100         {"zh_CN"                , C_GB18030     , C_GB18030},
1101         {"zh_HK"                , C_BIG5_HKSCS  , C_BIG5_HKSCS},
1102         {"zh_TW.eucTW"          , C_EUC_TW      , C_BIG5},
1103         {"zh_TW.EUC-TW"         , C_EUC_TW      , C_BIG5},
1104         {"zh_TW.Big5"           , C_BIG5        , C_BIG5},
1105         {"zh_TW"                , C_BIG5        , C_BIG5},
1106
1107         {"ru_RU.KOI8-R"         , C_KOI8_R      , C_KOI8_R},
1108         {"ru_RU.KOI8R"          , C_KOI8_R      , C_KOI8_R},
1109         {"ru_RU.CP1251"         , C_WINDOWS_1251, C_KOI8_R},
1110 #ifdef G_OS_WIN32
1111         {"ru_RU"                , C_WINDOWS_1251, C_KOI8_R},
1112 #else
1113         {"ru_RU"                , C_ISO_8859_5  , C_KOI8_R},
1114 #endif
1115         {"tg_TJ"                , C_KOI8_T      , C_KOI8_T},
1116         {"ru_UA"                , C_KOI8_U      , C_KOI8_U},
1117         {"uk_UA.CP1251"         , C_WINDOWS_1251, C_KOI8_U},
1118         {"uk_UA"                , C_KOI8_U      , C_KOI8_U},
1119
1120         {"be_BY"                , C_WINDOWS_1251, C_WINDOWS_1251},
1121         {"bg_BG"                , C_WINDOWS_1251, C_WINDOWS_1251},
1122
1123         {"yi_US"                , C_WINDOWS_1255, C_WINDOWS_1255},
1124
1125         {"af_ZA"                , C_ISO_8859_1  , C_ISO_8859_1},
1126         {"br_FR"                , C_ISO_8859_1  , C_ISO_8859_1},
1127         {"ca_ES"                , C_ISO_8859_1  , C_ISO_8859_1},
1128         {"da_DK"                , C_ISO_8859_1  , C_ISO_8859_1},
1129         {"de_AT"                , C_ISO_8859_1  , C_ISO_8859_1},
1130         {"de_BE"                , C_ISO_8859_1  , C_ISO_8859_1},
1131         {"de_CH"                , C_ISO_8859_1  , C_ISO_8859_1},
1132         {"de_DE"                , C_ISO_8859_1  , C_ISO_8859_1},
1133         {"de_LU"                , C_ISO_8859_1  , C_ISO_8859_1},
1134         {"en_AU"                , C_ISO_8859_1  , C_ISO_8859_1},
1135         {"en_BW"                , C_ISO_8859_1  , C_ISO_8859_1},
1136         {"en_CA"                , C_ISO_8859_1  , C_ISO_8859_1},
1137         {"en_DK"                , C_ISO_8859_1  , C_ISO_8859_1},
1138         {"en_GB"                , C_ISO_8859_1  , C_ISO_8859_1},
1139         {"en_HK"                , C_ISO_8859_1  , C_ISO_8859_1},
1140         {"en_IE"                , C_ISO_8859_1  , C_ISO_8859_1},
1141         {"en_NZ"                , C_ISO_8859_1  , C_ISO_8859_1},
1142         {"en_PH"                , C_ISO_8859_1  , C_ISO_8859_1},
1143         {"en_SG"                , C_ISO_8859_1  , C_ISO_8859_1},
1144         {"en_US"                , C_ISO_8859_1  , C_ISO_8859_1},
1145         {"en_ZA"                , C_ISO_8859_1  , C_ISO_8859_1},
1146         {"en_ZW"                , C_ISO_8859_1  , C_ISO_8859_1},
1147         {"es_AR"                , C_ISO_8859_1  , C_ISO_8859_1},
1148         {"es_BO"                , C_ISO_8859_1  , C_ISO_8859_1},
1149         {"es_CL"                , C_ISO_8859_1  , C_ISO_8859_1},
1150         {"es_CO"                , C_ISO_8859_1  , C_ISO_8859_1},
1151         {"es_CR"                , C_ISO_8859_1  , C_ISO_8859_1},
1152         {"es_DO"                , C_ISO_8859_1  , C_ISO_8859_1},
1153         {"es_EC"                , C_ISO_8859_1  , C_ISO_8859_1},
1154         {"es_ES"                , C_ISO_8859_1  , C_ISO_8859_1},
1155         {"es_GT"                , C_ISO_8859_1  , C_ISO_8859_1},
1156         {"es_HN"                , C_ISO_8859_1  , C_ISO_8859_1},
1157         {"es_MX"                , C_ISO_8859_1  , C_ISO_8859_1},
