src/codeconv.c

   1 /*
   2  * Sylpheed -- a GTK+ based, lightweight, and fast e-mail client
   3  * Copyright (C) 1999-2005 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 "defs.h"
  25
  26 #include <glib.h>
  27 #include <glib/gi18n.h>
  28 #include <string.h>
  29 #include <ctype.h>
  30 #include <stdlib.h>
  31 #include <errno.h>
  32
  33 #if HAVE_LOCALE_H
  34 #  include <locale.h>
  35 #endif
  36
  37 #include <iconv.h>
  38
  39 #include "codeconv.h"
  40 #include "unmime.h"
  41 #include "base64.h"
  42 #include "quoted-printable.h"
  43 #include "utils.h"
  44 #include "prefs_common.h"
  45
  46 typedef enum
  47 {
  48         JIS_ASCII,
  49         JIS_KANJI,
  50         JIS_HWKANA,
  51         JIS_AUXKANJI
  52 } JISState;
  53
  54 #define SUBST_CHAR      0x5f;
  55 #define ESC             '\033'
  56
  57 #define iseuckanji(c) \
  58         (((c) & 0xff) >= 0xa1 && ((c) & 0xff) <= 0xfe)
  59 #define iseuchwkana1(c) \
  60         (((c) & 0xff) == 0x8e)
  61 #define iseuchwkana2(c) \
  62         (((c) & 0xff) >= 0xa1 && ((c) & 0xff) <= 0xdf)
  63 #define iseucaux(c) \
  64         (((c) & 0xff) == 0x8f)
  65 #define issjiskanji1(c) \
  66         ((((c) & 0xff) >= 0x81 && ((c) & 0xff) <= 0x9f) || \
  67          (((c) & 0xff) >= 0xe0 && ((c) & 0xff) <= 0xfc))
  68 #define issjiskanji2(c) \
  69         ((((c) & 0xff) >= 0x40 && ((c) & 0xff) <= 0x7e) || \
  70          (((c) & 0xff) >= 0x80 && ((c) & 0xff) <= 0xfc))
  71 #define issjishwkana(c) \
  72         (((c) & 0xff) >= 0xa1 && ((c) & 0xff) <= 0xdf)
  73
  74 #define K_IN()                          \
  75         if (state != JIS_KANJI) {       \
  76                 *out++ = ESC;           \
  77                 *out++ = '$';           \
  78                 *out++ = 'B';           \
  79                 state = JIS_KANJI;      \
  80         }
  81
  82 #define K_OUT()                         \
  83         if (state != JIS_ASCII) {       \
  84                 *out++ = ESC;           \
  85                 *out++ = '(';           \
  86                 *out++ = 'B';           \
  87                 state = JIS_ASCII;      \
  88         }
  89
  90 #define HW_IN()                         \
  91         if (state != JIS_HWKANA) {      \
  92                 *out++ = ESC;           \
  93                 *out++ = '(';           \
  94                 *out++ = 'I';           \
  95                 state = JIS_HWKANA;     \
  96         }
  97
  98 #define AUX_IN()                        \
  99         if (state != JIS_AUXKANJI) {    \
 100                 *out++ = ESC;           \
 101                 *out++ = '$';           \
 102                 *out++ = '(';           \
 103                 *out++ = 'D';           \
 104                 state = JIS_AUXKANJI;   \
 105         }
 106
 107 static void conv_jistoeuc(gchar *outbuf, gint outlen, const gchar *inbuf);
 108 static void conv_euctojis(gchar *outbuf, gint outlen, const gchar *inbuf);
 109 static void conv_sjistoeuc(gchar *outbuf, gint outlen, const gchar *inbuf);
 110
 111 static void conv_jistoutf8(gchar *outbuf, gint outlen, const gchar *inbuf);
 112 static void conv_sjistoutf8(gchar *outbuf, gint outlen, const gchar *inbuf);
 113 static void conv_euctoutf8(gchar *outbuf, gint outlen, const gchar *inbuf);
 114 static void conv_anytoutf8(gchar *outbuf, gint outlen, const gchar *inbuf);
 115
 116 static void conv_utf8toeuc(gchar *outbuf, gint outlen, const gchar *inbuf);
 117 static void conv_utf8tojis(gchar *outbuf, gint outlen, const gchar *inbuf);
 118
 119 static void conv_unreadable_8bit(gchar *str);
 120
 121 static void conv_jistodisp(gchar *outbuf, gint outlen, const gchar *inbuf);
 122 static void conv_sjistodisp(gchar *outbuf, gint outlen, const gchar *inbuf);
 123 static void conv_euctodisp(gchar *outbuf, gint outlen, const gchar *inbuf);
 124
 125 static void conv_anytodisp(gchar *outbuf, gint outlen, const gchar *inbuf);
 126 static void conv_ustodisp(gchar *outbuf, gint outlen, const gchar *inbuf);
 127 static void conv_noconv(gchar *outbuf, gint outlen, const gchar *inbuf);
 128
 129 static void conv_jistoeuc(gchar *outbuf, gint outlen, const gchar *inbuf)
 130 {
 131         const guchar *in = inbuf;
 132         guchar *out = outbuf;
 133         JISState state = JIS_ASCII;
 134
 135         while (*in != '\0') {
 136                 if (*in == ESC) {
 137                         in++;
 138                         if (*in == '$') {
 139                                 if (*(in + 1) == '@' || *(in + 1) == 'B') {
 140                                         state = JIS_KANJI;
 141                                         in += 2;
 142                                 } else if (*(in + 1) == '(' &&
 143                                            *(in + 2) == 'D') {
 144                                         state = JIS_AUXKANJI;
 145                                         in += 3;
 146                                 } else {
 147                                         /* unknown escape sequence */
 148                                         state = JIS_ASCII;
 149                                 }
 150                         } else if (*in == '(') {
 151                                 if (*(in + 1) == 'B' || *(in + 1) == 'J') {
 152                                         state = JIS_ASCII;
 153                                         in += 2;
 154                                 } else if (*(in + 1) == 'I') {
 155                                         state = JIS_HWKANA;
 156                                         in += 2;
 157                                 } else {
 158                                         /* unknown escape sequence */
 159                                         state = JIS_ASCII;
 160                                 }
 161                         } else {
 162                                 /* unknown escape sequence */
 163                                 state = JIS_ASCII;
 164                         }
 165                 } else if (*in == 0x0e) {
 166                         state = JIS_HWKANA;
 167                         in++;
 168                 } else if (*in == 0x0f) {
 169                         state = JIS_ASCII;
 170                         in++;
 171                 } else {
 172                         switch (state) {
 173                         case JIS_ASCII:
 174                                 *out++ = *in++;
 175                                 break;
 176                         case JIS_KANJI:
 177                                 *out++ = *in++ | 0x80;
 178                                 if (*in == '\0') break;
 179                                 *out++ = *in++ | 0x80;
 180                                 break;
 181                         case JIS_HWKANA:
 182                                 *out++ = 0x8e;
 183                                 *out++ = *in++ | 0x80;
 184                                 break;
 185                         case JIS_AUXKANJI:
 186                                 *out++ = 0x8f;
 187                                 *out++ = *in++ | 0x80;
 188                                 if (*in == '\0') break;
 189                                 *out++ = *in++ | 0x80;
 190                                 break;
 191                         }
 192                 }
 193         }
 194
 195         *out = '\0';
 196 }
 197
 198 #define JIS_HWDAKUTEN           0x5e
 199 #define JIS_HWHANDAKUTEN        0x5f
 200
 201 static gint conv_jis_hantozen(guchar *outbuf, guchar jis_code, guchar sound_sym)
 202 {
 203         static guint16 h2z_tbl[] = {
 204                 /* 0x20 - 0x2f */
 205                 0x0000, 0x2123, 0x2156, 0x2157, 0x2122, 0x2126, 0x2572, 0x2521,
 206                 0x2523, 0x2525, 0x2527, 0x2529, 0x2563, 0x2565, 0x2567, 0x2543,
 207                 /* 0x30 - 0x3f */
 208                 0x213c, 0x2522, 0x2524, 0x2526, 0x2528, 0x252a, 0x252b, 0x252d,
 209                 0x252f, 0x2531, 0x2533, 0x2535, 0x2537, 0x2539, 0x253b, 0x253d,
 210                 /* 0x40 - 0x4f */
 211                 0x253f, 0x2541, 0x2544, 0x2546, 0x2548, 0x254a, 0x254b, 0x254c,
 212                 0x254d, 0x254e, 0x254f, 0x2552, 0x2555, 0x2558, 0x255b, 0x255e,
 213                 /* 0x50 - 0x5f */
 214                 0x255f, 0x2560, 0x2561, 0x2562, 0x2564, 0x2566, 0x2568, 0x2569,
 215                 0x256a, 0x256b, 0x256c, 0x256d, 0x256f, 0x2573, 0x212b, 0x212c
