src/codeconv.c

   1 /*
   2  * Sylpheed -- a GTK+ based, lightweight, and fast e-mail client
   3  * Copyright (C) 1999-2007 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 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., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, 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 "codeconv.h"
  38 #include "unmime.h"
  39 #include "base64.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 CharSet conv_get_locale_charset                  (void);
 115 static CharSet conv_get_outgoing_charset                (void);
 116 static CharSet conv_guess_ja_encoding(const gchar *str);
 117 static gboolean conv_is_ja_locale                       (void);
 118
 119 static void conv_jistoeuc(gchar *outbuf, gint outlen, const gchar *inbuf);
 120 static void conv_euctojis(gchar *outbuf, gint outlen, const gchar *inbuf);
 121 static void conv_sjistoeuc(gchar *outbuf, gint outlen, const gchar *inbuf);
 122
 123 static void conv_jistoutf8(gchar *outbuf, gint outlen, const gchar *inbuf);
 124 static void conv_sjistoutf8(gchar *outbuf, gint outlen, const gchar *inbuf);
 125 static void conv_euctoutf8(gchar *outbuf, gint outlen, const gchar *inbuf);
 126 static void conv_anytoutf8(gchar *outbuf, gint outlen, const gchar *inbuf);
 127
 128 static void conv_utf8toeuc(gchar *outbuf, gint outlen, const gchar *inbuf);
 129 static void conv_utf8tojis(gchar *outbuf, gint outlen, const gchar *inbuf);
 130
 131 static void conv_unreadable_8bit(gchar *str);
 132
 133 static void conv_jistodisp(gchar *outbuf, gint outlen, const gchar *inbuf);
 134 static void conv_sjistodisp(gchar *outbuf, gint outlen, const gchar *inbuf);
 135 static void conv_euctodisp(gchar *outbuf, gint outlen, const gchar *inbuf);
 136
 137 static void conv_anytodisp(gchar *outbuf, gint outlen, const gchar *inbuf);
 138 static void conv_ustodisp(gchar *outbuf, gint outlen, const gchar *inbuf);
 139 static void conv_noconv(gchar *outbuf, gint outlen, const gchar *inbuf);
 140
 141 static gboolean strict_mode = FALSE;
 142
 143 void codeconv_set_strict(gboolean mode)
 144 {
 145         strict_mode = mode;
 146 }
 147
 148 static void conv_jistoeuc(gchar *outbuf, gint outlen, const gchar *inbuf)
 149 {
 150         const guchar *in = inbuf;
 151         guchar *out = outbuf;
 152         JISState state = JIS_ASCII;
 153
 154         while (*in != '\0') {
 155                 if (*in == ESC) {
 156                         in++;
 157                         if (*in == '$') {
 158                                 if (*(in + 1) == '@' || *(in + 1) == 'B') {
 159                                         state = JIS_KANJI;
 160                                         in += 2;
 161                                 } else if (*(in + 1) == '(' &&
 162                                            *(in + 2) == 'D') {
 163                                         state = JIS_AUXKANJI;
 164                                         in += 3;
 165                                 } else {
 166                                         /* unknown escape sequence */
 167                                         state = JIS_ASCII;
 168                                 }
 169                         } else if (*in == '(') {
 170                                 if (*(in + 1) == 'B' || *(in + 1) == 'J') {
 171                                         state = JIS_ASCII;
 172                                         in += 2;
 173                                 } else if (*(in + 1) == 'I') {
 174                                         state = JIS_HWKANA;
 175                                         in += 2;
 176                                 } else {
 177                                         /* unknown escape sequence */
 178                                         state = JIS_ASCII;
 179                                 }
 180                         } else {
 181                                 /* unknown escape sequence */
 182                                 state = JIS_ASCII;
 183                         }
 184                 } else if (*in == 0x0e) {
 185                         state = JIS_HWKANA;
 186                         in++;
 187                 } else if (*in == 0x0f) {
 188                         state = JIS_ASCII;
 189                         in++;
 190                 } else {
 191                         switch (state) {
 192                         case JIS_ASCII:
 193                                 *out++ = *in++;
 194                                 break;
 195                         case JIS_KANJI:
 196                                 *out++ = *in++ | 0x80;
 197                                 if (*in == '\0') break;
 198                                 *out++ = *in++ | 0x80;
 199                                 break;
 200                         case JIS_HWKANA:
 201                                 *out++ = 0x8e;
 202                                 *out++ = *in++ | 0x80;
 203                                 break;
 204                         case JIS_AUXKANJI:
 205                                 *out++ = 0x8f;
 206                                 *out++ = *in++ | 0x80;
 207                                 if (*in == '\0') break;
 208                                 *out++ = *in++ | 0x80;
 209                                 break;
 210                         }
 211                 }
 212         }
 213
 214         *out = '\0';
 215 }
 216
 217 #define JIS_HWDAKUTEN           0x5e
 218 #define JIS_HWHANDAKUTEN        0x5f
 219
 220 static gint conv_jis_hantozen(guchar *outbuf, guchar jis_code, guchar sound_sym)
 221 {
 222         static guint16 h2z_tbl[] = {
 223                 /* 0x20 - 0x2f */
 224                 0x0000, 0x2123, 0x2156, 0x2157, 0x2122, 0x2126, 0x2572, 0x2521,
 225                 0x2523, 0x2525, 0x2527, 0x2529, 0x2563, 0x2565, 0x2567, 0x2543,
 226                 /* 0x30 - 0x3f */
 227                 0x213c, 0x2522, 0x2524, 0x2526, 0x2528, 0x252a, 0x252b, 0x252d,
 228                 0x252f, 0x2531, 0x2533, 0x2535, 0x2537, 0x2539, 0x253b, 0x253d,
 229                 /* 0x40 - 0x4f */
 230                 0x253f, 0x2541, 0x2544, 0x2546, 0x2548, 0x254a, 0x254b, 0x254c,
 231                 0x254d, 0x254e, 0x254f, 0x2552, 0x2555, 0x2558, 0x255b, 0x255e,
 232                 /* 0x50 - 0x5f */
 233                 0x255f, 0x2560, 0x2561, 0x2562, 0x2564, 0x2566, 0x2568, 0x2569,
 234                 0x256a, 0x256b, 0x256c, 0x256d, 0x256f, 0x2573, 0x212b, 0x212c
 235         };
 236
 237         static guint16 dakuten_tbl[] = {
 238                 /* 0x30 - 0x3f */
 239                 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x252c, 0x252e,
 240                 0x2530, 0x2532, 0x2534, 0x2536, 0x2538, 0x253a, 0x253c, 0x253e,
 241                 /* 0x40 - 0x4f */
 242                 0x2540, 0x2542, 0x2545, 0x2547, 0x2549, 0x0000, 0x0000, 0x0000,
 243                 0x0000, 0x0000, 0x2550, 0x2553, 0x2556, 0x2559, 0x255c, 0x0000
 244         };
 245
 246         static guint16 handakuten_tbl[] = {
 247                 /* 0x4a - 0x4e */
 248                 0x2551, 0x2554, 0x2557, 0x255a, 0x255d
 249         };
 250
 251         guint16 out_code;
 252
 253         jis_code &= 0x7f;
 254         sound_sym &= 0x7f;
 255
 256         if (jis_code < 0x21 || jis_code > 0x5f)
 257                 return 0;
 258
 259         if (sound_sym == JIS_HWDAKUTEN &&
 260             jis_code >= 0x36 && jis_code <= 0x4e) {
 261                 out_code = dakuten_tbl[jis_code - 0x30];
 262                 if (out_code != 0) {
 263                         *outbuf = out_code >> 8;
 264                         *(outbuf + 1) = out_code & 0xff;
 265                         return 2;
 266                 }
 267         }
 268
 269         if (sound_sym == JIS_HWHANDAKUTEN &&
 270             jis_code >= 0x4a && jis_code <= 0x4e) {
 271                 out_code = handakuten_tbl[jis_code - 0x4a];
 272                 *outbuf = out_code >> 8;
 273                 *(outbuf + 1) = out_code & 0xff;
 274                 return 2;
 275         }
 276
 277         out_code = h2z_tbl[jis_code - 0x20];
 278         *outbuf = out_code >> 8;
 279         *(outbuf + 1) = out_code & 0xff;
 280         return 1;