1158         {"es_NI"                , C_ISO_8859_1  , C_ISO_8859_1},
1159         {"es_PA"                , C_ISO_8859_1  , C_ISO_8859_1},
1160         {"es_PE"                , C_ISO_8859_1  , C_ISO_8859_1},
1161         {"es_PR"                , C_ISO_8859_1  , C_ISO_8859_1},
1162         {"es_PY"                , C_ISO_8859_1  , C_ISO_8859_1},
1163         {"es_SV"                , C_ISO_8859_1  , C_ISO_8859_1},
1164         {"es_US"                , C_ISO_8859_1  , C_ISO_8859_1},
1165         {"es_UY"                , C_ISO_8859_1  , C_ISO_8859_1},
1166         {"es_VE"                , C_ISO_8859_1  , C_ISO_8859_1},
1167         {"et_EE"                , C_ISO_8859_1  , C_ISO_8859_1},
1168         {"eu_ES"                , C_ISO_8859_1  , C_ISO_8859_1},
1169         {"fi_FI"                , C_ISO_8859_1  , C_ISO_8859_1},
1170         {"fo_FO"                , C_ISO_8859_1  , C_ISO_8859_1},
1171         {"fr_BE"                , C_ISO_8859_1  , C_ISO_8859_1},
1172         {"fr_CA"                , C_ISO_8859_1  , C_ISO_8859_1},
1173         {"fr_CH"                , C_ISO_8859_1  , C_ISO_8859_1},
1174         {"fr_FR"                , C_ISO_8859_1  , C_ISO_8859_1},
1175         {"fr_LU"                , C_ISO_8859_1  , C_ISO_8859_1},
1176         {"ga_IE"                , C_ISO_8859_1  , C_ISO_8859_1},
1177         {"gl_ES"                , C_ISO_8859_1  , C_ISO_8859_1},
1178         {"gv_GB"                , C_ISO_8859_1  , C_ISO_8859_1},
1179         {"id_ID"                , C_ISO_8859_1  , C_ISO_8859_1},
1180         {"is_IS"                , C_ISO_8859_1  , C_ISO_8859_1},
1181         {"it_CH"                , C_ISO_8859_1  , C_ISO_8859_1},
1182         {"it_IT"                , C_ISO_8859_1  , C_ISO_8859_1},
1183         {"kl_GL"                , C_ISO_8859_1  , C_ISO_8859_1},
1184         {"kw_GB"                , C_ISO_8859_1  , C_ISO_8859_1},
1185         {"ms_MY"                , C_ISO_8859_1  , C_ISO_8859_1},
1186         {"nl_BE"                , C_ISO_8859_1  , C_ISO_8859_1},
1187         {"nl_NL"                , C_ISO_8859_1  , C_ISO_8859_1},
1188         {"nb_NO"                , C_ISO_8859_1  , C_ISO_8859_1},
1189         {"nn_NO"                , C_ISO_8859_1  , C_ISO_8859_1},
1190         {"no_NO"                , C_ISO_8859_1  , C_ISO_8859_1},
1191         {"oc_FR"                , C_ISO_8859_1  , C_ISO_8859_1},
1192         {"pt_BR"                , C_ISO_8859_1  , C_ISO_8859_1},
1193         {"pt_PT"                , C_ISO_8859_1  , C_ISO_8859_1},
1194         {"sq_AL"                , C_ISO_8859_1  , C_ISO_8859_1},
1195         {"sv_FI"                , C_ISO_8859_1  , C_ISO_8859_1},
1196         {"sv_SE"                , C_ISO_8859_1  , C_ISO_8859_1},
1197         {"tl_PH"                , C_ISO_8859_1  , C_ISO_8859_1},
1198         {"uz_UZ"                , C_ISO_8859_1  , C_ISO_8859_1},
1199         {"wa_BE"                , C_ISO_8859_1  , C_ISO_8859_1},
1200
1201         {"bs_BA"                , C_ISO_8859_2  , C_ISO_8859_2},
1202         {"cs_CZ"                , C_ISO_8859_2  , C_ISO_8859_2},
1203         {"hr_HR"                , C_ISO_8859_2  , C_ISO_8859_2},
1204         {"hu_HU"                , C_ISO_8859_2  , C_ISO_8859_2},
1205         {"pl_PL"                , C_ISO_8859_2  , C_ISO_8859_2},
1206         {"ro_RO"                , C_ISO_8859_2  , C_ISO_8859_2},
1207         {"sk_SK"                , C_ISO_8859_2  , C_ISO_8859_2},
1208         {"sl_SI"                , C_ISO_8859_2  , C_ISO_8859_2},
1209