 216         };
 217
 218         static guint16 dakuten_tbl[] = {
 219                 /* 0x30 - 0x3f */
 220                 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x252c, 0x252e,
 221                 0x2530, 0x2532, 0x2534, 0x2536, 0x2538, 0x253a, 0x253c, 0x253e,
 222                 /* 0x40 - 0x4f */
 223                 0x2540, 0x2542, 0x2545, 0x2547, 0x2549, 0x0000, 0x0000, 0x0000,
 224                 0x0000, 0x0000, 0x2550, 0x2553, 0x2556, 0x2559, 0x255c, 0x0000
 225         };
 226
 227         static guint16 handakuten_tbl[] = {
 228                 /* 0x4a - 0x4e */
 229                 0x2551, 0x2554, 0x2557, 0x255a, 0x255d
 230         };
 231
 232         guint16 out_code;
 233
 234         jis_code &= 0x7f;
 235         sound_sym &= 0x7f;
 236
 237         if (jis_code < 0x21 || jis_code > 0x5f)
 238                 return 0;
 239
 240         if (sound_sym == JIS_HWDAKUTEN &&
 241             jis_code >= 0x36 && jis_code <= 0x4e) {
 242                 out_code = dakuten_tbl[jis_code - 0x30];
 243                 if (out_code != 0) {
 244                         *outbuf = out_code >> 8;
 245                         *(outbuf + 1) = out_code & 0xff;
 246                         return 2;
 247                 }
 248         }
 249
 250         if (sound_sym == JIS_HWHANDAKUTEN &&
 251             jis_code >= 0x4a && jis_code <= 0x4e) {
 252                 out_code = handakuten_tbl[jis_code - 0x4a];
 253                 *outbuf = out_code >> 8;
 254                 *(outbuf + 1) = out_code & 0xff;
 255                 return 2;
 256         }
 257
 258         out_code = h2z_tbl[jis_code - 0x20];
 259         *outbuf = out_code >> 8;
 260         *(outbuf + 1) = out_code & 0xff;
 261         return 1;
 262 }
 263
 264 static void conv_euctojis(gchar *outbuf, gint outlen, const gchar *inbuf)
 265 {
 266         const guchar *in = inbuf;
 267         guchar *out = outbuf;
 268         JISState state = JIS_ASCII;
 269
 270         while (*in != '\0') {
 271                 if (IS_ASCII(*in)) {
 272                         K_OUT();
 273                         *out++ = *in++;
 274                 } else if (iseuckanji(*in)) {
 275                         if (iseuckanji(*(in + 1))) {
 276                                 K_IN();
 277                                 *out++ = *in++ & 0x7f;
 278                                 *out++ = *in++ & 0x7f;
 279                         } else {
 280                                 K_OUT();
 281                                 *out++ = SUBST_CHAR;
 282                                 in++;
 283                                 if (*in != '\0' && !IS_ASCII(*in)) {
 284                                         *out++ = SUBST_CHAR;
 285                                         in++;
 286                                 }
 287                         }
 288                 } else if (iseuchwkana1(*in)) {
 289                         if (iseuchwkana2(*(in + 1))) {
 290                                 if (prefs_common.allow_jisx0201_kana) {
 291                                         HW_IN();
 292                                         in++;
 293                                         *out++ = *in++ & 0x7f;
 294                                 } else {
 295                                         guchar jis_ch[2];
 296                                         gint len;
 297
 298                                         if (iseuchwkana1(*(in + 2)) &&
 299                                             iseuchwkana2(*(in + 3)))
 300                                                 len = conv_jis_hantozen
 301                                                         (jis_ch,
 302                                                          *(in + 1), *(in + 3));
 303                                         else
 304                                                 len = conv_jis_hantozen
 305                                                         (jis_ch,
 306                                                          *(in + 1), '\0');
 307                                         if (len == 0)
 308                                                 in += 2;
 309                                         else {
 310                                                 K_IN();
 311                                                 in += len * 2;
 312                                                 *out++ = jis_ch[0];
 313                                                 *out++ = jis_ch[1];
 314                                         }
 315                                 }
 316                         } else {
 317                                 K_OUT();
 318                                 in++;
 319                                 if (*in != '\0' && !IS_ASCII(*in)) {
 320                                         *out++ = SUBST_CHAR;
 321                                         in++;
 322                                 }
 323                         }
 324                 } else if (iseucaux(*in)) {
 325                         in++;
 326                         if (iseuckanji(*in) && iseuckanji(*(in + 1))) {
 327                                 AUX_IN();
 328                                 *out++ = *in++ & 0x7f;
 329                                 *out++ = *in++ & 0x7f;
 330                         } else {
 331                                 K_OUT();
 332                                 if (*in != '\0' && !IS_ASCII(*in)) {
 333                                         *out++ = SUBST_CHAR;
 334                                         in++;
 335                                         if (*in != '\0' && !IS_ASCII(*in)) {
 336                                                 *out++ = SUBST_CHAR;
 337                                                 in++;
 338                                         }
 339                                 }
 340                         }
 341                 } else {
 342                         K_OUT();
 343                         *out++ = SUBST_CHAR;
 344                         in++;
 345                 }
 346         }
 347
 348         K_OUT();
 349         *out = '\0';
 350 }
 351
 352 static void conv_sjistoeuc(gchar *outbuf, gint outlen, const gchar *inbuf)
 353 {
 354         const guchar *in = inbuf;
 355         guchar *out = outbuf;
 356
 357         while (*in != '\0') {
 358                 if (IS_ASCII(*in)) {
 359                         *out++ = *in++;
 360                 } else if (issjiskanji1(*in)) {
 361                         if (issjiskanji2(*(in + 1))) {
 362                                 guchar out1 = *in;
 363                                 guchar out2 = *(in + 1);
 364                                 guchar row;
 365
 366                                 row = out1 < 0xa0 ? 0x70 : 0xb0;
 367                                 if (out2 < 0x9f) {
 368                                         out1 = (out1 - row) * 2 - 1;
 369                                         out2 -= out2 > 0x7f ? 0x20 : 0x1f;
 370                                 } else {
 371                                         out1 = (out1 - row) * 2;
 372                                         out2 -= 0x7e;
 373                                 }
 374
 375                                 *out++ = out1 | 0x80;
 376                                 *out++ = out2 | 0x80;
 377                                 in += 2;
 378                         } else {
 379                                 *out++ = SUBST_CHAR;
 380                                 in++;
 381                                 if (*in != '\0' && !IS_ASCII(*in)) {
 382                                         *out++ = SUBST_CHAR;
 383                                         in++;
 384                                 }
 385                         }
 386                 } else if (issjishwkana(*in)) {
 387                         *out++ = 0x8e;
 388                         *out++ = *in++;
 389                 } else {
 390                         *out++ = SUBST_CHAR;
 391                         in++;
 392                 }
 393         }
 394
 395         *out = '\0';
 396 }
 397
 398 static void conv_jistoutf8(gchar *outbuf, gint outlen, const gchar *inbuf)
 399 {
 400         gchar *eucstr;
 401
 402         Xalloca(eucstr, outlen, return);
 403
 404         conv_jistoeuc(eucstr, outlen, inbuf);
 405         conv_euctoutf8(outbuf, outlen, eucstr);
 406 }
 407
 408 static void conv_sjistoutf8(gchar *outbuf, gint outlen, const gchar *inbuf)