 281 }
 282
 283 static void conv_euctojis(gchar *outbuf, gint outlen, const gchar *inbuf)
 284 {
 285         const guchar *in = inbuf;
 286         guchar *out = outbuf;
 287         JISState state = JIS_ASCII;
 288
 289         while (*in != '\0') {
 290                 if (IS_ASCII(*in)) {
 291                         K_OUT();
 292                         *out++ = *in++;
 293                 } else if (iseuckanji(*in)) {
 294                         if (iseuckanji(*(in + 1))) {
 295                                 K_IN();
 296                                 *out++ = *in++ & 0x7f;
 297                                 *out++ = *in++ & 0x7f;
 298                         } else {
 299                                 K_OUT();
 300                                 *out++ = SUBST_CHAR;
 301                                 in++;
 302                                 if (*in != '\0' && !IS_ASCII(*in)) {
 303                                         *out++ = SUBST_CHAR;
 304                                         in++;
 305                                 }
 306                         }
 307                 } else if (iseuchwkana1(*in)) {
 308                         if (iseuchwkana2(*(in + 1))) {
 309                                 if (prefs_common.allow_jisx0201_kana) {
 310                                         HW_IN();
 311                                         in++;
 312                                         *out++ = *in++ & 0x7f;
 313                                 } else {
 314                                         guchar jis_ch[2];
 315                                         gint len;
 316
 317                                         if (iseuchwkana1(*(in + 2)) &&
 318                                             iseuchwkana2(*(in + 3)))
 319                                                 len = conv_jis_hantozen
 320                                                         (jis_ch,
 321                                                          *(in + 1), *(in + 3));
 322                                         else
 323                                                 len = conv_jis_hantozen
 324                                                         (jis_ch,
 325                                                          *(in + 1), '\0');
 326                                         if (len == 0)
 327                                                 in += 2;
 328                                         else {
 329                                                 K_IN();
 330                                                 in += len * 2;
 331                                                 *out++ = jis_ch[0];
 332                                                 *out++ = jis_ch[1];
 333                                         }
 334                                 }
 335                         } else {
 336                                 K_OUT();
 337                                 in++;
 338                                 if (*in != '\0' && !IS_ASCII(*in)) {
 339                                         *out++ = SUBST_CHAR;
 340                                         in++;
 341                                 }
 342                         }
 343                 } else if (iseucaux(*in)) {
 344                         in++;
 345                         if (iseuckanji(*in) && iseuckanji(*(in + 1))) {
 346                                 AUX_IN();
 347                                 *out++ = *in++ & 0x7f;
 348                                 *out++ = *in++ & 0x7f;
 349                         } else {
 350                                 K_OUT();
 351                                 if (*in != '\0' && !IS_ASCII(*in)) {
 352                                         *out++ = SUBST_CHAR;
 353                                         in++;
 354                                         if (*in != '\0' && !IS_ASCII(*in)) {
 355                                                 *out++ = SUBST_CHAR;
 356                                                 in++;
 357                                         }
 358                                 }
 359                         }
 360                 } else {
 361                         K_OUT();
 362                         *out++ = SUBST_CHAR;
 363                         in++;
 364                 }
 365         }
 366
 367         K_OUT();
 368         *out = '\0';
 369 }
 370
 371 static void conv_sjistoeuc(gchar *outbuf, gint outlen, const gchar *inbuf)
 372 {
 373         const guchar *in = inbuf;
 374         guchar *out = outbuf;
 375
 376         while (*in != '\0') {
 377                 if (IS_ASCII(*in)) {
 378                         *out++ = *in++;
 379                 } else if (issjiskanji1(*in)) {
 380                         if (issjiskanji2(*(in + 1))) {
 381                                 guchar out1 = *in;
 382                                 guchar out2 = *(in + 1);
 383                                 guchar row;
 384
 385                                 row = out1 < 0xa0 ? 0x70 : 0xb0;
 386                                 if (out2 < 0x9f) {
 387                                         out1 = (out1 - row) * 2 - 1;
 388                                         out2 -= out2 > 0x7f ? 0x20 : 0x1f;
 389                                 } else {
 390                                         out1 = (out1 - row) * 2;
 391                                         out2 -= 0x7e;
 392                                 }
 393
 394                                 *out++ = out1 | 0x80;
 395                                 *out++ = out2 | 0x80;
 396                                 in += 2;
 397                         } else {
 398                                 *out++ = SUBST_CHAR;
 399                                 in++;
 400                                 if (*in != '\0' && !IS_ASCII(*in)) {
 401                                         *out++ = SUBST_CHAR;
 402                                         in++;
 403                                 }
 404                         }
 405                 } else if (issjishwkana(*in)) {
 406                         *out++ = 0x8e;
 407                         *out++ = *in++;
 408                 } else {
 409                         *out++ = SUBST_CHAR;
 410                         in++;
 411                 }
 412         }
 413
 414         *out = '\0';
 415 }
 416
 417 static void conv_jistoutf8(gchar *outbuf, gint outlen, const gchar *inbuf)
 418 {
 419         gchar *eucstr;
 420
 421         Xalloca(eucstr, outlen, return);
 422
 423         conv_jistoeuc(eucstr, outlen, inbuf);
 424         conv_euctoutf8(outbuf, outlen, eucstr);
 425 }
 426
 427 static void conv_sjistoutf8(gchar *outbuf, gint outlen, const gchar *inbuf)
 428 {
 429         gchar *tmpstr;
 430
 431         tmpstr = conv_iconv_strdup(inbuf, CS_SHIFT_JIS, CS_UTF_8);
 432         if (tmpstr) {
 433                 strncpy2(outbuf, tmpstr, outlen);
 434                 g_free(tmpstr);
 435         } else
 436                 strncpy2(outbuf, inbuf, outlen);
 437 }
 438
 439 static void conv_euctoutf8(gchar *outbuf, gint outlen, const gchar *inbuf)
 440 {
 441         static iconv_t cd = (iconv_t)-1;
 442         static gboolean iconv_ok = TRUE;
 443         gchar *tmpstr;
 444
 445         if (cd == (iconv_t)-1) {
 446                 if (!iconv_ok) {
 447                         strncpy2(outbuf, inbuf, outlen);
 448                         return;
 449                 }
 450                 cd = iconv_open(CS_UTF_8, CS_EUC_JP_MS);
 451                 if (cd == (iconv_t)-1) {
 452                         cd = iconv_open(CS_UTF_8, CS_EUC_JP);
 453                         if (cd == (iconv_t)-1) {
 454                                 g_warning("conv_euctoutf8(): %s\n",
 455                                           g_strerror(errno));
 456                                 iconv_ok = FALSE;
 457                                 strncpy2(outbuf, inbuf, outlen);
 458                                 return;
 459                         }
 460                 }
 461         }
 462
 463         tmpstr = conv_iconv_strdup_with_cd(inbuf, cd);
 464         if (tmpstr) {
 465                 strncpy2(outbuf, tmpstr, outlen);