1210         {"sr_YU@cyrillic"       , C_ISO_8859_5  , C_ISO_8859_5},
1211         {"sr_YU"                , C_ISO_8859_2  , C_ISO_8859_2},
1212
1213         {"mt_MT"                , C_ISO_8859_3  , C_ISO_8859_3},
1214
1215         {"lt_LT.iso88594"       , C_ISO_8859_4  , C_ISO_8859_4},
1216         {"lt_LT.ISO8859-4"      , C_ISO_8859_4  , C_ISO_8859_4},
1217         {"lt_LT.ISO_8859-4"     , C_ISO_8859_4  , C_ISO_8859_4},
1218         {"lt_LT"                , C_ISO_8859_13 , C_ISO_8859_13},
1219
1220         {"mk_MK"                , C_ISO_8859_5  , C_ISO_8859_5},
1221
1222         {"ar_AE"                , C_ISO_8859_6  , C_ISO_8859_6},
1223         {"ar_BH"                , C_ISO_8859_6  , C_ISO_8859_6},
1224         {"ar_DZ"                , C_ISO_8859_6  , C_ISO_8859_6},
1225         {"ar_EG"                , C_ISO_8859_6  , C_ISO_8859_6},
1226         {"ar_IQ"                , C_ISO_8859_6  , C_ISO_8859_6},
1227         {"ar_JO"                , C_ISO_8859_6  , C_ISO_8859_6},
1228         {"ar_KW"                , C_ISO_8859_6  , C_ISO_8859_6},
1229         {"ar_LB"                , C_ISO_8859_6  , C_ISO_8859_6},
1230         {"ar_LY"                , C_ISO_8859_6  , C_ISO_8859_6},
1231         {"ar_MA"                , C_ISO_8859_6  , C_ISO_8859_6},
1232         {"ar_OM"                , C_ISO_8859_6  , C_ISO_8859_6},
1233         {"ar_QA"                , C_ISO_8859_6  , C_ISO_8859_6},
1234         {"ar_SA"                , C_ISO_8859_6  , C_ISO_8859_6},
1235         {"ar_SD"                , C_ISO_8859_6  , C_ISO_8859_6},
1236         {"ar_SY"                , C_ISO_8859_6  , C_ISO_8859_6},
1237         {"ar_TN"                , C_ISO_8859_6  , C_ISO_8859_6},
1238         {"ar_YE"                , C_ISO_8859_6  , C_ISO_8859_6},
1239
1240         {"el_GR"                , C_ISO_8859_7  , C_ISO_8859_7},
1241         {"he_IL"                , C_ISO_8859_8  , C_ISO_8859_8},
1242         {"iw_IL"                , C_ISO_8859_8  , C_ISO_8859_8},
1243         {"tr_TR"                , C_ISO_8859_9  , C_ISO_8859_9},
1244
1245         {"lv_LV"                , C_ISO_8859_13 , C_ISO_8859_13},
1246         {"mi_NZ"                , C_ISO_8859_13 , C_ISO_8859_13},
1247
1248         {"cy_GB"                , C_ISO_8859_14 , C_ISO_8859_14},
1249
1250         {"ar_IN"                , C_UTF_8       , C_UTF_8},
1251         {"en_IN"                , C_UTF_8       , C_UTF_8},
1252         {"se_NO"                , C_UTF_8       , C_UTF_8},
1253         {"ta_IN"                , C_UTF_8       , C_UTF_8},
1254         {"te_IN"                , C_UTF_8       , C_UTF_8},
1255         {"ur_PK"                , C_UTF_8       , C_UTF_8},
1256
1257         {"th_TH"                , C_TIS_620     , C_TIS_620},
1258         /* {"th_TH"             , C_WINDOWS_874}, */
1259         /* {"th_TH"             , C_ISO_8859_11}, */
1260
1261         {"ka_GE"                , C_GEORGIAN_PS , C_GEORGIAN_PS},
1262         {"vi_VN.TCVN"           , C_TCVN5712_1  , C_TCVN5712_1},
1263
1264         {"C"                    , C_US_ASCII    , C_US_ASCII},
1265         {"POSIX"                , C_US_ASCII    , C_US_ASCII},
1266         {"ANSI_X3.4-1968"       , C_US_ASCII    , C_US_ASCII},
1267 };
1268
1269 static GHashTable *conv_get_charset_to_str_table(void)
1270 {
1271         static GHashTable *table;
1272         gint i;
1273
1274         if (table)
1275                 return table;
1276
1277         table = g_hash_table_new(NULL, g_direct_equal);
1278
1279         for (i = 0; i < sizeof(charsets) / sizeof(charsets[0]); i++) {