 409 {
 410         gchar *tmpstr;
 411
 412         tmpstr = conv_iconv_strdup(inbuf, CS_SHIFT_JIS, CS_UTF_8);
 413         if (tmpstr) {
 414                 strncpy2(outbuf, tmpstr, outlen);
 415                 g_free(tmpstr);
 416         } else
 417                 strncpy2(outbuf, inbuf, outlen);
 418 }
 419
 420 static void conv_euctoutf8(gchar *outbuf, gint outlen, const gchar *inbuf)
 421 {
 422         static iconv_t cd = (iconv_t)-1;
 423         static gboolean iconv_ok = TRUE;
 424         gchar *tmpstr;
 425
 426         if (cd == (iconv_t)-1) {
 427                 if (!iconv_ok) {
 428                         strncpy2(outbuf, inbuf, outlen);
 429                         return;
 430                 }
 431                 cd = iconv_open(CS_UTF_8, CS_EUC_JP_MS);
 432                 if (cd == (iconv_t)-1) {
 433                         cd = iconv_open(CS_UTF_8, CS_EUC_JP);
 434                         if (cd == (iconv_t)-1) {
 435                                 g_warning("conv_euctoutf8(): %s\n",
 436                                           g_strerror(errno));
 437                                 iconv_ok = FALSE;
 438                                 strncpy2(outbuf, inbuf, outlen);
 439                                 return;
 440                         }
 441                 }
 442         }
 443
 444         tmpstr = conv_iconv_strdup_with_cd(inbuf, cd);
 445         if (tmpstr) {
 446                 strncpy2(outbuf, tmpstr, outlen);
 447                 g_free(tmpstr);
 448         } else
 449                 strncpy2(outbuf, inbuf, outlen);
 450 }
 451
 452 static void conv_anytoutf8(gchar *outbuf, gint outlen, const gchar *inbuf)
 453 {
 454         switch (conv_guess_ja_encoding(inbuf)) {
 455         case C_ISO_2022_JP:
 456                 conv_jistoutf8(outbuf, outlen, inbuf);
 457                 break;
 458         case C_SHIFT_JIS:
 459                 conv_sjistoutf8(outbuf, outlen, inbuf);
 460                 break;
 461         case C_EUC_JP:
 462                 conv_euctoutf8(outbuf, outlen, inbuf);
 463                 break;
 464         default:
 465                 strncpy2(outbuf, inbuf, outlen);
 466                 break;
 467         }
 468 }
 469
 470 static void conv_utf8toeuc(gchar *outbuf, gint outlen, const gchar *inbuf)
 471 {
 472         static iconv_t cd = (iconv_t)-1;
 473         static gboolean iconv_ok = TRUE;
 474         gchar *tmpstr;
 475
 476         if (cd == (iconv_t)-1) {
 477                 if (!iconv_ok) {
 478                         strncpy2(outbuf, inbuf, outlen);
 479                         return;
 480                 }
 481                 cd = iconv_open(CS_EUC_JP_MS, CS_UTF_8);
 482                 if (cd == (iconv_t)-1) {
 483                         cd = iconv_open(CS_EUC_JP, CS_UTF_8);
 484                         if (cd == (iconv_t)-1) {
 485                                 g_warning("conv_utf8toeuc(): %s\n",
 486                                           g_strerror(errno));
 487                                 iconv_ok = FALSE;
 488                                 strncpy2(outbuf, inbuf, outlen);
 489                                 return;
 490                         }
 491                 }
 492         }
 493
 494         tmpstr = conv_iconv_strdup_with_cd(inbuf, cd);
 495         if (tmpstr) {
 496                 strncpy2(outbuf, tmpstr, outlen);
 497                 g_free(tmpstr);
 498         } else
 499                 strncpy2(outbuf, inbuf, outlen);
 500 }
 501
 502 static void conv_utf8tojis(gchar *outbuf, gint outlen, const gchar *inbuf)
 503 {
 504         gchar *eucstr;
 505
 506         Xalloca(eucstr, outlen, return);
 507
 508         conv_utf8toeuc(eucstr, outlen, inbuf);
 509         conv_euctojis(outbuf, outlen, eucstr);
 510 }
 511
 512 static void conv_unreadable_8bit(gchar *str)
 513 {
 514         register guchar *p = str;
 515
 516         while (*p != '\0') {
 517                 /* convert CR+LF -> LF */
 518                 if (*p == '\r' && *(p + 1) == '\n')
 519                         memmove(p, p + 1, strlen(p));
 520                 else if (!IS_ASCII(*p)) *p = SUBST_CHAR;
 521                 p++;
 522         }
 523 }
 524
 525 #define NCV     '\0'
 526
 527 void conv_mb_alnum(gchar *str)
 528 {
 529         static guchar char_tbl[] = {
 530                 /* 0xa0 - 0xaf */
 531                 NCV, ' ', NCV, NCV, ',', '.', NCV, ':',
 532                 ';', '?', '!', NCV, NCV, NCV, NCV, NCV,
 533                 /* 0xb0 - 0xbf */
 534                 NCV, NCV, NCV, NCV, NCV, NCV, NCV, NCV,
 535                 NCV, NCV, NCV, NCV, NCV, NCV, NCV, NCV,
 536                 /* 0xc0 - 0xcf */
 537                 NCV, NCV, NCV, NCV, NCV, NCV, NCV, NCV,
 538                 NCV, NCV, '(', ')', NCV, NCV, '[', ']',
 539                 /* 0xd0 - 0xdf */
 540                 '{', '}', NCV, NCV, NCV, NCV, NCV, NCV,
 541                 NCV, NCV, NCV, NCV, '+', '-', NCV, NCV,
 542                 /* 0xe0 - 0xef */
 543                 NCV, '=', NCV, '<', '>', NCV, NCV, NCV,
 544                 NCV, NCV, NCV, NCV, NCV, NCV, NCV, NCV
 545         };
 546
 547         register guchar *p = str;
 548         register gint len;
 549
 550         len = strlen(str);
 551
 552         while (len > 1) {
 553                 if (*p == 0xa3) {
 554                         register guchar ch = *(p + 1);
 555
 556                         if (ch >= 0xb0 && ch <= 0xfa) {
 557                                 /* [a-zA-Z] */
 558                                 *p = ch & 0x7f;
 559                                 p++;
 560                                 len--;
 561                                 memmove(p, p + 1, len);
 562                                 len--;
 563                         } else  {
 564                                 p += 2;
 565                                 len -= 2;
 566                         }
 567                 } else if (*p == 0xa1) {
 568                         register guchar ch = *(p + 1);
 569
 570                         if (ch >= 0xa0 && ch <= 0xef &&
 571                             NCV != char_tbl[ch - 0xa0]) {
 572                                 *p = char_tbl[ch - 0xa0];
 573                                 p++;
 574                                 len--;
 575                                 memmove(p, p + 1, len);
 576                                 len--;
 577                         } else {
 578                                 p += 2;
 579                                 len -= 2;
 580                         }
 581                 } else if (iseuckanji(*p)) {
 582                         p += 2;
 583                         len -= 2;
 584                 } else {
 585                         p++;
 586                         len--;
 587                 }
 588         }
 589 }
 590
 591 CharSet conv_guess_ja_encoding(const gchar *str)
 592 {
 593         const guchar *p = str;
 594         CharSet guessed = C_US_ASCII;
 595
 596         while (*p != '\0') {
 597                 if (*p == ESC && (*(p + 1) == '$' || *(p + 1) == '(')) {
 598                         if (guessed == C_US_ASCII)
 599                                 return C_ISO_2022_JP;
 600                         p += 2;
 601                 } else if (IS_ASCII(*p)) {
 602                         p++;
 603                 } else if (iseuckanji(*p) && iseuckanji(*(p + 1))) {
 604                         if (*p >= 0xfd && *p <= 0xfe)
 605                                 return C_EUC_JP;
 606                         else if (guessed == C_SHIFT_JIS) {
 607                                 if ((issjiskanji1(*p) &&
 608                                      issjiskanji2(*(p + 1))) ||
 609                                     issjishwkana(*p))
 610                                         guessed = C_SHIFT_JIS;
 611                                 else
 612                                         guessed = C_EUC_JP;
 613                         } else
 614                                 guessed = C_EUC_JP;
 615                         p += 2;
 616                 } else if (issjiskanji1(*p) && issjiskanji2(*(p + 1))) {
 617                         if (iseuchwkana1(*p) && iseuchwkana2(*(p + 1)))
 618                                 guessed = C_SHIFT_JIS;
 619                         else
 620                                 return C_SHIFT_JIS;
 621                         p += 2;