 466                 g_free(tmpstr);
 467         } else
 468                 strncpy2(outbuf, inbuf, outlen);
 469 }
 470
 471 static void conv_anytoutf8(gchar *outbuf, gint outlen, const gchar *inbuf)
 472 {
 473         switch (conv_guess_ja_encoding(inbuf)) {
 474         case C_ISO_2022_JP:
 475                 conv_jistoutf8(outbuf, outlen, inbuf);
 476                 break;
 477         case C_SHIFT_JIS:
 478                 conv_sjistoutf8(outbuf, outlen, inbuf);
 479                 break;
 480         case C_EUC_JP:
 481                 conv_euctoutf8(outbuf, outlen, inbuf);
 482                 break;
 483         default:
 484                 strncpy2(outbuf, inbuf, outlen);
 485                 break;
 486         }
 487 }
 488
 489 static void conv_utf8toeuc(gchar *outbuf, gint outlen, const gchar *inbuf)
 490 {
 491         static iconv_t cd = (iconv_t)-1;
 492         static gboolean iconv_ok = TRUE;
 493         gchar *tmpstr;
 494
 495         if (cd == (iconv_t)-1) {
 496                 if (!iconv_ok) {
 497                         strncpy2(outbuf, inbuf, outlen);
 498                         return;
 499                 }
 500                 cd = iconv_open(CS_EUC_JP_MS, CS_UTF_8);
 501                 if (cd == (iconv_t)-1) {
 502                         cd = iconv_open(CS_EUC_JP, CS_UTF_8);
 503                         if (cd == (iconv_t)-1) {
 504                                 g_warning("conv_utf8toeuc(): %s\n",
 505                                           g_strerror(errno));
 506                                 iconv_ok = FALSE;
 507                                 strncpy2(outbuf, inbuf, outlen);
 508                                 return;
 509                         }
 510                 }
 511         }
 512
 513         tmpstr = conv_iconv_strdup_with_cd(inbuf, cd);
 514         if (tmpstr) {
 515                 strncpy2(outbuf, tmpstr, outlen);
 516                 g_free(tmpstr);
 517         } else
 518                 strncpy2(outbuf, inbuf, outlen);
 519 }
 520
 521 static void conv_utf8tojis(gchar *outbuf, gint outlen, const gchar *inbuf)
 522 {
 523         gchar *eucstr;
 524
 525         Xalloca(eucstr, outlen, return);
 526
 527         conv_utf8toeuc(eucstr, outlen, inbuf);
 528         conv_euctojis(outbuf, outlen, eucstr);
 529 }
 530
 531 static void conv_unreadable_8bit(gchar *str)
 532 {
 533         register guchar *p = str;
 534
 535         while (*p != '\0') {
 536                 /* convert CR+LF -> LF */
 537                 if (*p == '\r' && *(p + 1) == '\n')
 538                         memmove(p, p + 1, strlen(p));
 539                 else if (!IS_ASCII(*p)) *p = SUBST_CHAR;
 540                 p++;
 541         }
 542 }
 543
 544 static CharSet conv_guess_ja_encoding(const gchar *str)
 545 {
 546         const guchar *p = str;
 547         CharSet guessed = C_US_ASCII;
 548
 549         while (*p != '\0') {
 550                 if (*p == ESC && (*(p + 1) == '$' || *(p + 1) == '(')) {
 551                         if (guessed == C_US_ASCII)
 552                                 return C_ISO_2022_JP;
 553                         p += 2;
 554                 } else if (IS_ASCII(*p)) {
 555                         p++;
 556                 } else if (iseuckanji(*p) && iseuckanji(*(p + 1))) {
 557                         if (*p >= 0xfd && *p <= 0xfe)
 558                                 return C_EUC_JP;
 559                         else if (guessed == C_SHIFT_JIS) {
 560                                 if ((issjiskanji1(*p) &&
 561                                      issjiskanji2(*(p + 1))) ||
 562                                     issjishwkana(*p))
 563                                         guessed = C_SHIFT_JIS;
 564                                 else
 565                                         guessed = C_EUC_JP;
 566                         } else
 567                                 guessed = C_EUC_JP;
 568                         p += 2;
 569                 } else if (issjiskanji1(*p) && issjiskanji2(*(p + 1))) {
 570                         if (iseuchwkana1(*p) && iseuchwkana2(*(p + 1)))
 571                                 guessed = C_SHIFT_JIS;
 572                         else
 573                                 return C_SHIFT_JIS;
 574                         p += 2;
 575                 } else if (issjishwkana(*p)) {
 576                         guessed = C_SHIFT_JIS;
 577                         p++;
 578                 } else {
 579                         p++;
 580                 }
 581         }
 582
 583         return guessed;
 584 }
 585
 586 static void conv_jistodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
 587 {
 588         conv_jistoutf8(outbuf, outlen, inbuf);
 589 }
 590
 591 static void conv_sjistodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
 592 {
 593         conv_sjistoutf8(outbuf, outlen, inbuf);
 594 }
 595
 596 static void conv_euctodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
 597 {
 598         conv_euctoutf8(outbuf, outlen, inbuf);
 599 }
 600
 601 void conv_utf8todisp(gchar *outbuf, gint outlen, const gchar *inbuf)
 602 {
 603         if (g_utf8_validate(inbuf, -1, NULL) == TRUE)
 604                 strncpy2(outbuf, inbuf, outlen);
 605         else
 606                 conv_ustodisp(outbuf, outlen, inbuf);
 607 }
 608
 609 static void conv_anytodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
 610 {
 611         conv_anytoutf8(outbuf, outlen, inbuf);
 612         if (g_utf8_validate(outbuf, -1, NULL) != TRUE)
 613                 conv_unreadable_8bit(outbuf);
 614 }
 615
 616 static void conv_ustodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
 617 {
 618         strncpy2(outbuf, inbuf, outlen);
 619         conv_unreadable_8bit(outbuf);
 620 }
 621
 622 void conv_localetodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
 623 {
 624         gchar *tmpstr;
 625
 626         codeconv_set_strict(TRUE);
 627         tmpstr = conv_iconv_strdup(inbuf, conv_get_locale_charset_str(),
 628                                    CS_INTERNAL);
 629         codeconv_set_strict(FALSE);
 630         if (tmpstr && g_utf8_validate(tmpstr, -1, NULL)) {
 631                 strncpy2(outbuf, tmpstr, outlen);
 632                 g_free(tmpstr);
 633                 return;
 634         } else if (tmpstr && !g_utf8_validate(tmpstr, -1, NULL)) {
 635                 g_free(tmpstr);
 636                 codeconv_set_strict(TRUE);
 637                 tmpstr = conv_iconv_strdup(inbuf,
 638                                 conv_get_locale_charset_str_no_utf8(),
 639                                 CS_INTERNAL);
 640                 codeconv_set_strict(FALSE);
 641         }
 642         if (tmpstr && g_utf8_validate(tmpstr, -1, NULL)) {
 643                 strncpy2(outbuf, tmpstr, outlen);
 644                 g_free(tmpstr);
 645                 return;
 646         } else {
 647                 g_free(tmpstr);
 648                 conv_utf8todisp(outbuf, outlen, inbuf);
 649         }
 650 }
 651
 652 static void conv_noconv(gchar *outbuf, gint outlen, const gchar *inbuf)
 653 {
 654         strncpy2(outbuf, inbuf, outlen);
 655 }
 656
 657 static const gchar *
 658 conv_get_fallback_for_private_encoding(const gchar *encoding)
 659 {
 660         if (encoding && (encoding[0] == 'X' || encoding[0] == 'x') &&
 661             encoding[1] == '-') {
 662                 if (!g_ascii_strcasecmp(encoding, CS_X_GBK))
 663                         return CS_GBK;
 664         }
 665
 666         return encoding;
 667 }
 668
 669 CodeConverter *conv_code_converter_new(const gchar *src_charset)
 670 {
 671         CodeConverter *conv;
 672
 673         src_charset = conv_get_fallback_for_private_encoding(src_charset);
 674
 675         conv = g_new0(CodeConverter, 1);
 676         conv->code_conv_func = conv_get_code_conv_func(src_charset, NULL);
 677         conv->charset_str = g_strdup(src_charset);
 678         conv->charset = conv_get_charset_from_str(src_charset);
 679
 680         return conv;
 681 }
 682
 683 void conv_code_converter_destroy(CodeConverter *conv)
 684 {
 685         g_free(conv->charset_str);
 686         g_free(conv);
 687 }
 688