1280                 if (g_hash_table_lookup(table, GUINT_TO_POINTER(charsets[i].charset))
1281                     == NULL) {
1282                         g_hash_table_insert
1283                                 (table, GUINT_TO_POINTER(charsets[i].charset),
1284                                  charsets[i].name);
1285                 }
1286         }
1287
1288         return table;
1289 }
1290
1291 static GHashTable *conv_get_charset_from_str_table(void)
1292 {
1293         static GHashTable *table;
1294         gint i;
1295
1296         if (table)
1297                 return table;
1298
1299         table = g_hash_table_new(str_case_hash, str_case_equal);
1300
1301         for (i = 0; i < sizeof(charsets) / sizeof(charsets[0]); i++) {
1302                 g_hash_table_insert(table, charsets[i].name,
1303                                     GUINT_TO_POINTER(charsets[i].charset));
1304         }
1305
1306         return table;
1307 }
1308
1309 const gchar *conv_get_charset_str(CharSet charset)
1310 {
1311         GHashTable *table;
1312
1313         table = conv_get_charset_to_str_table();
1314         return g_hash_table_lookup(table, GUINT_TO_POINTER(charset));
1315 }
1316
1317 CharSet conv_get_charset_from_str(const gchar *charset)
1318 {
1319         GHashTable *table;
1320
1321         if (!charset) return C_AUTO;
1322
1323         table = conv_get_charset_from_str_table();
1324         return GPOINTER_TO_UINT(g_hash_table_lookup(table, charset));
1325 }
1326
1327 static CharSet conv_get_locale_charset(void)
1328 {
1329         static CharSet cur_charset = C_UNINITIALIZED;
1330         const gchar *cur_locale;
1331         const gchar *p;
1332         gint i;
1333
1334         if (cur_charset != C_UNINITIALIZED)
1335                 return cur_charset;
1336
1337         cur_locale = conv_get_current_locale();
1338         if (!cur_locale) {
1339                 cur_charset = C_US_ASCII;
1340                 return cur_charset;
1341         }
1342
1343         if (strcasestr(cur_locale, "UTF-8") ||
1344             strcasestr(cur_locale, "utf8")) {
1345                 cur_charset = C_UTF_8;
1346                 return cur_charset;
1347         }
1348
1349         if ((p = strcasestr(cur_locale, "@euro")) && p[5] == '\0') {
1350                 cur_charset = C_ISO_8859_15;
1351                 return cur_charset;
1352         }
1353
1354         for (i = 0; i < sizeof(locale_table) / sizeof(locale_table[0]); i++) {
1355                 const gchar *p;
1356
1357                 /* "ja_JP.EUC" matches with "ja_JP.eucJP", "ja_JP.EUC" and
1358                    "ja_JP". "ja_JP" matches with "ja_JP.xxxx" and "ja" */
1359                 if (!g_ascii_strncasecmp(cur_locale, locale_table[i].locale,
1360                                  strlen(locale_table[i].locale))) {
1361                         cur_charset = locale_table[i].charset;
1362                         return cur_charset;
1363                 } else if ((p = strchr(locale_table[i].locale, '_')) &&
1364                          !strchr(p + 1, '.')) {
1365                         if (strlen(cur_locale) == 2 &&
1366                             !g_ascii_strncasecmp(cur_locale, locale_table[i].locale, 2)) {
1367                                 cur_charset = locale_table[i].charset;
1368                                 return cur_charset;
1369                         }
1370                 }
1371         }
1372
1373         cur_charset = C_AUTO;
1374         return cur_charset;
1375 }
1376