 622                 } else if (issjishwkana(*p)) {
 623                         guessed = C_SHIFT_JIS;
 624                         p++;
 625                 } else {
 626                         p++;
 627                 }
 628         }
 629
 630         return guessed;
 631 }
 632
 633 static void conv_jistodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
 634 {
 635         conv_jistoutf8(outbuf, outlen, inbuf);
 636 }
 637
 638 static void conv_sjistodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
 639 {
 640         conv_sjistoutf8(outbuf, outlen, inbuf);
 641 }
 642
 643 static void conv_euctodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
 644 {
 645         conv_euctoutf8(outbuf, outlen, inbuf);
 646 }
 647
 648 void conv_utf8todisp(gchar *outbuf, gint outlen, const gchar *inbuf)
 649 {
 650         if (g_utf8_validate(inbuf, -1, NULL) == TRUE)
 651                 strncpy2(outbuf, inbuf, outlen);
 652         else
 653                 conv_ustodisp(outbuf, outlen, inbuf);
 654 }
 655
 656 static void conv_anytodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
 657 {
 658         conv_anytoutf8(outbuf, outlen, inbuf);
 659         if (g_utf8_validate(outbuf, -1, NULL) != TRUE)
 660                 conv_unreadable_8bit(outbuf);
 661 }
 662
 663 static void conv_ustodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
 664 {
 665         strncpy2(outbuf, inbuf, outlen);
 666         conv_unreadable_8bit(outbuf);
 667 }
 668
 669 void conv_localetodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
 670 {
 671         gchar *tmpstr;
 672
 673         tmpstr = conv_iconv_strdup(inbuf, conv_get_locale_charset_str(),
 674                                    CS_INTERNAL);
 675         if (tmpstr) {
 676                 strncpy2(outbuf, tmpstr, outlen);
 677                 g_free(tmpstr);
 678         } else
 679                 conv_utf8todisp(outbuf, outlen, inbuf);
 680 }
 681
 682 static void conv_noconv(gchar *outbuf, gint outlen, const gchar *inbuf)
 683 {
 684         strncpy2(outbuf, inbuf, outlen);
 685 }
 686
 687 static const gchar *
 688 conv_get_fallback_for_private_encoding(const gchar *encoding)
 689 {
 690         if (encoding && (encoding[0] == 'X' || encoding[0] == 'x') &&
 691             encoding[1] == '-') {
 692                 if (!g_ascii_strcasecmp(encoding, CS_X_GBK))
 693                         return CS_GBK;
 694         }
 695
 696         return encoding;
 697 }
 698
 699 CodeConverter *conv_code_converter_new(const gchar *src_charset)
 700 {
 701         CodeConverter *conv;
 702
 703         src_charset = conv_get_fallback_for_private_encoding(src_charset);
 704
 705         conv = g_new0(CodeConverter, 1);
 706         conv->code_conv_func = conv_get_code_conv_func(src_charset, NULL);
 707         conv->charset_str = g_strdup(src_charset);
 708         conv->charset = conv_get_charset_from_str(src_charset);
 709
 710         return conv;
 711 }
 712
 713 void conv_code_converter_destroy(CodeConverter *conv)
 714 {
 715         g_free(conv->charset_str);
 716         g_free(conv);
 717 }
 718
 719 gint conv_convert(CodeConverter *conv, gchar *outbuf, gint outlen,
 720                   const gchar *inbuf)
 721 {
 722         if (conv->code_conv_func != conv_noconv)
 723                 conv->code_conv_func(outbuf, outlen, inbuf);
 724         else {
 725                 gchar *str;
 726
 727                 str = conv_iconv_strdup(inbuf, conv->charset_str, NULL);
 728                 if (!str)
 729                         return -1;
 730                 else {
 731                         strncpy2(outbuf, str, outlen);
 732                         g_free(str);
 733                 }
 734         }
 735
 736         return 0;
 737 }
 738
 739 gchar *conv_codeset_strdup(const gchar *inbuf,
 740                            const gchar *src_code, const gchar *dest_code)
 741 {
 742         gchar *buf;
 743         size_t len;
 744         CodeConvFunc conv_func;
 745
 746         src_code = conv_get_fallback_for_private_encoding(src_code);
 747         conv_func = conv_get_code_conv_func(src_code, dest_code);
 748         if (conv_func != conv_noconv) {
 749                 len = (strlen(inbuf) + 1) * 3;
 750                 buf = g_malloc(len);
 751                 if (!buf) return NULL;
 752
 753                 conv_func(buf, len, inbuf);
 754                 return g_realloc(buf, strlen(buf) + 1);
 755         }
 756
 757         return conv_iconv_strdup(inbuf, src_code, dest_code);
 758 }
 759
 760 CodeConvFunc conv_get_code_conv_func(const gchar *src_charset_str,
 761                                      const gchar *dest_charset_str)
 762 {
 763         CodeConvFunc code_conv = conv_noconv;
 764         CharSet src_charset;
 765         CharSet dest_charset;
 766
 767         if (!src_charset_str)
 768                 src_charset = conv_get_locale_charset();
 769         else
 770                 src_charset = conv_get_charset_from_str(src_charset_str);
 771
 772         /* auto detection mode */
 773         if (!src_charset_str && !dest_charset_str) {
 774                 if (conv_is_ja_locale())
 775                         return conv_anytodisp;
 776                 else
 777                         return conv_noconv;
 778         }
 779
 780         dest_charset = conv_get_charset_from_str(dest_charset_str);
 781
 782         if (dest_charset == C_US_ASCII)
 783                 return conv_ustodisp;
 784
 785         switch (src_charset) {
 786         case C_US_ASCII:
 787         case C_ISO_8859_1:
 788         case C_ISO_8859_2:
 789         case C_ISO_8859_3:
 790         case C_ISO_8859_4:
 791         case C_ISO_8859_5:
 792         case C_ISO_8859_6:
 793         case C_ISO_8859_7:
 794         case C_ISO_8859_8:
 795         case C_ISO_8859_9:
 796         case C_ISO_8859_10:
 797         case C_ISO_8859_11:
 798         case C_ISO_8859_13:
 799         case C_ISO_8859_14:
 800         case C_ISO_8859_15:
 801                 break;
 802         case C_ISO_2022_JP:
 803         case C_ISO_2022_JP_2:
 804         case C_ISO_2022_JP_3:
 805                 if (dest_charset == C_AUTO)
 806                         code_conv = conv_jistodisp;
 807                 else if (dest_charset == C_EUC_JP)
 808                         code_conv = conv_jistoeuc;
 809                 else if (dest_charset == C_UTF_8)
 810                         code_conv = conv_jistoutf8;
 811                 break;
 812         case C_SHIFT_JIS:
 813                 if (dest_charset == C_AUTO)
 814                         code_conv = conv_sjistodisp;
 815                 else if (dest_charset == C_EUC_JP)
 816                         code_conv = conv_sjistoeuc;
 817                 else if (dest_charset == C_UTF_8)
 818                         code_conv = conv_sjistoutf8;
 819                 break;
 820         case C_EUC_JP:
 821                 if (dest_charset == C_AUTO)
 822                         code_conv = conv_euctodisp;
 823                 else if (dest_charset == C_ISO_2022_JP   ||
 824                          dest_charset == C_ISO_2022_JP_2 ||
 825                          dest_charset == C_ISO_2022_JP_3)
 826                         code_conv = conv_euctojis;
 827                 else if (dest_charset == C_UTF_8)
 828                         code_conv = conv_euctoutf8;
 829                 break;
 830         case C_UTF_8:
 831                 if (dest_charset == C_EUC_JP)
 832                         code_conv = conv_utf8toeuc;
 833                 else if (dest_charset == C_ISO_2022_JP   ||
 834                          dest_charset == C_ISO_2022_JP_2 ||
 835                          dest_charset == C_ISO_2022_JP_3)
 836                         code_conv = conv_utf8tojis;
 837                 break;
 838         default:
 839                 break;
 840         }
 841
 842         return code_conv;
 843 }
 844
 845 gchar *conv_iconv_strdup(const gchar *inbuf,
 846                          const gchar *src_code, const gchar *dest_code)
 847 {
 848         iconv_t cd;
 849         gchar *outbuf;
 850
 851         if (!src_code)
 852                 src_code = conv_get_outgoing_charset_str();
 853         if (!dest_code)
 854                 dest_code = CS_INTERNAL;
 855
 856         /* don't convert if src and dest codeset are identical */
 857         if (!strcasecmp(src_code, dest_code))
 858                 return g_strdup(inbuf);