 689 gint conv_convert(CodeConverter *conv, gchar *outbuf, gint outlen,
 690                   const gchar *inbuf)
 691 {
 692         if (conv->code_conv_func != conv_noconv)
 693                 conv->code_conv_func(outbuf, outlen, inbuf);
 694         else {
 695                 gchar *str;
 696
 697                 str = conv_iconv_strdup(inbuf, conv->charset_str, NULL);
 698                 if (!str)
 699                         return -1;
 700                 else {
 701                         strncpy2(outbuf, str, outlen);
 702                         g_free(str);
 703                 }
 704         }
 705
 706         return 0;
 707 }
 708
 709 gchar *conv_codeset_strdup(const gchar *inbuf,
 710                            const gchar *src_code, const gchar *dest_code)
 711 {
 712         gchar *buf;
 713         size_t len;
 714         CodeConvFunc conv_func;
 715
 716         src_code = conv_get_fallback_for_private_encoding(src_code);
 717         conv_func = conv_get_code_conv_func(src_code, dest_code);
 718         if (conv_func != conv_noconv) {
 719                 len = (strlen(inbuf) + 1) * 3;
 720                 buf = g_malloc(len);
 721                 if (!buf) return NULL;
 722
 723                 conv_func(buf, len, inbuf);
 724                 return g_realloc(buf, strlen(buf) + 1);
 725         }
 726
 727         return conv_iconv_strdup(inbuf, src_code, dest_code);
 728 }
 729
 730 static CodeConvFunc conv_get_code_conv_func(const gchar *src_charset_str,
 731                                      const gchar *dest_charset_str)
 732 {
 733         CodeConvFunc code_conv = conv_noconv;
 734         CharSet src_charset;
 735         CharSet dest_charset;
 736
 737         if (!src_charset_str)
 738                 src_charset = conv_get_locale_charset();
 739         else
 740                 src_charset = conv_get_charset_from_str(src_charset_str);
 741
 742         /* auto detection mode */
 743         if (!src_charset_str && !dest_charset_str) {
 744                 if (conv_is_ja_locale())
 745                         return conv_anytodisp;
 746                 else
 747                         return conv_noconv;
 748         }
 749
 750         dest_charset = conv_get_charset_from_str(dest_charset_str);
 751
 752         if (dest_charset == C_US_ASCII)
 753                 return conv_ustodisp;
 754
 755         switch (src_charset) {
 756         case C_US_ASCII:
 757         case C_ISO_8859_1:
 758         case C_ISO_8859_2:
 759         case C_ISO_8859_3:
 760         case C_ISO_8859_4:
 761         case C_ISO_8859_5:
 762         case C_ISO_8859_6:
 763         case C_ISO_8859_7:
 764         case C_ISO_8859_8:
 765         case C_ISO_8859_9:
 766         case C_ISO_8859_10:
 767         case C_ISO_8859_11:
 768         case C_ISO_8859_13:
 769         case C_ISO_8859_14:
 770         case C_ISO_8859_15:
 771                 break;
 772         case C_ISO_2022_JP:
 773         case C_ISO_2022_JP_2:
 774         case C_ISO_2022_JP_3:
 775                 if (dest_charset == C_AUTO)
 776                         code_conv = conv_jistodisp;
 777                 else if (dest_charset == C_EUC_JP)
 778                         code_conv = conv_jistoeuc;
 779                 else if (dest_charset == C_UTF_8)
 780                         code_conv = conv_jistoutf8;
 781                 break;
 782         case C_SHIFT_JIS:
 783                 if (dest_charset == C_AUTO)
 784                         code_conv = conv_sjistodisp;
 785                 else if (dest_charset == C_EUC_JP)
 786                         code_conv = conv_sjistoeuc;
 787                 else if (dest_charset == C_UTF_8)
 788                         code_conv = conv_sjistoutf8;
 789                 break;
 790         case C_EUC_JP:
 791                 if (dest_charset == C_AUTO)
 792                         code_conv = conv_euctodisp;
 793                 else if (dest_charset == C_ISO_2022_JP   ||
 794                          dest_charset == C_ISO_2022_JP_2 ||
 795                          dest_charset == C_ISO_2022_JP_3)
 796                         code_conv = conv_euctojis;
 797                 else if (dest_charset == C_UTF_8)
 798                         code_conv = conv_euctoutf8;
 799                 break;
 800         case C_UTF_8:
 801                 if (dest_charset == C_EUC_JP)
 802                         code_conv = conv_utf8toeuc;
 803                 else if (dest_charset == C_ISO_2022_JP   ||
 804                          dest_charset == C_ISO_2022_JP_2 ||
 805                          dest_charset == C_ISO_2022_JP_3)
 806                         code_conv = conv_utf8tojis;
 807                 break;
 808         default:
 809                 break;
 810         }
 811
 812         return code_conv;
 813 }
 814
 815 gchar *conv_iconv_strdup(const gchar *inbuf,
 816                          const gchar *src_code, const gchar *dest_code)
 817 {
 818         iconv_t cd;
 819         gchar *outbuf;
 820
 821         if (!src_code && !dest_code &&
 822             g_utf8_validate(inbuf, -1, NULL))
 823                 return g_strdup(inbuf);
 824
 825         if (!src_code)
 826                 src_code = conv_get_outgoing_charset_str();
 827         if (!dest_code)
 828                 dest_code = CS_INTERNAL;
 829
 830         /* don't convert if src and dest codeset are identical */
 831         if (!strcasecmp(src_code, dest_code))
 832                 return g_strdup(inbuf);
 833
 834         /* don't convert if dest codeset is US-ASCII */
 835         if (!strcasecmp(src_code, CS_US_ASCII))
 836                 return g_strdup(inbuf);
 837
 838         /* don't convert if dest codeset is US-ASCII */
 839         if (!strcasecmp(dest_code, CS_US_ASCII))
 840                 return g_strdup(inbuf);
 841
 842         cd = iconv_open(dest_code, src_code);
 843         if (cd == (iconv_t)-1)
 844                 return NULL;
 845
 846         outbuf = conv_iconv_strdup_with_cd(inbuf, cd);
 847
 848         iconv_close(cd);
 849
 850         return outbuf;
 851 }
 852
 853 gchar *conv_iconv_strdup_with_cd(const gchar *inbuf, iconv_t cd)
 854 {
 855         const gchar *inbuf_p;
 856         gchar *outbuf;
 857         gchar *outbuf_p;
 858         size_t in_size;
 859         size_t in_left;
 860         size_t out_size;
 861         size_t out_left;
 862         size_t n_conv;
 863         size_t len;
 864
 865         inbuf_p = inbuf;
 866         in_size = strlen(inbuf);
 867         in_left = in_size;
 868         out_size = (in_size + 1) * 2;
 869         outbuf = g_malloc(out_size);
 870         outbuf_p = outbuf;
 871         out_left = out_size;
 872
 873 #define EXPAND_BUF()                            \
 874 {                                               \
 875         len = outbuf_p - outbuf;                \
 876         out_size *= 2;                          \
 877         outbuf = g_realloc(outbuf, out_size);   \
 878         outbuf_p = outbuf + len;                \
 879         out_left = out_size - len;              \
 880 }
 881
 882         while ((n_conv = iconv(cd, (ICONV_CONST gchar **)&inbuf_p, &in_left,
 883                                &outbuf_p, &out_left)) == (size_t)-1) {
 884                 if (EILSEQ == errno) {
 885                         if (strict_mode) {
 886                                 g_free(outbuf);
 887                                 return NULL;
 888                         }
 889                         //g_print("iconv(): at %d: %s\n", in_size - in_left, g_strerror(errno));
 890                         inbuf_p++;
 891                         in_left--;
 892                         if (out_left == 0) {
 893                                 EXPAND_BUF();
 894                         }
 895                         *outbuf_p++ = SUBST_CHAR;
 896                         out_left--;
 897                 } else if (EINVAL == errno) {
 898                         break;
 899                 } else if (E2BIG == errno) {
 900                         EXPAND_BUF();
 901                 } else {
 902                         g_warning("conv_iconv_strdup(): %s\n",
 903                                   g_strerror(errno));
 904                         break;