1377 static CharSet conv_get_locale_charset_no_utf8(void)
1378 {
1379         static CharSet cur_charset = C_UNINITIALIZED;
1380         const gchar *cur_locale;
1381         const gchar *p;
1382         gint i;
1383
1384         if (prefs_common.broken_are_utf8) {
1385                 cur_charset = C_UTF_8;
1386                 return cur_charset;
1387         }
1388
1389         cur_locale = conv_get_current_locale();
1390         if (!cur_locale) {
1391                 cur_charset = C_US_ASCII;
1392                 return cur_charset;
1393         }
1394
1395         if (strcasestr(cur_locale, "UTF-8") ||
1396             strcasestr(cur_locale, "utf8")) {
1397                 cur_charset = C_UTF_8;
1398                 return cur_charset;
1399         }
1400
1401         if ((p = strcasestr(cur_locale, "@euro")) && p[5] == '\0') {
1402                 cur_charset = C_ISO_8859_15;
1403                 return cur_charset;
1404         }
1405
1406         for (i = 0; i < sizeof(locale_table) / sizeof(locale_table[0]); i++) {
1407                 const gchar *p;
1408
1409                 /* "ja_JP.EUC" matches with "ja_JP.eucJP", "ja_JP.EUC" and
1410                    "ja_JP". "ja_JP" matches with "ja_JP.xxxx" and "ja" */
1411                 if (!g_ascii_strncasecmp(cur_locale, locale_table[i].locale,
1412                                  strlen(locale_table[i].locale))) {
1413                         cur_charset = locale_table[i].charset;
1414                         return cur_charset;
1415                 } else if ((p = strchr(locale_table[i].locale, '_')) &&
1416                          !strchr(p + 1, '.')) {
1417                         if (strlen(cur_locale) == 2 &&
1418                             !g_ascii_strncasecmp(cur_locale, locale_table[i].locale, 2)) {
1419                                 cur_charset = locale_table[i].charset;
1420                                 return cur_charset;
1421                         }
1422                 }
1423         }
1424
1425         cur_charset = C_AUTO;
1426         return cur_charset;
1427 }
1428
1429 const gchar *conv_get_locale_charset_str(void)
1430 {
1431         static const gchar *codeset = NULL;
1432
1433         if (!codeset)
1434                 codeset = conv_get_charset_str(conv_get_locale_charset());
1435
1436         return codeset ? codeset : CS_INTERNAL;
1437 }
1438
1439 const gchar *conv_get_locale_charset_str_no_utf8(void)
1440 {
1441         static const gchar *codeset = NULL;
1442
1443         if (!codeset)
1444                 codeset = conv_get_charset_str(conv_get_locale_charset_no_utf8());
1445
1446         return codeset ? codeset : CS_INTERNAL;
1447 }
1448
1449 static CharSet conv_get_outgoing_charset(void)
1450 {
1451         static CharSet out_charset = C_UNINITIALIZED;
1452         const gchar *cur_locale;
1453         const gchar *p;
1454         gint i;
1455
1456         if (out_charset != C_UNINITIALIZED)
1457                 return out_charset;
1458
1459         cur_locale = conv_get_current_locale();
1460         if (!cur_locale) {
1461                 out_charset = C_AUTO;
1462                 return out_charset;
1463         }
1464
1465         if (strcasestr(cur_locale, "UTF-8") ||
1466             strcasestr(cur_locale, "utf8")) {
1467                 out_charset = C_UTF_8;
1468                 return out_charset;
1469         }
1470
1471         if ((p = strcasestr(cur_locale, "@euro")) && p[5] == '\0') {
1472                 out_charset = C_ISO_8859_15;
1473                 return out_charset;