 859
 860         /* don't convert if current codeset is US-ASCII */
 861         if (!strcasecmp(dest_code, CS_US_ASCII))
 862                 return g_strdup(inbuf);
 863
 864         cd = iconv_open(dest_code, src_code);
 865         if (cd == (iconv_t)-1)
 866                 return NULL;
 867
 868         outbuf = conv_iconv_strdup_with_cd(inbuf, cd);
 869
 870         iconv_close(cd);
 871
 872         return outbuf;
 873 }
 874
 875 gchar *conv_iconv_strdup_with_cd(const gchar *inbuf, iconv_t cd)
 876 {
 877         const gchar *inbuf_p;
 878         gchar *outbuf;
 879         gchar *outbuf_p;
 880         size_t in_size;
 881         size_t in_left;
 882         size_t out_size;
 883         size_t out_left;
 884         size_t n_conv;
 885         size_t len;
 886
 887         inbuf_p = inbuf;
 888         in_size = strlen(inbuf);
 889         in_left = in_size;
 890         out_size = (in_size + 1) * 2;
 891         outbuf = g_malloc(out_size);
 892         outbuf_p = outbuf;
 893         out_left = out_size;
 894
 895 #define EXPAND_BUF()                            \
 896 {                                               \
 897         len = outbuf_p - outbuf;                \
 898         out_size *= 2;                          \
 899         outbuf = g_realloc(outbuf, out_size);   \
 900         outbuf_p = outbuf + len;                \
 901         out_left = out_size - len;              \
 902 }
 903
 904         while ((n_conv = iconv(cd, (ICONV_CONST gchar **)&inbuf_p, &in_left,
 905                                &outbuf_p, &out_left)) == (size_t)-1) {
 906                 if (EILSEQ == errno) {
 907                         //g_print("iconv(): at %d: %s\n", in_size - in_left, g_strerror(errno));
 908                         inbuf_p++;
 909                         in_left--;
 910                         if (out_left == 0) {
 911                                 EXPAND_BUF();
 912                         }
 913                         *outbuf_p++ = SUBST_CHAR;
 914                         out_left--;
 915                 } else if (EINVAL == errno) {
 916                         break;
 917                 } else if (E2BIG == errno) {
 918                         EXPAND_BUF();
 919                 } else {
 920                         g_warning("conv_iconv_strdup(): %s\n",
 921                                   g_strerror(errno));
 922                         break;
 923                 }
 924         }
 925
 926         while ((n_conv = iconv(cd, NULL, NULL, &outbuf_p, &out_left)) ==
 927                (size_t)-1) {
 928                 if (E2BIG == errno) {
 929                         EXPAND_BUF();
 930                 } else {
 931                         g_warning("conv_iconv_strdup(): %s\n",
 932                                   g_strerror(errno));
 933                         break;
 934                 }
 935         }
 936
 937 #undef EXPAND_BUF
 938
 939         len = outbuf_p - outbuf;
 940         outbuf = g_realloc(outbuf, len + 1);
 941         outbuf[len] = '\0';
 942
 943         return outbuf;
 944 }
 945
 946 static const struct {
 947         CharSet charset;
 948         gchar *const name;
 949 } charsets[] = {
 950         {C_US_ASCII,            CS_US_ASCII},
 951         {C_US_ASCII,            CS_ANSI_X3_4_1968},
 952         {C_UTF_8,               CS_UTF_8},
 953         {C_UTF_7,               CS_UTF_7},
 954         {C_ISO_8859_1,          CS_ISO_8859_1},
 955         {C_ISO_8859_2,          CS_ISO_8859_2},
 956         {C_ISO_8859_3,          CS_ISO_8859_3},
 957         {C_ISO_8859_4,          CS_ISO_8859_4},
 958         {C_ISO_8859_5,          CS_ISO_8859_5},
 959         {C_ISO_8859_6,          CS_ISO_8859_6},
 960         {C_ISO_8859_7,          CS_ISO_8859_7},
 961         {C_ISO_8859_8,          CS_ISO_8859_8},
 962         {C_ISO_8859_9,          CS_ISO_8859_9},
 963         {C_ISO_8859_10,         CS_ISO_8859_10},
 964         {C_ISO_8859_11,         CS_ISO_8859_11},
 965         {C_ISO_8859_13,         CS_ISO_8859_13},
 966         {C_ISO_8859_14,         CS_ISO_8859_14},
 967         {C_ISO_8859_15,         CS_ISO_8859_15},
 968         {C_BALTIC,              CS_BALTIC},
 969         {C_CP1250,              CS_CP1250},
 970         {C_CP1251,              CS_CP1251},
 971         {C_CP1252,              CS_CP1252},
 972         {C_CP1253,              CS_CP1253},
 973         {C_CP1254,              CS_CP1254},
 974         {C_CP1255,              CS_CP1255},
 975         {C_CP1256,              CS_CP1256},
 976         {C_CP1257,              CS_CP1257},
 977         {C_CP1258,              CS_CP1258},
 978         {C_WINDOWS_1250,        CS_WINDOWS_1250},
 979         {C_WINDOWS_1251,        CS_WINDOWS_1251},
 980         {C_WINDOWS_1252,        CS_WINDOWS_1252},
 981         {C_WINDOWS_1253,        CS_WINDOWS_1253},
 982         {C_WINDOWS_1254,        CS_WINDOWS_1254},
 983         {C_WINDOWS_1255,        CS_WINDOWS_1255},
 984         {C_WINDOWS_1256,        CS_WINDOWS_1256},
 985         {C_WINDOWS_1257,        CS_WINDOWS_1257},
 986         {C_WINDOWS_1258,        CS_WINDOWS_1258},
 987         {C_KOI8_R,              CS_KOI8_R},
 988         {C_KOI8_T,              CS_KOI8_T},
 989         {C_KOI8_U,              CS_KOI8_U},
 990         {C_ISO_2022_JP,         CS_ISO_2022_JP},
 991         {C_ISO_2022_JP_2,       CS_ISO_2022_JP_2},
 992         {C_ISO_2022_JP_3,       CS_ISO_2022_JP_3},
 993         {C_EUC_JP,              CS_EUC_JP},
 994         {C_EUC_JP,              CS_EUCJP},
 995         {C_EUC_JP_MS,           CS_EUC_JP_MS},
 996         {C_SHIFT_JIS,           CS_SHIFT_JIS},
 997         {C_SHIFT_JIS,           CS_SHIFT__JIS},
 998         {C_SHIFT_JIS,           CS_SJIS},
 999         {C_ISO_2022_KR,         CS_ISO_2022_KR},
1000         {C_EUC_KR,              CS_EUC_KR},
1001         {C_ISO_2022_CN,         CS_ISO_2022_CN},
1002         {C_EUC_CN,              CS_EUC_CN},
1003         {C_GB2312,              CS_GB2312},
1004         {C_GBK,                 CS_GBK},
1005         {C_EUC_TW,              CS_EUC_TW},
1006         {C_BIG5,                CS_BIG5},
1007         {C_BIG5_HKSCS,          CS_BIG5_HKSCS},
1008         {C_TIS_620,             CS_TIS_620},
1009         {C_WINDOWS_874,         CS_WINDOWS_874},
1010         {C_GEORGIAN_PS,         CS_GEORGIAN_PS},
1011         {C_TCVN5712_1,          CS_TCVN5712_1},
1012 };
1013
1014 static const struct {
1015         gchar *const locale;
1016         CharSet charset;
1017         CharSet out_charset;
1018 } locale_table[] = {
1019         {"ja_JP.eucJP"  , C_EUC_JP      , C_ISO_2022_JP},
1020         {"ja_JP.EUC-JP" , C_EUC_JP      , C_ISO_2022_JP},
1021         {"ja_JP.EUC"    , C_EUC_JP      , C_ISO_2022_JP},
1022         {"ja_JP.ujis"   , C_EUC_JP      , C_ISO_2022_JP},
1023         {"ja_JP.SJIS"   , C_SHIFT_JIS   , C_ISO_2022_JP},
1024         {"ja_JP.JIS"    , C_ISO_2022_JP , C_ISO_2022_JP},
1025         {"ja_JP"        , C_EUC_JP      , C_ISO_2022_JP},
1026         {"ko_KR.EUC-KR" , C_EUC_KR      , C_EUC_KR},
1027         {"ko_KR"        , C_EUC_KR      , C_EUC_KR},
1028         {"zh_CN.GB2312" , C_GB2312      , C_GB2312},
1029         {"zh_CN.GBK"    , C_GBK         , C_GBK},
1030         {"zh_CN"        , C_GB2312      , C_GB2312},
1031         {"zh_HK"        , C_BIG5_HKSCS  , C_BIG5_HKSCS},
1032         {"zh_TW.eucTW"  , C_EUC_TW      , C_BIG5},
1033         {"zh_TW.EUC-TW" , C_EUC_TW      , C_BIG5},
1034         {"zh_TW.Big5"   , C_BIG5        , C_BIG5},
1035         {"zh_TW"        , C_BIG5        , C_BIG5},
1036
1037         {"ru_RU.KOI8-R" , C_KOI8_R      , C_KOI8_R},
1038         {"ru_RU.KOI8R"  , C_KOI8_R      , C_KOI8_R},
1039         {"ru_RU.CP1251" , C_WINDOWS_1251, C_KOI8_R},
1040         {"ru_RU"        , C_ISO_8859_5  , C_KOI8_R},
1041         {"tg_TJ"        , C_KOI8_T      , C_KOI8_T},
1042         {"ru_UA"        , C_KOI8_U      , C_KOI8_U},
1043         {"uk_UA.CP1251" , C_WINDOWS_1251, C_KOI8_U},
1044         {"uk_UA"        , C_KOI8_U      , C_KOI8_U},
1045
1046         {"be_BY"        , C_WINDOWS_1251, C_WINDOWS_1251},
1047         {"bg_BG"        , C_WINDOWS_1251, C_WINDOWS_1251},
1048
1049         {"yi_US"        , C_WINDOWS_1255, C_WINDOWS_1255},
1050
1051         {"af_ZA"        , C_ISO_8859_1  , C_ISO_8859_1},
1052         {"br_FR"        , C_ISO_8859_1  , C_ISO_8859_1},
1053         {"ca_ES"        , C_ISO_8859_1  , C_ISO_8859_1},