 905                 }
 906         }
 907
 908         while ((n_conv = iconv(cd, NULL, NULL, &outbuf_p, &out_left)) ==
 909                (size_t)-1) {
 910                 if (E2BIG == errno) {
 911                         EXPAND_BUF();
 912                 } else {
 913                         g_warning("conv_iconv_strdup(): %s\n",
 914                                   g_strerror(errno));
 915                         break;
 916                 }
 917         }
 918
 919 #undef EXPAND_BUF
 920
 921         len = outbuf_p - outbuf;
 922         outbuf = g_realloc(outbuf, len + 1);
 923         outbuf[len] = '\0';
 924
 925         return outbuf;
 926 }
 927
 928 static const struct {
 929         CharSet charset;
 930         gchar *const name;
 931 } charsets[] = {
 932         {C_US_ASCII,            CS_US_ASCII},
 933         {C_US_ASCII,            CS_ANSI_X3_4_1968},
 934         {C_UTF_8,               CS_UTF_8},
 935         {C_UTF_7,               CS_UTF_7},
 936         {C_ISO_8859_1,          CS_ISO_8859_1},
 937         {C_ISO_8859_2,          CS_ISO_8859_2},
 938         {C_ISO_8859_3,          CS_ISO_8859_3},
 939         {C_ISO_8859_4,          CS_ISO_8859_4},
 940         {C_ISO_8859_5,          CS_ISO_8859_5},
 941         {C_ISO_8859_6,          CS_ISO_8859_6},
 942         {C_ISO_8859_7,          CS_ISO_8859_7},
 943         {C_ISO_8859_8,          CS_ISO_8859_8},
 944         {C_ISO_8859_9,          CS_ISO_8859_9},
 945         {C_ISO_8859_10,         CS_ISO_8859_10},
 946         {C_ISO_8859_11,         CS_ISO_8859_11},
 947         {C_ISO_8859_13,         CS_ISO_8859_13},
 948         {C_ISO_8859_14,         CS_ISO_8859_14},
 949         {C_ISO_8859_15,         CS_ISO_8859_15},
 950         {C_BALTIC,              CS_BALTIC},
 951         {C_CP1250,              CS_CP1250},
 952         {C_CP1251,              CS_CP1251},
 953         {C_CP1252,              CS_CP1252},
 954         {C_CP1253,              CS_CP1253},
 955         {C_CP1254,              CS_CP1254},
 956         {C_CP1255,              CS_CP1255},
 957         {C_CP1256,              CS_CP1256},
 958         {C_CP1257,              CS_CP1257},
 959         {C_CP1258,              CS_CP1258},
 960         {C_WINDOWS_1250,        CS_WINDOWS_1250},
 961         {C_WINDOWS_1251,        CS_WINDOWS_1251},
 962         {C_WINDOWS_1252,        CS_WINDOWS_1252},
 963         {C_WINDOWS_1253,        CS_WINDOWS_1253},
 964         {C_WINDOWS_1254,        CS_WINDOWS_1254},
 965         {C_WINDOWS_1255,        CS_WINDOWS_1255},
 966         {C_WINDOWS_1256,        CS_WINDOWS_1256},
 967         {C_WINDOWS_1257,        CS_WINDOWS_1257},
 968         {C_WINDOWS_1258,        CS_WINDOWS_1258},
 969         {C_KOI8_R,              CS_KOI8_R},
 970         {C_KOI8_T,              CS_KOI8_T},
 971         {C_KOI8_U,              CS_KOI8_U},
 972         {C_ISO_2022_JP,         CS_ISO_2022_JP},
 973         {C_ISO_2022_JP_2,       CS_ISO_2022_JP_2},
 974         {C_ISO_2022_JP_3,       CS_ISO_2022_JP_3},
 975         {C_EUC_JP,              CS_EUC_JP},
 976         {C_EUC_JP,              CS_EUCJP},
 977         {C_EUC_JP_MS,           CS_EUC_JP_MS},
 978         {C_SHIFT_JIS,           CS_SHIFT_JIS},
 979         {C_SHIFT_JIS,           CS_SHIFT__JIS},
 980         {C_SHIFT_JIS,           CS_SJIS},
 981         {C_ISO_2022_KR,         CS_ISO_2022_KR},
 982         {C_EUC_KR,              CS_EUC_KR},
 983         {C_ISO_2022_CN,         CS_ISO_2022_CN},
 984         {C_EUC_CN,              CS_EUC_CN},
 985         {C_GB2312,              CS_GB2312},
 986         {C_GBK,                 CS_GBK},
 987         {C_EUC_TW,              CS_EUC_TW},
 988         {C_BIG5,                CS_BIG5},
 989         {C_BIG5_HKSCS,          CS_BIG5_HKSCS},
 990         {C_TIS_620,             CS_TIS_620},
 991         {C_WINDOWS_874,         CS_WINDOWS_874},
 992         {C_GEORGIAN_PS,         CS_GEORGIAN_PS},
 993         {C_TCVN5712_1,          CS_TCVN5712_1},
 994 };
 995
 996 static const struct {
 997         gchar *const locale;
 998         CharSet charset;
 999         CharSet out_charset;
1000 } locale_table[] = {
1001         {"ja_JP.eucJP"  , C_EUC_JP      , C_ISO_2022_JP},
1002         {"ja_JP.EUC-JP" , C_EUC_JP      , C_ISO_2022_JP},
1003         {"ja_JP.EUC"    , C_EUC_JP      , C_ISO_2022_JP},
1004         {"ja_JP.ujis"   , C_EUC_JP      , C_ISO_2022_JP},
1005         {"ja_JP.SJIS"   , C_SHIFT_JIS   , C_ISO_2022_JP},
1006         {"ja_JP.JIS"    , C_ISO_2022_JP , C_ISO_2022_JP},
1007 #ifdef G_OS_WIN32
1008         {"ja_JP"        , C_SHIFT_JIS   , C_ISO_2022_JP},
1009 #else
1010         {"ja_JP"        , C_EUC_JP      , C_ISO_2022_JP},
1011 #endif
1012         {"ko_KR.EUC-KR" , C_EUC_KR      , C_EUC_KR},
1013         {"ko_KR"        , C_EUC_KR      , C_EUC_KR},
1014         {"zh_CN.GB2312" , C_GB2312      , C_GB2312},
1015         {"zh_CN.GBK"    , C_GBK         , C_GBK},
1016         {"zh_CN"        , C_GB2312      , C_GB2312},
1017         {"zh_HK"        , C_BIG5_HKSCS  , C_BIG5_HKSCS},
1018         {"zh_TW.eucTW"  , C_EUC_TW      , C_BIG5},
1019         {"zh_TW.EUC-TW" , C_EUC_TW      , C_BIG5},
1020         {"zh_TW.Big5"   , C_BIG5        , C_BIG5},
1021         {"zh_TW"        , C_BIG5        , C_BIG5},
1022
1023         {"ru_RU.KOI8-R" , C_KOI8_R      , C_KOI8_R},
1024         {"ru_RU.KOI8R"  , C_KOI8_R      , C_KOI8_R},
1025         {"ru_RU.CP1251" , C_WINDOWS_1251, C_KOI8_R},
1026         {"ru_RU"        , C_ISO_8859_5  , C_KOI8_R},
1027         {"tg_TJ"        , C_KOI8_T      , C_KOI8_T},
1028         {"ru_UA"        , C_KOI8_U      , C_KOI8_U},
1029         {"uk_UA.CP1251" , C_WINDOWS_1251, C_KOI8_U},
1030         {"uk_UA"        , C_KOI8_U      , C_KOI8_U},
1031
1032         {"be_BY"        , C_WINDOWS_1251, C_WINDOWS_1251},
1033         {"bg_BG"        , C_WINDOWS_1251, C_WINDOWS_1251},
1034
1035         {"yi_US"        , C_WINDOWS_1255, C_WINDOWS_1255},
1036
1037         {"af_ZA"        , C_ISO_8859_1  , C_ISO_8859_1},
1038         {"br_FR"        , C_ISO_8859_1  , C_ISO_8859_1},
1039         {"ca_ES"        , C_ISO_8859_1  , C_ISO_8859_1},
1040         {"da_DK"        , C_ISO_8859_1  , C_ISO_8859_1},
1041         {"de_AT"        , C_ISO_8859_1  , C_ISO_8859_1},
1042         {"de_BE"        , C_ISO_8859_1  , C_ISO_8859_1},
1043         {"de_CH"        , C_ISO_8859_1  , C_ISO_8859_1},
1044         {"de_DE"        , C_ISO_8859_1  , C_ISO_8859_1},
1045         {"de_LU"        , C_ISO_8859_1  , C_ISO_8859_1},
1046         {"en_AU"        , C_ISO_8859_1  , C_ISO_8859_1},
1047         {"en_BW"        , C_ISO_8859_1  , C_ISO_8859_1},
1048         {"en_CA"        , C_ISO_8859_1  , C_ISO_8859_1},
1049         {"en_DK"        , C_ISO_8859_1  , C_ISO_8859_1},
1050         {"en_GB"        , C_ISO_8859_1  , C_ISO_8859_1},
1051         {"en_HK"        , C_ISO_8859_1  , C_ISO_8859_1},
1052         {"en_IE"        , C_ISO_8859_1  , C_ISO_8859_1},
1053         {"en_NZ"        , C_ISO_8859_1  , C_ISO_8859_1},
1054         {"en_PH"        , C_ISO_8859_1  , C_ISO_8859_1},
1055         {"en_SG"        , C_ISO_8859_1  , C_ISO_8859_1},
1056         {"en_US"        , C_ISO_8859_1  , C_ISO_8859_1},
1057         {"en_ZA"        , C_ISO_8859_1  , C_ISO_8859_1},
1058         {"en_ZW"        , C_ISO_8859_1  , C_ISO_8859_1},
1059         {"es_AR"        , C_ISO_8859_1  , C_ISO_8859_1},
1060         {"es_BO"        , C_ISO_8859_1  , C_ISO_8859_1},
1061         {"es_CL"        , C_ISO_8859_1  , C_ISO_8859_1},
1062         {"es_CO"        , C_ISO_8859_1  , C_ISO_8859_1},
1063         {"es_CR"        , C_ISO_8859_1  , C_ISO_8859_1},
1064         {"es_DO"        , C_ISO_8859_1  , C_ISO_8859_1},
1065         {"es_EC"        , C_ISO_8859_1  , C_ISO_8859_1},
1066         {"es_ES"        , C_ISO_8859_1  , C_ISO_8859_1},
1067         {"es_GT"        , C_ISO_8859_1  , C_ISO_8859_1},
1068         {"es_HN"        , C_ISO_8859_1  , C_ISO_8859_1},
1069         {"es_MX"        , C_ISO_8859_1  , C_ISO_8859_1},
1070         {"es_NI"        , C_ISO_8859_1  , C_ISO_8859_1},
1071         {"es_PA"        , C_ISO_8859_1  , C_ISO_8859_1},
1072         {"es_PE"        , C_ISO_8859_1  , C_ISO_8859_1},
1073         {"es_PR"        , C_ISO_8859_1  , C_ISO_8859_1},