1474         }
1475
1476         for (i = 0; i < sizeof(locale_table) / sizeof(locale_table[0]); i++) {
1477                 const gchar *p;
1478
1479                 if (!g_ascii_strncasecmp(cur_locale, locale_table[i].locale,
1480                                  strlen(locale_table[i].locale))) {
1481                         out_charset = locale_table[i].out_charset;
1482                         break;
1483                 } else if ((p = strchr(locale_table[i].locale, '_')) &&
1484                          !strchr(p + 1, '.')) {
1485                         if (strlen(cur_locale) == 2 &&
1486                             !g_ascii_strncasecmp(cur_locale, locale_table[i].locale, 2)) {
1487                                 out_charset = locale_table[i].out_charset;
1488                                 break;
1489                         }
1490                 }
1491         }
1492
1493         return out_charset;
1494 }
1495
1496 const gchar *conv_get_outgoing_charset_str(void)
1497 {
1498         CharSet out_charset;
1499         const gchar *str;
1500
1501         out_charset = conv_get_outgoing_charset();
1502         str = conv_get_charset_str(out_charset);
1503
1504         return str ? str : CS_UTF_8;
1505 }
1506
1507 const gchar *conv_get_current_locale(void)
1508 {
1509         const gchar *cur_locale;
1510
1511 #ifdef G_OS_WIN32
1512         cur_locale = g_win32_getlocale();
1513 #else
1514         cur_locale = g_getenv("LC_ALL");
1515         if (!cur_locale) cur_locale = g_getenv("LC_CTYPE");
1516         if (!cur_locale) cur_locale = g_getenv("LANG");
1517         if (!cur_locale) cur_locale = setlocale(LC_CTYPE, NULL);
1518 #endif /* G_OS_WIN32 */
1519
1520         debug_print("current locale: %s\n",
1521                     cur_locale ? cur_locale : "(none)");
1522
1523         return cur_locale;
1524 }
1525
1526 static gboolean conv_is_ja_locale(void)
1527 {
1528         static gint is_ja_locale = -1;
1529         const gchar *cur_locale;
1530
1531         if (is_ja_locale != -1)
1532                 return is_ja_locale != 0;
1533
1534         is_ja_locale = 0;
1535         cur_locale = conv_get_current_locale();
1536         if (cur_locale) {
1537                 if (g_ascii_strncasecmp(cur_locale, "ja", 2) == 0)
1538                         is_ja_locale = 1;
1539         }
1540
1541         return is_ja_locale != 0;
1542 }
1543
1544 gchar *conv_unmime_header(const gchar *str, const gchar *default_encoding,
1545                            gboolean addr_field)
1546 {
1547         gchar buf[BUFFSIZE];
1548
1549         if (is_ascii_str(str))
1550                 return unmime_header(str, addr_field);
1551
1552         if (default_encoding) {
1553                 gchar *utf8_buf;
1554
1555                 utf8_buf = conv_codeset_strdup
1556                         (str, default_encoding, CS_INTERNAL);
1557                 if (utf8_buf) {
1558                         gchar *decoded_str;
1559
1560                         decoded_str = unmime_header(utf8_buf, addr_field);
1561                         g_free(utf8_buf);
1562                         return decoded_str;
1563                 }
1564         }
1565
1566         if (conv_is_ja_locale())
1567                 conv_anytodisp(buf, sizeof(buf), str);
1568         else
1569                 conv_localetodisp(buf, sizeof(buf), str);
1570
1571         return unmime_header(buf, addr_field);
1572 }
1573
1574 #define MAX_LINELEN             76
1575 #define MAX_HARD_LINELEN        996
1576 #define MIMESEP_BEGIN           "=?"