1054         {"da_DK"        , C_ISO_8859_1  , C_ISO_8859_1},
1055         {"de_AT"        , C_ISO_8859_1  , C_ISO_8859_1},
1056         {"de_BE"        , C_ISO_8859_1  , C_ISO_8859_1},
1057         {"de_CH"        , C_ISO_8859_1  , C_ISO_8859_1},
1058         {"de_DE"        , C_ISO_8859_1  , C_ISO_8859_1},
1059         {"de_LU"        , C_ISO_8859_1  , C_ISO_8859_1},
1060         {"en_AU"        , C_ISO_8859_1  , C_ISO_8859_1},
1061         {"en_BW"        , C_ISO_8859_1  , C_ISO_8859_1},
1062         {"en_CA"        , C_ISO_8859_1  , C_ISO_8859_1},
1063         {"en_DK"        , C_ISO_8859_1  , C_ISO_8859_1},
1064         {"en_GB"        , C_ISO_8859_1  , C_ISO_8859_1},
1065         {"en_HK"        , C_ISO_8859_1  , C_ISO_8859_1},
1066         {"en_IE"        , C_ISO_8859_1  , C_ISO_8859_1},
1067         {"en_NZ"        , C_ISO_8859_1  , C_ISO_8859_1},
1068         {"en_PH"        , C_ISO_8859_1  , C_ISO_8859_1},
1069         {"en_SG"        , C_ISO_8859_1  , C_ISO_8859_1},
1070         {"en_US"        , C_ISO_8859_1  , C_ISO_8859_1},
1071         {"en_ZA"        , C_ISO_8859_1  , C_ISO_8859_1},
1072         {"en_ZW"        , C_ISO_8859_1  , C_ISO_8859_1},
1073         {"es_AR"        , C_ISO_8859_1  , C_ISO_8859_1},
1074         {"es_BO"        , C_ISO_8859_1  , C_ISO_8859_1},
1075         {"es_CL"        , C_ISO_8859_1  , C_ISO_8859_1},
1076         {"es_CO"        , C_ISO_8859_1  , C_ISO_8859_1},
1077         {"es_CR"        , C_ISO_8859_1  , C_ISO_8859_1},
1078         {"es_DO"        , C_ISO_8859_1  , C_ISO_8859_1},
1079         {"es_EC"        , C_ISO_8859_1  , C_ISO_8859_1},
1080         {"es_ES"        , C_ISO_8859_1  , C_ISO_8859_1},
1081         {"es_GT"        , C_ISO_8859_1  , C_ISO_8859_1},
1082         {"es_HN"        , C_ISO_8859_1  , C_ISO_8859_1},
1083         {"es_MX"        , C_ISO_8859_1  , C_ISO_8859_1},
1084         {"es_NI"        , C_ISO_8859_1  , C_ISO_8859_1},
1085         {"es_PA"        , C_ISO_8859_1  , C_ISO_8859_1},
1086         {"es_PE"        , C_ISO_8859_1  , C_ISO_8859_1},
1087         {"es_PR"        , C_ISO_8859_1  , C_ISO_8859_1},
1088         {"es_PY"        , C_ISO_8859_1  , C_ISO_8859_1},
1089         {"es_SV"        , C_ISO_8859_1  , C_ISO_8859_1},
1090         {"es_US"        , C_ISO_8859_1  , C_ISO_8859_1},
1091         {"es_UY"        , C_ISO_8859_1  , C_ISO_8859_1},
1092         {"es_VE"        , C_ISO_8859_1  , C_ISO_8859_1},
1093         {"et_EE"        , C_ISO_8859_1  , C_ISO_8859_1},
1094         {"eu_ES"        , C_ISO_8859_1  , C_ISO_8859_1},
1095         {"fi_FI"        , C_ISO_8859_1  , C_ISO_8859_1},
1096         {"fo_FO"        , C_ISO_8859_1  , C_ISO_8859_1},
1097         {"fr_BE"        , C_ISO_8859_1  , C_ISO_8859_1},
1098         {"fr_CA"        , C_ISO_8859_1  , C_ISO_8859_1},
1099         {"fr_CH"        , C_ISO_8859_1  , C_ISO_8859_1},
1100         {"fr_FR"        , C_ISO_8859_1  , C_ISO_8859_1},
1101         {"fr_LU"        , C_ISO_8859_1  , C_ISO_8859_1},
1102         {"ga_IE"        , C_ISO_8859_1  , C_ISO_8859_1},
1103         {"gl_ES"        , C_ISO_8859_1  , C_ISO_8859_1},
1104         {"gv_GB"        , C_ISO_8859_1  , C_ISO_8859_1},
1105         {"id_ID"        , C_ISO_8859_1  , C_ISO_8859_1},
1106         {"is_IS"        , C_ISO_8859_1  , C_ISO_8859_1},
1107         {"it_CH"        , C_ISO_8859_1  , C_ISO_8859_1},
1108         {"it_IT"        , C_ISO_8859_1  , C_ISO_8859_1},
1109         {"kl_GL"        , C_ISO_8859_1  , C_ISO_8859_1},
1110         {"kw_GB"        , C_ISO_8859_1  , C_ISO_8859_1},
1111         {"ms_MY"        , C_ISO_8859_1  , C_ISO_8859_1},
1112         {"nl_BE"        , C_ISO_8859_1  , C_ISO_8859_1},
1113         {"nl_NL"        , C_ISO_8859_1  , C_ISO_8859_1},
1114         {"nn_NO"        , C_ISO_8859_1  , C_ISO_8859_1},
1115         {"no_NO"        , C_ISO_8859_1  , C_ISO_8859_1},
1116         {"oc_FR"        , C_ISO_8859_1  , C_ISO_8859_1},
1117         {"pt_BR"        , C_ISO_8859_1  , C_ISO_8859_1},
1118         {"pt_PT"        , C_ISO_8859_1  , C_ISO_8859_1},
1119         {"sq_AL"        , C_ISO_8859_1  , C_ISO_8859_1},
1120         {"sv_FI"        , C_ISO_8859_1  , C_ISO_8859_1},
1121         {"sv_SE"        , C_ISO_8859_1  , C_ISO_8859_1},
1122         {"tl_PH"        , C_ISO_8859_1  , C_ISO_8859_1},
1123         {"uz_UZ"        , C_ISO_8859_1  , C_ISO_8859_1},
1124         {"wa_BE"        , C_ISO_8859_1  , C_ISO_8859_1},
1125
1126         {"bs_BA"        , C_ISO_8859_2  , C_ISO_8859_2},
1127         {"cs_CZ"        , C_ISO_8859_2  , C_ISO_8859_2},
1128         {"hr_HR"        , C_ISO_8859_2  , C_ISO_8859_2},
1129         {"hu_HU"        , C_ISO_8859_2  , C_ISO_8859_2},
1130         {"pl_PL"        , C_ISO_8859_2  , C_ISO_8859_2},
1131         {"ro_RO"        , C_ISO_8859_2  , C_ISO_8859_2},
1132         {"sk_SK"        , C_ISO_8859_2  , C_ISO_8859_2},
1133         {"sl_SI"        , C_ISO_8859_2  , C_ISO_8859_2},
1134
1135         {"sr_YU@cyrillic"       , C_ISO_8859_5  , C_ISO_8859_5},
1136         {"sr_YU"                , C_ISO_8859_2  , C_ISO_8859_2},
1137
1138         {"mt_MT"                , C_ISO_8859_3  , C_ISO_8859_3},
1139
1140         {"lt_LT.iso88594"       , C_ISO_8859_4  , C_ISO_8859_4},
1141         {"lt_LT.ISO8859-4"      , C_ISO_8859_4  , C_ISO_8859_4},
1142         {"lt_LT.ISO_8859-4"     , C_ISO_8859_4  , C_ISO_8859_4},
1143         {"lt_LT"                , C_ISO_8859_13 , C_ISO_8859_13},
1144
1145         {"mk_MK"        , C_ISO_8859_5  , C_ISO_8859_5},
1146
1147         {"ar_AE"        , C_ISO_8859_6  , C_ISO_8859_6},
1148         {"ar_BH"        , C_ISO_8859_6  , C_ISO_8859_6},
1149         {"ar_DZ"        , C_ISO_8859_6  , C_ISO_8859_6},
1150         {"ar_EG"        , C_ISO_8859_6  , C_ISO_8859_6},
1151         {"ar_IQ"        , C_ISO_8859_6  , C_ISO_8859_6},
1152         {"ar_JO"        , C_ISO_8859_6  , C_ISO_8859_6},
1153         {"ar_KW"        , C_ISO_8859_6  , C_ISO_8859_6},
1154         {"ar_LB"        , C_ISO_8859_6  , C_ISO_8859_6},
1155         {"ar_LY"        , C_ISO_8859_6  , C_ISO_8859_6},
1156         {"ar_MA"        , C_ISO_8859_6  , C_ISO_8859_6},
1157         {"ar_OM"        , C_ISO_8859_6  , C_ISO_8859_6},
1158         {"ar_QA"        , C_ISO_8859_6  , C_ISO_8859_6},
1159         {"ar_SA"        , C_ISO_8859_6  , C_ISO_8859_6},
1160         {"ar_SD"        , C_ISO_8859_6  , C_ISO_8859_6},
1161         {"ar_SY"        , C_ISO_8859_6  , C_ISO_8859_6},
1162         {"ar_TN"        , C_ISO_8859_6  , C_ISO_8859_6},
1163         {"ar_YE"        , C_ISO_8859_6  , C_ISO_8859_6},
1164
1165         {"el_GR"        , C_ISO_8859_7  , C_ISO_8859_7},
1166         {"he_IL"        , C_ISO_8859_8  , C_ISO_8859_8},
1167         {"iw_IL"        , C_ISO_8859_8  , C_ISO_8859_8},
1168         {"tr_TR"        , C_ISO_8859_9  , C_ISO_8859_9},
1169
1170         {"lv_LV"        , C_ISO_8859_13 , C_ISO_8859_13},
1171         {"mi_NZ"        , C_ISO_8859_13 , C_ISO_8859_13},
1172
1173         {"cy_GB"        , C_ISO_8859_14 , C_ISO_8859_14},
1174
1175         {"ar_IN"        , C_UTF_8       , C_UTF_8},
1176         {"en_IN"        , C_UTF_8       , C_UTF_8},
1177         {"se_NO"        , C_UTF_8       , C_UTF_8},
1178         {"ta_IN"        , C_UTF_8       , C_UTF_8},
1179         {"te_IN"        , C_UTF_8       , C_UTF_8},
1180         {"ur_PK"        , C_UTF_8       , C_UTF_8},
1181
1182         {"th_TH"        , C_TIS_620     , C_TIS_620},
1183         /* {"th_TH"     , C_WINDOWS_874}, */
1184         /* {"th_TH"     , C_ISO_8859_11}, */
1185
1186         {"ka_GE"        , C_GEORGIAN_PS , C_GEORGIAN_PS},
1187         {"vi_VN.TCVN"   , C_TCVN5712_1  , C_TCVN5712_1},
1188
1189         {"C"                    , C_US_ASCII    , C_US_ASCII},
1190         {"POSIX"                , C_US_ASCII    , C_US_ASCII},
1191         {"ANSI_X3.4-1968"       , C_US_ASCII    , C_US_ASCII},
1192 };
1193
1194 static GHashTable *conv_get_charset_to_str_table(void)
1195 {
1196         static GHashTable *table;
1197         gint i;
1198
1199         if (table)
1200                 return table;
1201
1202         table = g_hash_table_new(NULL, g_direct_equal);
1203
1204         for (i = 0; i < sizeof(charsets) / sizeof(charsets[0]); i++) {
1205                 if (g_hash_table_lookup(table, GUINT_TO_POINTER(charsets[i].charset))
1206                     == NULL) {
1207                         g_hash_table_insert
1208                                 (table, GUINT_TO_POINTER(charsets[i].charset),