1074         {"es_PY"        , C_ISO_8859_1  , C_ISO_8859_1},
1075         {"es_SV"        , C_ISO_8859_1  , C_ISO_8859_1},
1076         {"es_US"        , C_ISO_8859_1  , C_ISO_8859_1},
1077         {"es_UY"        , C_ISO_8859_1  , C_ISO_8859_1},
1078         {"es_VE"        , C_ISO_8859_1  , C_ISO_8859_1},
1079         {"et_EE"        , C_ISO_8859_1  , C_ISO_8859_1},
1080         {"eu_ES"        , C_ISO_8859_1  , C_ISO_8859_1},
1081         {"fi_FI"        , C_ISO_8859_1  , C_ISO_8859_1},
1082         {"fo_FO"        , C_ISO_8859_1  , C_ISO_8859_1},
1083         {"fr_BE"        , C_ISO_8859_1  , C_ISO_8859_1},
1084         {"fr_CA"        , C_ISO_8859_1  , C_ISO_8859_1},
1085         {"fr_CH"        , C_ISO_8859_1  , C_ISO_8859_1},
1086         {"fr_FR"        , C_ISO_8859_1  , C_ISO_8859_1},
1087         {"fr_LU"        , C_ISO_8859_1  , C_ISO_8859_1},
1088         {"ga_IE"        , C_ISO_8859_1  , C_ISO_8859_1},
1089         {"gl_ES"        , C_ISO_8859_1  , C_ISO_8859_1},
1090         {"gv_GB"        , C_ISO_8859_1  , C_ISO_8859_1},
1091         {"id_ID"        , C_ISO_8859_1  , C_ISO_8859_1},
1092         {"is_IS"        , C_ISO_8859_1  , C_ISO_8859_1},
1093         {"it_CH"        , C_ISO_8859_1  , C_ISO_8859_1},
1094         {"it_IT"        , C_ISO_8859_1  , C_ISO_8859_1},
1095         {"kl_GL"        , C_ISO_8859_1  , C_ISO_8859_1},
1096         {"kw_GB"        , C_ISO_8859_1  , C_ISO_8859_1},
1097         {"ms_MY"        , C_ISO_8859_1  , C_ISO_8859_1},
1098         {"nl_BE"        , C_ISO_8859_1  , C_ISO_8859_1},
1099         {"nl_NL"        , C_ISO_8859_1  , C_ISO_8859_1},
1100         {"nn_NO"        , C_ISO_8859_1  , C_ISO_8859_1},
1101         {"no_NO"        , C_ISO_8859_1  , C_ISO_8859_1},
1102         {"oc_FR"        , C_ISO_8859_1  , C_ISO_8859_1},
1103         {"pt_BR"        , C_ISO_8859_1  , C_ISO_8859_1},
1104         {"pt_PT"        , C_ISO_8859_1  , C_ISO_8859_1},
1105         {"sq_AL"        , C_ISO_8859_1  , C_ISO_8859_1},
1106         {"sv_FI"        , C_ISO_8859_1  , C_ISO_8859_1},
1107         {"sv_SE"        , C_ISO_8859_1  , C_ISO_8859_1},
1108         {"tl_PH"        , C_ISO_8859_1  , C_ISO_8859_1},
1109         {"uz_UZ"        , C_ISO_8859_1  , C_ISO_8859_1},
1110         {"wa_BE"        , C_ISO_8859_1  , C_ISO_8859_1},
1111
1112         {"bs_BA"        , C_ISO_8859_2  , C_ISO_8859_2},
1113         {"cs_CZ"        , C_ISO_8859_2  , C_ISO_8859_2},
1114         {"hr_HR"        , C_ISO_8859_2  , C_ISO_8859_2},
1115         {"hu_HU"        , C_ISO_8859_2  , C_ISO_8859_2},
1116         {"pl_PL"        , C_ISO_8859_2  , C_ISO_8859_2},
1117         {"ro_RO"        , C_ISO_8859_2  , C_ISO_8859_2},
1118         {"sk_SK"        , C_ISO_8859_2  , C_ISO_8859_2},
1119         {"sl_SI"        , C_ISO_8859_2  , C_ISO_8859_2},
1120
1121         {"sr_YU@cyrillic"       , C_ISO_8859_5  , C_ISO_8859_5},
1122         {"sr_YU"                , C_ISO_8859_2  , C_ISO_8859_2},
1123
1124         {"mt_MT"                , C_ISO_8859_3  , C_ISO_8859_3},
1125
1126         {"lt_LT.iso88594"       , C_ISO_8859_4  , C_ISO_8859_4},
1127         {"lt_LT.ISO8859-4"      , C_ISO_8859_4  , C_ISO_8859_4},
1128         {"lt_LT.ISO_8859-4"     , C_ISO_8859_4  , C_ISO_8859_4},
1129         {"lt_LT"                , C_ISO_8859_13 , C_ISO_8859_13},
1130
1131         {"mk_MK"        , C_ISO_8859_5  , C_ISO_8859_5},
1132
1133         {"ar_AE"        , C_ISO_8859_6  , C_ISO_8859_6},
1134         {"ar_BH"        , C_ISO_8859_6  , C_ISO_8859_6},
1135         {"ar_DZ"        , C_ISO_8859_6  , C_ISO_8859_6},
1136         {"ar_EG"        , C_ISO_8859_6  , C_ISO_8859_6},
1137         {"ar_IQ"        , C_ISO_8859_6  , C_ISO_8859_6},
1138         {"ar_JO"        , C_ISO_8859_6  , C_ISO_8859_6},
1139         {"ar_KW"        , C_ISO_8859_6  , C_ISO_8859_6},
1140         {"ar_LB"        , C_ISO_8859_6  , C_ISO_8859_6},
1141         {"ar_LY"        , C_ISO_8859_6  , C_ISO_8859_6},
1142         {"ar_MA"        , C_ISO_8859_6  , C_ISO_8859_6},
1143         {"ar_OM"        , C_ISO_8859_6  , C_ISO_8859_6},
1144         {"ar_QA"        , C_ISO_8859_6  , C_ISO_8859_6},
1145         {"ar_SA"        , C_ISO_8859_6  , C_ISO_8859_6},
1146         {"ar_SD"        , C_ISO_8859_6  , C_ISO_8859_6},
1147         {"ar_SY"        , C_ISO_8859_6  , C_ISO_8859_6},
1148         {"ar_TN"        , C_ISO_8859_6  , C_ISO_8859_6},
1149         {"ar_YE"        , C_ISO_8859_6  , C_ISO_8859_6},
1150
1151         {"el_GR"        , C_ISO_8859_7  , C_ISO_8859_7},
1152         {"he_IL"        , C_ISO_8859_8  , C_ISO_8859_8},
1153         {"iw_IL"        , C_ISO_8859_8  , C_ISO_8859_8},
1154         {"tr_TR"        , C_ISO_8859_9  , C_ISO_8859_9},
1155
1156         {"lv_LV"        , C_ISO_8859_13 , C_ISO_8859_13},
1157         {"mi_NZ"        , C_ISO_8859_13 , C_ISO_8859_13},
1158
1159         {"cy_GB"        , C_ISO_8859_14 , C_ISO_8859_14},
1160
1161         {"ar_IN"        , C_UTF_8       , C_UTF_8},
1162         {"en_IN"        , C_UTF_8       , C_UTF_8},
1163         {"se_NO"        , C_UTF_8       , C_UTF_8},
1164         {"ta_IN"        , C_UTF_8       , C_UTF_8},
1165         {"te_IN"        , C_UTF_8       , C_UTF_8},
1166         {"ur_PK"        , C_UTF_8       , C_UTF_8},
1167
1168         {"th_TH"        , C_TIS_620     , C_TIS_620},
1169         /* {"th_TH"     , C_WINDOWS_874}, */
1170         /* {"th_TH"     , C_ISO_8859_11}, */
1171
1172         {"ka_GE"        , C_GEORGIAN_PS , C_GEORGIAN_PS},
1173         {"vi_VN.TCVN"   , C_TCVN5712_1  , C_TCVN5712_1},
1174
1175         {"C"                    , C_US_ASCII    , C_US_ASCII},
1176         {"POSIX"                , C_US_ASCII    , C_US_ASCII},
1177         {"ANSI_X3.4-1968"       , C_US_ASCII    , C_US_ASCII},
1178 };
1179
1180 static GHashTable *conv_get_charset_to_str_table(void)
1181 {
1182         static GHashTable *table;
1183         gint i;
1184
1185         if (table)
1186                 return table;
1187
1188         table = g_hash_table_new(NULL, g_direct_equal);
1189
1190         for (i = 0; i < sizeof(charsets) / sizeof(charsets[0]); i++) {
1191                 if (g_hash_table_lookup(table, GUINT_TO_POINTER(charsets[i].charset))
1192                     == NULL) {
1193                         g_hash_table_insert
1194                                 (table, GUINT_TO_POINTER(charsets[i].charset),
1195                                  charsets[i].name);
1196                 }
1197         }
1198
1199         return table;
1200 }
1201
1202 static GHashTable *conv_get_charset_from_str_table(void)
1203 {
1204         static GHashTable *table;
1205         gint i;
1206
1207         if (table)
1208                 return table;
1209
1210         table = g_hash_table_new(str_case_hash, str_case_equal);
1211
1212         for (i = 0; i < sizeof(charsets) / sizeof(charsets[0]); i++) {
1213                 g_hash_table_insert(table, charsets[i].name,
1214                                     GUINT_TO_POINTER(charsets[i].charset));
1215         }
1216
1217         return table;
1218 }
1219
1220 const gchar *conv_get_charset_str(CharSet charset)
1221 {
1222         GHashTable *table;
1223
1224         table = conv_get_charset_to_str_table();
1225         return g_hash_table_lookup(table, GUINT_TO_POINTER(charset));
1226 }
1227
1228 CharSet conv_get_charset_from_str(const gchar *charset)
1229 {
1230         GHashTable *table;
1231
1232         if (!charset) return C_AUTO;
1233
1234         table = conv_get_charset_from_str_table();
1235         return GPOINTER_TO_UINT(g_hash_table_lookup(table, charset));
1236 }
1237
1238 static CharSet conv_get_locale_charset(void)
1239 {
1240         static CharSet cur_charset = -1;
1241         const gchar *cur_locale;
1242         const gchar *p;
1243         gint i;
1244
1245         if (cur_charset != -1)
1246                 return cur_charset;
1247
1248         cur_locale = conv_get_current_locale();
1249         if (!cur_locale) {
1250                 cur_charset = C_US_ASCII;
1251                 return cur_charset;
1252         }
1253
1254         if (strcasestr(cur_locale, "UTF-8")) {
1255                 cur_charset = C_UTF_8;
1256                 return cur_charset;
1257         }
1258
1259         if ((p = strcasestr(cur_locale, "@euro")) && p[5] == '\0') {