1577 #define MIMESEP_END             "?="
1578
1579 #define LBREAK_IF_REQUIRED(cond, is_plain_text)                         \
1580 {                                                                       \
1581         if (len - (destp - (guchar *)dest) < MAX_LINELEN + 2) {         \
1582                 *destp = '\0';                                          \
1583                 return;                                                 \
1584         }                                                               \
1585                                                                         \
1586         if ((cond) && *srcp) {                                          \
1587                 if (destp > (guchar *)dest && left < MAX_LINELEN - 1) { \
1588                         if (isspace(*(destp - 1)))                      \
1589                                 destp--;                                \
1590                         else if (is_plain_text && isspace(*srcp))       \
1591                                 srcp++;                                 \
1592                         if (*srcp) {                                    \
1593                                 *destp++ = '\n';                        \
1594                                 *destp++ = ' ';                         \
1595                                 left = MAX_LINELEN - 1;                 \
1596                         }                                               \
1597                 } else if (destp == (guchar *)dest && left < 7) {       \
1598                         if (is_plain_text && isspace(*srcp))            \
1599                                 srcp++;                                 \
1600                         if (*srcp) {                                    \
1601                                 *destp++ = '\n';                        \
1602                                 *destp++ = ' ';                         \
1603                                 left = MAX_LINELEN - 1;                 \
1604                         }                                               \
1605                 }                                                       \
1606         }                                                               \
1607 }
1608
1609 #define B64LEN(len) ((len) / 3 * 4 + ((len) % 3 ? 4 : 0))
1610
1611 void conv_encode_header_full(gchar *dest, gint len, const gchar *src,
1612                         gint header_len, gboolean addr_field,
1613                         const gchar *out_encoding_)
1614 {
1615         const gchar *cur_encoding;
1616         const gchar *out_encoding;
1617         gint mimestr_len;
1618         gchar *mimesep_enc;
1619         gint left;
1620         const guchar *srcp = src;
1621         guchar *destp = dest;
1622         gboolean use_base64;
1623
1624         cm_return_if_fail(g_utf8_validate(src, -1, NULL) == TRUE);
1625         cm_return_if_fail(destp != NULL);
1626
1627         if (MB_CUR_MAX > 1) {
1628                 use_base64 = TRUE;
1629                 mimesep_enc = "?B?";
1630         } else {
1631                 use_base64 = FALSE;
1632                 mimesep_enc = "?Q?";
1633         }
1634
1635         cur_encoding = CS_INTERNAL;
1636
1637         if (out_encoding_)
1638                 out_encoding = out_encoding_;
1639         else
1640                 out_encoding = conv_get_outgoing_charset_str();
1641
1642         if (!strcmp(out_encoding, CS_US_ASCII))
1643                 out_encoding = CS_ISO_8859_1;
1644
1645         mimestr_len = strlen(MIMESEP_BEGIN) + strlen(out_encoding) +
1646                 strlen(mimesep_enc) + strlen(MIMESEP_END);
1647
1648         left = MAX_LINELEN - header_len;
1649
1650         while (*srcp) {
1651                 LBREAK_IF_REQUIRED(left <= 0, TRUE);
1652
1653                 while (isspace(*srcp)) {
1654                         *destp++ = *srcp++;
1655                         left--;
1656                         LBREAK_IF_REQUIRED(left <= 0, TRUE);
1657                 }
1658
1659                 /* output as it is if the next word is ASCII string */
1660                 if (!is_next_nonascii(srcp)) {
1661                         gint word_len;
1662
1663                         word_len = get_next_word_len(srcp);
1664                         LBREAK_IF_REQUIRED(left < word_len, TRUE);
1665                         while (word_len > 0) {
1666                                 LBREAK_IF_REQUIRED(left + (MAX_HARD_LINELEN - MAX_LINELEN) <= 0, TRUE)
1667                                 *destp++ = *srcp++;
1668                                 left--;
1669                                 word_len--;
1670                         }
1671
1672                         continue;
1673                 }
1674
1675                 /* don't include parentheses and quotes in encoded strings */
1676                 if (addr_field && (*srcp == '(' || *srcp == ')' || *srcp == '"')) {
1677                         LBREAK_IF_REQUIRED(left < 2, FALSE);
1678                         *destp++ = *srcp++;
1679                         left--;
1680                 }
1681
1682                 while (1) {
1683                         gint mb_len = 0;
1684                         gint cur_len = 0;
1685                         gchar *part_str;
1686                         gchar *out_str;
1687                         gchar *enc_str;
1688                         const guchar *p = srcp;
1689                         gint out_str_len;
1690                         gint out_enc_str_len;
1691                         gint mime_block_len;
1692                         gboolean cont = FALSE;
1693
1694                         while (*p != '\0') {
1695                                 if (isspace(*p) && !is_next_nonascii(p + 1))
1696                                         break;
1697                                 /* don't include parentheses in encoded
1698                                    strings */
1699                                 if (addr_field && (*p == '(' || *p == ')' || *p == '"'))
1700                                         break;
1701