1209                                  charsets[i].name);
1210                 }
1211         }
1212
1213         return table;
1214 }
1215
1216 static GHashTable *conv_get_charset_from_str_table(void)
1217 {
1218         static GHashTable *table;
1219         gint i;
1220
1221         if (table)
1222                 return table;
1223
1224         table = g_hash_table_new(str_case_hash, str_case_equal);
1225
1226         for (i = 0; i < sizeof(charsets) / sizeof(charsets[0]); i++) {
1227                 g_hash_table_insert(table, charsets[i].name,
1228                                     GUINT_TO_POINTER(charsets[i].charset));
1229         }
1230
1231         return table;
1232 }
1233
1234 const gchar *conv_get_charset_str(CharSet charset)
1235 {
1236         GHashTable *table;
1237
1238         table = conv_get_charset_to_str_table();
1239         return g_hash_table_lookup(table, GUINT_TO_POINTER(charset));
1240 }
1241
1242 CharSet conv_get_charset_from_str(const gchar *charset)
1243 {
1244         GHashTable *table;
1245
1246         if (!charset) return C_AUTO;
1247
1248         table = conv_get_charset_from_str_table();
1249         return GPOINTER_TO_UINT(g_hash_table_lookup(table, charset));
1250 }
1251
1252 CharSet conv_get_locale_charset(void)
1253 {
1254         static CharSet cur_charset = -1;
1255         const gchar *cur_locale;
1256         const gchar *p;
1257         gint i;
1258
1259         if (cur_charset != -1)
1260                 return cur_charset;
1261
1262         cur_locale = conv_get_current_locale();
1263         if (!cur_locale) {
1264                 cur_charset = C_US_ASCII;
1265                 return cur_charset;
1266         }
1267
1268         if (strcasestr(cur_locale, "UTF-8")) {
1269                 cur_charset = C_UTF_8;
1270                 return cur_charset;
1271         }
1272
1273         if ((p = strcasestr(cur_locale, "@euro")) && p[5] == '\0') {
1274                 cur_charset = C_ISO_8859_15;
1275                 return cur_charset;
1276         }
1277
1278         for (i = 0; i < sizeof(locale_table) / sizeof(locale_table[0]); i++) {
1279                 const gchar *p;
1280
1281                 /* "ja_JP.EUC" matches with "ja_JP.eucJP", "ja_JP.EUC" and
1282                    "ja_JP". "ja_JP" matches with "ja_JP.xxxx" and "ja" */
1283                 if (!g_ascii_strncasecmp(cur_locale, locale_table[i].locale,
1284                                  strlen(locale_table[i].locale))) {
1285                         cur_charset = locale_table[i].charset;
1286                         return cur_charset;
1287                 } else if ((p = strchr(locale_table[i].locale, '_')) &&
1288                          !strchr(p + 1, '.')) {
1289                         if (strlen(cur_locale) == 2 &&
1290                             !g_ascii_strncasecmp(cur_locale, locale_table[i].locale, 2)) {
1291                                 cur_charset = locale_table[i].charset;
1292                                 return cur_charset;
1293                         }
1294                 }
1295         }
1296
1297         cur_charset = C_AUTO;
1298         return cur_charset;
1299 }
1300
1301 const gchar *conv_get_locale_charset_str(void)
1302 {
1303         static const gchar *codeset = NULL;
1304
1305         if (!codeset)
1306                 codeset = conv_get_charset_str(conv_get_locale_charset());
1307
1308         return codeset ? codeset : CS_INTERNAL;
1309 }
1310
1311 CharSet conv_get_internal_charset(void)
1312 {
1313         return C_INTERNAL;
1314 }
1315
1316 const gchar *conv_get_internal_charset_str(void)
1317 {
1318         return CS_INTERNAL;
1319 }
1320
1321 CharSet conv_get_outgoing_charset(void)
1322 {
1323         static CharSet out_charset = -1;
1324         const gchar *cur_locale;
1325         const gchar *p;
1326         gint i;
1327
1328         if (out_charset != -1)
1329                 return out_charset;
1330
1331         cur_locale = conv_get_current_locale();
1332         if (!cur_locale) {
1333                 out_charset = C_AUTO;
1334                 return out_charset;
1335         }
1336
1337         if (strcasestr(cur_locale, "UTF-8")) {
1338                 out_charset = C_UTF_8;
1339                 return out_charset;
1340         }
1341
1342         if ((p = strcasestr(cur_locale, "@euro")) && p[5] == '\0') {
1343                 out_charset = C_ISO_8859_15;
1344                 return out_charset;
1345         }
1346
1347         for (i = 0; i < sizeof(locale_table) / sizeof(locale_table[0]); i++) {
1348                 const gchar *p;
1349
1350                 if (!g_ascii_strncasecmp(cur_locale, locale_table[i].locale,
1351                                  strlen(locale_table[i].locale))) {
1352                         out_charset = locale_table[i].out_charset;
1353                         break;
1354                 } else if ((p = strchr(locale_table[i].locale, '_')) &&
1355                          !strchr(p + 1, '.')) {
1356                         if (strlen(cur_locale) == 2 &&
1357                             !g_ascii_strncasecmp(cur_locale, locale_table[i].locale, 2)) {
1358                                 out_charset = locale_table[i].out_charset;
1359                                 break;
1360                         }
1361                 }
1362         }
1363
1364         return out_charset;
1365 }
1366
1367 const gchar *conv_get_outgoing_charset_str(void)
1368 {
1369         CharSet out_charset;
1370         const gchar *str;
1371
1372         out_charset = conv_get_outgoing_charset();
1373         str = conv_get_charset_str(out_charset);
1374
1375         return str ? str : CS_UTF_8;
1376 }
1377
1378 gboolean conv_is_multibyte_encoding(CharSet encoding)
1379 {
1380         switch (encoding) {
1381         case C_EUC_JP:
1382         case C_EUC_JP_MS:
1383         case C_EUC_KR:
1384         case C_EUC_TW:
1385         case C_EUC_CN:
1386         case C_ISO_2022_JP:
1387         case C_ISO_2022_JP_2:
1388         case C_ISO_2022_JP_3:
1389         case C_ISO_2022_KR:
1390         case C_ISO_2022_CN:
1391         case C_SHIFT_JIS:
1392         case C_GB2312:
1393         case C_GBK:
1394         case C_BIG5:
1395         case C_UTF_8:
1396         case C_UTF_7:
1397                 return TRUE;
1398         default:
1399                 return FALSE;
1400         }
1401 }
1402
1403 const gchar *conv_get_current_locale(void)
1404 {
1405         const gchar *cur_locale;
1406
1407         cur_locale = g_getenv("LC_ALL");
1408         if (!cur_locale) cur_locale = g_getenv("LC_CTYPE");
1409         if (!cur_locale) cur_locale = g_getenv("LANG");
1410         if (!cur_locale) cur_locale = setlocale(LC_CTYPE, NULL);
1411
1412         debug_print("current locale: %s\n",
1413                     cur_locale ? cur_locale : "(none)");
1414
1415         return cur_locale;
1416 }
1417
1418 gboolean conv_is_ja_locale(void)
1419 {
1420         static gint is_ja_locale = -1;
1421         const gchar *cur_locale;
1422
1423         if (is_ja_locale != -1)
1424                 return is_ja_locale != 0;
1425
1426         is_ja_locale = 0;
1427         cur_locale = conv_get_current_locale();
1428         if (cur_locale) {
1429                 if (g_ascii_strncasecmp(cur_locale, "ja", 2) == 0)
1430                         is_ja_locale = 1;
1431         }
1432
1433         return is_ja_locale != 0;
1434 }
1435
1436 gchar *conv_unmime_header(const gchar *str, const gchar *default_encoding)
1437 {
1438         gchar buf[BUFFSIZE];
1439
1440         if (is_ascii_str(str))
1441                 return unmime_header(str);
1442
1443         if (default_encoding) {
1444                 gchar *utf8_buf;
1445
1446                 utf8_buf = conv_codeset_strdup
1447                         (str, default_encoding, CS_INTERNAL);
1448                 if (utf8_buf) {
1449                         gchar *decoded_str;
1450
1451                         decoded_str = unmime_header(utf8_buf);
1452                         g_free(utf8_buf);
1453                         return decoded_str;
1454                 }
1455         }
1456
1457         if (conv_is_ja_locale())
1458                 conv_anytodisp(buf, sizeof(buf), str);
1459         else
1460                 conv_localetodisp(buf, sizeof(buf), str);
1461
1462         return unmime_header(buf);
1463 }
1464
1465 #define MAX_LINELEN             76
1466 #define MAX_HARD_LINELEN        996
1467 #define MIMESEP_BEGIN           "=?"