1260                 cur_charset = C_ISO_8859_15;
1261                 return cur_charset;
1262         }
1263
1264         for (i = 0; i < sizeof(locale_table) / sizeof(locale_table[0]); i++) {
1265                 const gchar *p;
1266
1267                 /* "ja_JP.EUC" matches with "ja_JP.eucJP", "ja_JP.EUC" and
1268                    "ja_JP". "ja_JP" matches with "ja_JP.xxxx" and "ja" */
1269                 if (!g_ascii_strncasecmp(cur_locale, locale_table[i].locale,
1270                                  strlen(locale_table[i].locale))) {
1271                         cur_charset = locale_table[i].charset;
1272                         return cur_charset;
1273                 } else if ((p = strchr(locale_table[i].locale, '_')) &&
1274                          !strchr(p + 1, '.')) {
1275                         if (strlen(cur_locale) == 2 &&
1276                             !g_ascii_strncasecmp(cur_locale, locale_table[i].locale, 2)) {
1277                                 cur_charset = locale_table[i].charset;
1278                                 return cur_charset;
1279                         }
1280                 }
1281         }
1282
1283         cur_charset = C_AUTO;
1284         return cur_charset;
1285 }
1286
1287 static CharSet conv_get_locale_charset_no_utf8(void)
1288 {
1289         static CharSet cur_charset = -1;
1290         const gchar *cur_locale;
1291         const gchar *p;
1292         gchar *tmp;
1293         gint i;
1294
1295         if (prefs_common.broken_are_utf8)
1296                 return conv_get_locale_charset();
1297
1298         if (cur_charset != -1)
1299                 return cur_charset;
1300
1301         cur_locale = conv_get_current_locale();
1302         if (!cur_locale) {
1303                 cur_charset = C_US_ASCII;
1304                 return cur_charset;
1305         }
1306
1307         if (strcasestr(cur_locale, "UTF-8")) {
1308                 tmp = g_strdup(cur_locale);
1309                 *(strcasestr(tmp, ".UTF-8")) = '\0';
1310                 cur_locale = tmp;
1311         }
1312
1313         if ((p = strcasestr(cur_locale, "@euro")) && p[5] == '\0') {
1314                 cur_charset = C_ISO_8859_15;
1315                 return cur_charset;
1316         }
1317
1318         for (i = 0; i < sizeof(locale_table) / sizeof(locale_table[0]); i++) {
1319                 const gchar *p;
1320
1321                 /* "ja_JP.EUC" matches with "ja_JP.eucJP", "ja_JP.EUC" and
1322                    "ja_JP". "ja_JP" matches with "ja_JP.xxxx" and "ja" */
1323                 if (!g_ascii_strncasecmp(cur_locale, locale_table[i].locale,
1324                                  strlen(locale_table[i].locale))) {
1325                         cur_charset = locale_table[i].charset;
1326                         return cur_charset;
1327                 } else if ((p = strchr(locale_table[i].locale, '_')) &&
1328                          !strchr(p + 1, '.')) {
1329                         if (strlen(cur_locale) == 2 &&
1330                             !g_ascii_strncasecmp(cur_locale, locale_table[i].locale, 2)) {
1331                                 cur_charset = locale_table[i].charset;
1332                                 return cur_charset;
1333                         }
1334                 }
1335         }
1336
1337         cur_charset = C_AUTO;
1338         return cur_charset;
1339 }
1340
1341 const gchar *conv_get_locale_charset_str(void)
1342 {
1343         static const gchar *codeset = NULL;
1344
1345         if (!codeset)
1346                 codeset = conv_get_charset_str(conv_get_locale_charset());
1347
1348         return codeset ? codeset : CS_INTERNAL;
1349 }
1350
1351 const gchar *conv_get_locale_charset_str_no_utf8(void)
1352 {
1353         static const gchar *codeset = NULL;
1354
1355         if (!codeset)
1356                 codeset = conv_get_charset_str(conv_get_locale_charset_no_utf8());
1357
1358         return codeset ? codeset : CS_INTERNAL;
1359 }
1360
1361 static CharSet conv_get_outgoing_charset(void)
1362 {
1363         static CharSet out_charset = -1;
1364         const gchar *cur_locale;
1365         const gchar *p;
1366         gint i;
1367
1368         if (out_charset != -1)
1369                 return out_charset;
1370
1371         cur_locale = conv_get_current_locale();
1372         if (!cur_locale) {
1373                 out_charset = C_AUTO;
1374                 return out_charset;
1375         }
1376
1377         if (strcasestr(cur_locale, "UTF-8")) {
1378                 out_charset = C_UTF_8;
1379                 return out_charset;
1380         }
1381
1382         if ((p = strcasestr(cur_locale, "@euro")) && p[5] == '\0') {
1383                 out_charset = C_ISO_8859_15;
1384                 return out_charset;
1385         }
1386
1387         for (i = 0; i < sizeof(locale_table) / sizeof(locale_table[0]); i++) {
1388                 const gchar *p;
1389
1390                 if (!g_ascii_strncasecmp(cur_locale, locale_table[i].locale,
1391                                  strlen(locale_table[i].locale))) {
1392                         out_charset = locale_table[i].out_charset;
1393                         break;
1394                 } else if ((p = strchr(locale_table[i].locale, '_')) &&
1395                          !strchr(p + 1, '.')) {
1396                         if (strlen(cur_locale) == 2 &&
1397                             !g_ascii_strncasecmp(cur_locale, locale_table[i].locale, 2)) {
1398                                 out_charset = locale_table[i].out_charset;
1399                                 break;
1400                         }
1401                 }
1402         }
1403
1404         return out_charset;
1405 }
1406
1407 const gchar *conv_get_outgoing_charset_str(void)
1408 {
1409         CharSet out_charset;
1410         const gchar *str;
1411
1412         out_charset = conv_get_outgoing_charset();
1413         str = conv_get_charset_str(out_charset);
1414
1415         return str ? str : CS_UTF_8;
1416 }
1417
1418 const gchar *conv_get_current_locale(void)
1419 {
1420         const gchar *cur_locale;
1421
1422 #ifdef G_OS_WIN32
1423         cur_locale = g_win32_getlocale();
1424 #else
1425         cur_locale = g_getenv("LC_ALL");
1426         if (!cur_locale) cur_locale = g_getenv("LC_CTYPE");
1427         if (!cur_locale) cur_locale = g_getenv("LANG");
1428         if (!cur_locale) cur_locale = setlocale(LC_CTYPE, NULL);
1429 #endif /* G_OS_WIN32 */
1430
1431         debug_print("current locale: %s\n",
1432                     cur_locale ? cur_locale : "(none)");
1433
1434         return cur_locale;
1435 }
1436
1437 static gboolean conv_is_ja_locale(void)
1438 {
1439         static gint is_ja_locale = -1;
1440         const gchar *cur_locale;
1441
1442         if (is_ja_locale != -1)
1443                 return is_ja_locale != 0;
1444
1445         is_ja_locale = 0;
1446         cur_locale = conv_get_current_locale();
1447         if (cur_locale) {
1448                 if (g_ascii_strncasecmp(cur_locale, "ja", 2) == 0)
1449                         is_ja_locale = 1;
1450         }
1451
1452         return is_ja_locale != 0;
1453 }
1454
1455 gchar *conv_unmime_header(const gchar *str, const gchar *default_encoding)
1456 {
1457         gchar buf[BUFFSIZE];
1458
1459         if (is_ascii_str(str))
1460                 return unmime_header(str);
1461
1462         if (default_encoding) {
1463                 gchar *utf8_buf;
1464
1465                 utf8_buf = conv_codeset_strdup
1466                         (str, default_encoding, CS_INTERNAL);
1467                 if (utf8_buf) {
1468                         gchar *decoded_str;
1469
1470                         decoded_str = unmime_header(utf8_buf);
1471                         g_free(utf8_buf);
1472                         return decoded_str;
1473                 }
1474         }
1475
1476         if (conv_is_ja_locale())
1477                 conv_anytodisp(buf, sizeof(buf), str);
1478         else
1479                 conv_localetodisp(buf, sizeof(buf), str);
1480
1481         return unmime_header(buf);
1482 }
1483
1484 #define MAX_LINELEN             76
1485 #define MAX_HARD_LINELEN        996
1486 #define MIMESEP_BEGIN           "=?"