1702                                 mb_len = g_utf8_skip[*p];
1703
1704                                 Xstrndup_a(part_str, srcp, cur_len + mb_len, );
1705                                 out_str = conv_codeset_strdup
1706                                         (part_str, cur_encoding, out_encoding);
1707                                 if (!out_str) {
1708                                         if (strict_mode) {
1709                                                 *dest = '\0';
1710                                                 return;
1711                                         } else {
1712                                                 g_warning("conv_encode_header_full(): code conversion failed");
1713                                                 conv_unreadable_8bit(part_str);
1714                                                 out_str = g_strdup(part_str);
1715                                         }
1716                                 }
1717                                 out_str_len = strlen(out_str);
1718
1719                                 if (use_base64)
1720                                         out_enc_str_len = B64LEN(out_str_len);
1721                                 else
1722                                         out_enc_str_len =
1723                                                 qp_get_q_encoding_len(out_str);
1724
1725                                 g_free(out_str);
1726
1727                                 if (mimestr_len + out_enc_str_len <= left) {
1728                                         cur_len += mb_len;
1729                                         p += mb_len;
1730                                 } else if (cur_len == 0) {
1731                                         left = 0;
1732                                         LBREAK_IF_REQUIRED(1, FALSE);
1733                                         continue;
1734                                 } else {
1735                                         cont = TRUE;
1736                                         break;
1737                                 }
1738                         }
1739
1740                         if (cur_len > 0) {
1741                                 Xstrndup_a(part_str, srcp, cur_len, );
1742                                 out_str = conv_codeset_strdup
1743                                         (part_str, cur_encoding, out_encoding);
1744                                 if (!out_str) {
1745                                         g_warning("conv_encode_header_full(): code conversion failed");
1746                                         conv_unreadable_8bit(part_str);
1747                                         out_str = g_strdup(part_str);
1748                                 }
1749                                 out_str_len = strlen(out_str);
1750
1751                                 if (use_base64)
1752                                         out_enc_str_len = B64LEN(out_str_len);
1753                                 else
1754                                         out_enc_str_len =
1755                                                 qp_get_q_encoding_len(out_str);
1756
1757                                 if (use_base64)
1758                                         enc_str = g_base64_encode(out_str, out_str_len);
1759                                 else {
1760                                         Xalloca(enc_str, out_enc_str_len + 1, );
1761                                         qp_q_encode(enc_str, out_str);
1762                                 }
1763
1764                                 g_free(out_str);
1765
1766                                 /* output MIME-encoded string block */
1767                                 mime_block_len = mimestr_len + strlen(enc_str);
1768                                 g_snprintf(destp, mime_block_len + 1,
1769                                            MIMESEP_BEGIN "%s%s%s" MIMESEP_END,
1770                                            out_encoding, mimesep_enc, enc_str);
1771
1772                                 if (use_base64)
1773                                         g_free(enc_str);
1774
1775                                 destp += mime_block_len;
1776                                 srcp += cur_len;
1777
1778                                 left -= mime_block_len;
1779                         }
1780
1781                         LBREAK_IF_REQUIRED(cont, FALSE);
1782
1783                         if (cur_len == 0)
1784                                 break;
1785                 }
1786         }
1787
1788         *destp = '\0';
1789 }
1790
1791 void conv_encode_header(gchar *dest, gint len, const gchar *src,
1792                         gint header_len, gboolean addr_field)
1793 {
1794         conv_encode_header_full(dest,len,src,header_len,addr_field,NULL);
1795 }
1796
1797 #undef LBREAK_IF_REQUIRED
1798 #undef B64LEN
1799
1800 gchar *conv_filename_from_utf8(const gchar *utf8_file)
1801 {
1802         gchar *fs_file;
1803         GError *error = NULL;
1804
1805         fs_file = g_filename_from_utf8(utf8_file, -1, NULL, NULL, &error);
1806         if (error) {
1807                 debug_print("failed to convert encoding of file name: %s\n",
1808                           error->message);
1809                 g_error_free(error);
1810         }
1811         if (!fs_file)
1812                 fs_file = g_strdup(utf8_file);
1813
1814         return fs_file;
1815 }
1816
1817 gchar *conv_filename_to_utf8(const gchar *fs_file)
1818 {
1819         gchar *utf8_file = NULL;
1820         GError *error = NULL;
1821
1822         utf8_file = g_filename_to_utf8(fs_file, -1, NULL, NULL, &error);
1823         if (error) {
1824                 g_warning("failed to convert encoding of file name: %s",
1825                           error->message);
1826                 g_error_free(error);
1827         }
1828
1829         if (!utf8_file || !g_utf8_validate(utf8_file, -1, NULL)) {
1830                 g_free(utf8_file);
1831                 utf8_file = g_strdup(fs_file);
1832                 conv_unreadable_8bit(utf8_file);
1833         }
1834
1835         return utf8_file;
1836 }