1468 #define MIMESEP_END             "?="
1469
1470 #define LBREAK_IF_REQUIRED(cond, is_plain_text)                         \
1471 {                                                                       \
1472         if (len - (destp - (guchar *)dest) < MAX_LINELEN + 2) {         \
1473                 *destp = '\0';                                          \
1474                 return;                                                 \
1475         }                                                               \
1476                                                                         \
1477         if ((cond) && *srcp) {                                          \
1478                 if (destp > (guchar *)dest && left < MAX_LINELEN - 1) { \
1479                         if (isspace(*(destp - 1)))                      \
1480                                 destp--;                                \
1481                         else if (is_plain_text && isspace(*srcp))       \
1482                                 srcp++;                                 \
1483                         if (*srcp) {                                    \
1484                                 *destp++ = '\n';                        \
1485                                 *destp++ = ' ';                         \
1486                                 left = MAX_LINELEN - 1;                 \
1487                         }                                               \
1488                 }                                                       \
1489         }                                                               \
1490 }
1491
1492 void conv_encode_header_full(gchar *dest, gint len, const gchar *src,
1493                         gint header_len, gboolean addr_field,
1494                         const gchar *out_encoding_)
1495 {
1496         const gchar *cur_encoding;
1497         const gchar *out_encoding;
1498         gint mimestr_len;
1499         gchar *mimesep_enc;
1500         gint left;
1501         const guchar *srcp = src;
1502         guchar *destp = dest;
1503         gboolean use_base64;
1504
1505         g_return_if_fail(g_utf8_validate(src, -1, NULL) == TRUE);
1506
1507         if (MB_CUR_MAX > 1) {
1508                 use_base64 = TRUE;
1509                 mimesep_enc = "?B?";
1510         } else {
1511                 use_base64 = FALSE;
1512                 mimesep_enc = "?Q?";
1513         }
1514
1515         cur_encoding = CS_INTERNAL;
1516
1517         if (out_encoding_)
1518                 out_encoding = out_encoding_;
1519         else
1520                 out_encoding = conv_get_outgoing_charset_str();
1521
1522         if (!strcmp(out_encoding, CS_US_ASCII))
1523                 out_encoding = CS_ISO_8859_1;
1524
1525         mimestr_len = strlen(MIMESEP_BEGIN) + strlen(out_encoding) +
1526                 strlen(mimesep_enc) + strlen(MIMESEP_END);
1527
1528         left = MAX_LINELEN - header_len;
1529
1530         while (*srcp) {
1531                 LBREAK_IF_REQUIRED(left <= 0, TRUE);
1532
1533                 while (isspace(*srcp)) {
1534                         *destp++ = *srcp++;
1535                         left--;
1536                         LBREAK_IF_REQUIRED(left <= 0, TRUE);
1537                 }
1538
1539                 /* output as it is if the next word is ASCII string */
1540                 if (!is_next_nonascii(srcp)) {
1541                         gint word_len;
1542
1543                         word_len = get_next_word_len(srcp);
1544                         LBREAK_IF_REQUIRED(left < word_len, TRUE);
1545                         while (word_len > 0) {
1546                                 LBREAK_IF_REQUIRED(left + (MAX_HARD_LINELEN - MAX_LINELEN) <= 0, TRUE)
1547                                 *destp++ = *srcp++;
1548                                 left--;
1549                                 word_len--;
1550                         }
1551
1552                         continue;
1553                 }
1554
1555                 /* don't include parentheses in encoded strings */
1556                 if (addr_field && (*srcp == '(' || *srcp == ')')) {
1557                         LBREAK_IF_REQUIRED(left < 2, FALSE);
1558                         *destp++ = *srcp++;
1559                         left--;
1560                 }
1561
1562                 while (1) {
1563                         gint mb_len = 0;
1564                         gint cur_len = 0;
1565                         gchar *part_str;
1566                         gchar *out_str;
1567                         gchar *enc_str;
1568                         const guchar *p = srcp;
1569                         gint out_str_len;
1570                         gint out_enc_str_len;
1571                         gint mime_block_len;
1572                         gboolean cont = FALSE;
1573
1574                         while (*p != '\0') {
1575                                 if (isspace(*p) && !is_next_nonascii(p + 1))
1576                                         break;
1577                                 /* don't include parentheses in encoded
1578                                    strings */
1579                                 if (addr_field && (*p == '(' || *p == ')'))
1580                                         break;
1581
1582                                 mb_len = g_utf8_skip[*p];
1583
1584                                 Xstrndup_a(part_str, srcp, cur_len + mb_len, );
1585                                 out_str = conv_codeset_strdup
1586                                         (part_str, cur_encoding, out_encoding);
1587                                 if (!out_str) {
1588                                         g_warning("conv_encode_header(): code conversion failed\n");
1589                                         conv_unreadable_8bit(part_str);
1590                                         out_str = g_strdup(part_str);
1591                                 }
1592                                 out_str_len = strlen(out_str);
1593
1594                                 if (use_base64)
1595                                         out_enc_str_len = B64LEN(out_str_len);
1596                                 else
1597                                         out_enc_str_len =
1598                                                 qp_get_q_encoding_len(out_str);
1599
1600                                 g_free(out_str);
1601
1602                                 if (mimestr_len + out_enc_str_len <= left) {
1603                                         cur_len += mb_len;
1604                                         p += mb_len;
1605                                 } else if (cur_len == 0) {
1606                                         LBREAK_IF_REQUIRED(1, FALSE);
1607                                         continue;
1608                                 } else {
1609                                         cont = TRUE;
1610                                         break;
1611                                 }
1612                         }
1613
1614                         if (cur_len > 0) {
1615                                 Xstrndup_a(part_str, srcp, cur_len, );
1616                                 out_str = conv_codeset_strdup
1617                                         (part_str, cur_encoding, out_encoding);
1618                                 if (!out_str) {
1619                                         g_warning("conv_encode_header(): code conversion failed\n");
1620                                         conv_unreadable_8bit(part_str);
1621                                         out_str = g_strdup(part_str);
1622                                 }
1623                                 out_str_len = strlen(out_str);
1624
1625                                 if (use_base64)
1626                                         out_enc_str_len = B64LEN(out_str_len);
1627                                 else
1628                                         out_enc_str_len =
1629                                                 qp_get_q_encoding_len(out_str);
1630
1631                                 Xalloca(enc_str, out_enc_str_len + 1, );
1632                                 if (use_base64)
1633                                         base64_encode(enc_str, out_str, out_str_len);
1634                                 else
1635                                         qp_q_encode(enc_str, out_str);
1636
1637                                 g_free(out_str);
1638
1639                                 /* output MIME-encoded string block */
1640                                 mime_block_len = mimestr_len + strlen(enc_str);
1641                                 g_snprintf(destp, mime_block_len + 1,
1642                                            MIMESEP_BEGIN "%s%s%s" MIMESEP_END,
1643                                            out_encoding, mimesep_enc, enc_str);
1644                                 destp += mime_block_len;
1645                                 srcp += cur_len;
1646
1647                                 left -= mime_block_len;
1648                         }
1649
1650                         LBREAK_IF_REQUIRED(cont, FALSE);
1651
1652                         if (cur_len == 0)
1653                                 break;
1654                 }
1655         }
1656
1657         *destp = '\0';
1658 }
1659
1660 void conv_encode_header(gchar *dest, gint len, const gchar *src,
1661                         gint header_len, gboolean addr_field)
1662 {
1663         conv_encode_header_full(dest,len,src,header_len,addr_field,NULL);
1664 }
1665
1666 #undef LBREAK_IF_REQUIRED
1667 gchar *conv_filename_from_utf8(const gchar *utf8_file)
1668 {
1669         gchar *fs_file;
1670         GError *error = NULL;
1671
1672         fs_file = g_filename_from_utf8(utf8_file, -1, NULL, NULL, &error);
1673         if (error) {
1674                 g_warning("failed to convert encoding of file name: %s\n",
1675                           error->message);
1676                 g_error_free(error);
1677         }
1678         if (!fs_file)
1679                 fs_file = g_strdup(utf8_file);
1680
1681         return fs_file;
1682 }
1683
1684 gchar *conv_filename_to_utf8(const gchar *fs_file)
1685 {
1686         gchar *utf8_file = NULL;
1687         GError *error = NULL;
1688
1689         utf8_file = g_filename_to_utf8(fs_file, -1, NULL, NULL, &error);
1690         if (error) {
1691                 g_warning("failed to convert encoding of file name: %s\n",
1692                           error->message);
1693                 g_error_free(error);
1694         }
1695
1696         if (!utf8_file || !g_utf8_validate(utf8_file, -1, NULL)) {
1697                 g_free(utf8_file);
1698                 utf8_file = g_strdup(fs_file);
1699                 conv_unreadable_8bit(utf8_file);
1700         }
1701
1702         return utf8_file;
1703 }