1487 #define MIMESEP_END             "?="
1488
1489 #define LBREAK_IF_REQUIRED(cond, is_plain_text)                         \
1490 {                                                                       \
1491         if (len - (destp - (guchar *)dest) < MAX_LINELEN + 2) {         \
1492                 *destp = '\0';                                          \
1493                 return;                                                 \
1494         }                                                               \
1495                                                                         \
1496         if ((cond) && *srcp) {                                          \
1497                 if (destp > (guchar *)dest && left < MAX_LINELEN - 1) { \
1498                         if (isspace(*(destp - 1)))                      \
1499                                 destp--;                                \
1500                         else if (is_plain_text && isspace(*srcp))       \
1501                                 srcp++;                                 \
1502                         if (*srcp) {                                    \
1503                                 *destp++ = '\n';                        \
1504                                 *destp++ = ' ';                         \
1505                                 left = MAX_LINELEN - 1;                 \
1506                         }                                               \
1507                 }                                                       \
1508         }                                                               \
1509 }
1510
1511 void conv_encode_header_full(gchar *dest, gint len, const gchar *src,
1512                         gint header_len, gboolean addr_field,
1513                         const gchar *out_encoding_)
1514 {
1515         const gchar *cur_encoding;
1516         const gchar *out_encoding;
1517         gint mimestr_len;
1518         gchar *mimesep_enc;
1519         gint left;
1520         const guchar *srcp = src;
1521         guchar *destp = dest;
1522         gboolean use_base64;
1523
1524         g_return_if_fail(g_utf8_validate(src, -1, NULL) == TRUE);
1525
1526         if (MB_CUR_MAX > 1) {
1527                 use_base64 = TRUE;
1528                 mimesep_enc = "?B?";
1529         } else {
1530                 use_base64 = FALSE;
1531                 mimesep_enc = "?Q?";
1532         }
1533
1534         cur_encoding = CS_INTERNAL;
1535
1536         if (out_encoding_)
1537                 out_encoding = out_encoding_;
1538         else
1539                 out_encoding = conv_get_outgoing_charset_str();
1540
1541         if (!strcmp(out_encoding, CS_US_ASCII))
1542                 out_encoding = CS_ISO_8859_1;
1543
1544         mimestr_len = strlen(MIMESEP_BEGIN) + strlen(out_encoding) +
1545                 strlen(mimesep_enc) + strlen(MIMESEP_END);
1546
1547         left = MAX_LINELEN - header_len;
1548
1549         while (*srcp) {
1550                 LBREAK_IF_REQUIRED(left <= 0, TRUE);
1551
1552                 while (isspace(*srcp)) {
1553                         *destp++ = *srcp++;
1554                         left--;
1555                         LBREAK_IF_REQUIRED(left <= 0, TRUE);
1556                 }
1557
1558                 /* output as it is if the next word is ASCII string */
1559                 if (!is_next_nonascii(srcp)) {
1560                         gint word_len;
1561
1562                         word_len = get_next_word_len(srcp);
1563                         LBREAK_IF_REQUIRED(left < word_len, TRUE);
1564                         while (word_len > 0) {
1565                                 LBREAK_IF_REQUIRED(left + (MAX_HARD_LINELEN - MAX_LINELEN) <= 0, TRUE)
1566                                 *destp++ = *srcp++;
1567                                 left--;
1568                                 word_len--;
1569                         }
1570
1571                         continue;
1572                 }
1573
1574                 /* don't include parentheses in encoded strings */
1575                 if (addr_field && (*srcp == '(' || *srcp == ')')) {
1576                         LBREAK_IF_REQUIRED(left < 2, FALSE);
1577                         *destp++ = *srcp++;
1578                         left--;
1579                 }
1580
1581                 while (1) {
1582                         gint mb_len = 0;
1583                         gint cur_len = 0;
1584                         gchar *part_str;
1585                         gchar *out_str;
1586                         gchar *enc_str;
1587                         const guchar *p = srcp;
1588                         gint out_str_len;
1589                         gint out_enc_str_len;
1590                         gint mime_block_len;
1591                         gboolean cont = FALSE;
1592
1593                         while (*p != '\0') {
1594                                 if (isspace(*p) && !is_next_nonascii(p + 1))
1595                                         break;
1596                                 /* don't include parentheses in encoded
1597                                    strings */
1598                                 if (addr_field && (*p == '(' || *p == ')'))
1599                                         break;
1600
1601                                 mb_len = g_utf8_skip[*p];
1602
1603                                 Xstrndup_a(part_str, srcp, cur_len + mb_len, );
1604                                 out_str = conv_codeset_strdup
1605                                         (part_str, cur_encoding, out_encoding);
1606                                 if (!out_str) {
1607                                         if (strict_mode) {
1608                                                 *dest = '\0';
1609                                                 return;
1610                                         } else {
1611                                                 g_warning("conv_encode_header(): code conversion failed\n");
1612                                                 conv_unreadable_8bit(part_str);
1613                                                 out_str = g_strdup(part_str);
1614                                         }
1615                                 }
1616                                 out_str_len = strlen(out_str);
1617
1618                                 if (use_base64)
1619                                         out_enc_str_len = B64LEN(out_str_len);
1620                                 else
1621                                         out_enc_str_len =
1622                                                 qp_get_q_encoding_len(out_str);
1623
1624                                 g_free(out_str);
1625
1626                                 if (mimestr_len + out_enc_str_len <= left) {
1627                                         cur_len += mb_len;
1628                                         p += mb_len;
1629                                 } else if (cur_len == 0) {
1630                                         LBREAK_IF_REQUIRED(1, FALSE);
1631                                         continue;
1632                                 } else {
1633                                         cont = TRUE;
1634                                         break;
1635                                 }
1636                         }
1637
1638                         if (cur_len > 0) {
1639                                 Xstrndup_a(part_str, srcp, cur_len, );
1640                                 out_str = conv_codeset_strdup
1641                                         (part_str, cur_encoding, out_encoding);
1642                                 if (!out_str) {
1643                                         g_warning("conv_encode_header(): code conversion failed\n");
1644                                         conv_unreadable_8bit(part_str);
1645                                         out_str = g_strdup(part_str);
1646                                 }
1647                                 out_str_len = strlen(out_str);
1648
1649                                 if (use_base64)
1650                                         out_enc_str_len = B64LEN(out_str_len);
1651                                 else
1652                                         out_enc_str_len =
1653                                                 qp_get_q_encoding_len(out_str);
1654
1655                                 Xalloca(enc_str, out_enc_str_len + 1, );
1656                                 if (use_base64)
1657                                         base64_encode(enc_str, out_str, out_str_len);
1658                                 else
1659                                         qp_q_encode(enc_str, out_str);
1660
1661                                 g_free(out_str);
1662
1663                                 /* output MIME-encoded string block */
1664                                 mime_block_len = mimestr_len + strlen(enc_str);
1665                                 g_snprintf(destp, mime_block_len + 1,
1666                                            MIMESEP_BEGIN "%s%s%s" MIMESEP_END,
1667                                            out_encoding, mimesep_enc, enc_str);
1668                                 destp += mime_block_len;
1669                                 srcp += cur_len;
1670
1671                                 left -= mime_block_len;
1672                         }
1673
1674                         LBREAK_IF_REQUIRED(cont, FALSE);
1675
1676                         if (cur_len == 0)
1677                                 break;
1678                 }
1679         }
1680
1681         *destp = '\0';
1682 }
1683
1684 void conv_encode_header(gchar *dest, gint len, const gchar *src,
1685                         gint header_len, gboolean addr_field)
1686 {
1687         conv_encode_header_full(dest,len,src,header_len,addr_field,NULL);
1688 }
1689
1690 #undef LBREAK_IF_REQUIRED
1691 gchar *conv_filename_from_utf8(const gchar *utf8_file)
1692 {
1693         gchar *fs_file;
1694         GError *error = NULL;
1695
1696         fs_file = g_filename_from_utf8(utf8_file, -1, NULL, NULL, &error);
1697         if (error) {
1698                 g_warning("failed to convert encoding of file name: %s\n",
1699                           error->message);
1700                 g_error_free(error);
1701         }
1702         if (!fs_file)
1703                 fs_file = g_strdup(utf8_file);
1704
1705         return fs_file;
1706 }
1707
1708 gchar *conv_filename_to_utf8(const gchar *fs_file)
1709 {
1710         gchar *utf8_file = NULL;
1711         GError *error = NULL;
1712
1713         utf8_file = g_filename_to_utf8(fs_file, -1, NULL, NULL, &error);
1714         if (error) {
1715                 g_warning("failed to convert encoding of file name: %s\n",
1716                           error->message);
1717                 g_error_free(error);
1718         }
1719
1720         if (!utf8_file || !g_utf8_validate(utf8_file, -1, NULL)) {
1721                 g_free(utf8_file);
1722                 utf8_file = g_strdup(fs_file);
1723                 conv_unreadable_8bit(utf8_file);
1724         }
1725
1726         return utf8_file;
1727 }