src/codeconv.c

   1 /*
   2  * Sylpheed -- a GTK+ based, lightweight, and fast e-mail client
   3  * Copyright (C) 1999-2001 Hiroyuki Yamamoto
   4  *
   5  * This program is free software; you can redistribute it and/or modify
   6  * it under the terms of the GNU General Public License as published by
   7  * the Free Software Foundation; either version 2 of the License, or
   8  * (at your option) any later version.
   9  *
  10  * This program is distributed in the hope that it will be useful,
  11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  13  * GNU General Public License for more details.
  14  *
  15  * You should have received a copy of the GNU General Public License
  16  * along with this program; if not, write to the Free Software
  17  * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
  18  */
  19
  20 #ifdef HAVE_CONFIG_H
  21 #  include "config.h"
  22 #endif
  23
  24 #include <glib.h>
  25 #include <string.h>
  26 #include <ctype.h>
  27 #include <stdlib.h>
  28
  29 #if (HAVE_WCTYPE_H && HAVE_WCHAR_H)
  30 #  include <wchar.h>
  31 #  include <wctype.h>
  32 #endif
  33
  34 #if HAVE_LOCALE_H
  35 #  include <locale.h>
  36 #endif
  37
  38 #if HAVE_LIBJCONV
  39 #  include <jconv.h>
  40 #endif
  41
  42 #include <kcc.h>
  43
  44 #include "intl.h"
  45 #include "codeconv.h"
  46 #include "unmime.h"
  47 #include "base64.h"
  48 #include "utils.h"
  49 #include "prefs_common.h"
  50
  51 #define iskanji(c) \
  52         (((c) & 0xff) >= 0xa1 && ((c) & 0xff) <= 0xfe)
  53 #define iseucss(c) \
  54         (((c) & 0xff) == 0x8e || ((c) & 0xff) == 0x8f)
  55 #define isunprintablekanji(c) \
  56         (((c) & 0xff) >= 0xa9 && ((c) & 0xff) <= 0xaf)
  57
  58 void conv_jistoeuc(gchar *outbuf, gint outlen, const gchar *inbuf)
  59 {
  60         KCC_filter(outbuf, "EUC", (gchar *)inbuf, "JISBB", 0, 0, 0);
  61 }
  62
  63 void conv_euctojis(gchar *outbuf, gint outlen, const gchar *inbuf)
  64 {
  65         size_t inlen, len;
  66
  67         inlen = strlen(inbuf);
  68         if (iskanji(inbuf[inlen - 1]) || iseucss(inbuf[inlen - 1])) {
  69                 /* if tail end of the string is not ended with ascii,
  70                    add dummy return code. */
  71                 gchar *tmpin, *tmpout;
  72
  73                 /* length of original string + '\n' + '\0' */
  74                 tmpin = alloca(inlen + 2);
  75                 if (tmpin == NULL) {
  76                         g_warning(_("can't allocate memory\n"));
  77                         KCC_filter(outbuf, "JISBB", (gchar *)inbuf, "EUC",
  78                                    0, 0, 0);
  79                         return;
  80                 }
  81                 strcpy(tmpin, inbuf);
  82                 tmpin[inlen] = '\n';
  83                 tmpin[inlen + 1] = '\0';
  84
  85                 tmpout = alloca(outlen + 1);
  86                 if (tmpout == NULL) {
  87                         g_warning(_("can't allocate memory\n"));
  88                         KCC_filter(outbuf, "JISBB", (gchar *)inbuf, "EUC",
  89                                    0, 0, 0);
  90                         return;
  91                 }
  92
  93                 KCC_filter(tmpout, "JISBB", tmpin, "EUC", 0, 0, 0);
  94                 len = strlen(tmpout);
  95                 if (tmpout[len - 1] == '\n')
  96                         tmpout[len - 1] = '\0';
  97                 strncpy2(outbuf, tmpout, outlen);
  98         } else
  99                 KCC_filter(outbuf, "JISBB", (gchar *)inbuf, "EUC", 0, 0, 0);
 100 }
 101
 102 void conv_sjistoeuc(gchar *outbuf, gint outlen, const gchar *inbuf)
 103 {
 104         KCC_filter(outbuf, "EUC", (gchar *)inbuf, "SJIS", 0, 0, 0);
 105 }
 106
 107 void conv_anytoeuc(gchar *outbuf, gint outlen, const gchar *inbuf)
 108 {
 109         KCC_filter(outbuf, "EUC", (gchar *)inbuf, "AUTO", 0, 0, 0);
 110 }
 111
 112 void conv_anytojis(gchar *outbuf, gint outlen, const gchar *inbuf)
 113 {
 114         KCC_filter(outbuf, "JISBB", (gchar *)inbuf, "AUTO", 0, 0, 0);
 115 }
 116
 117 #define SUBST_CHAR      '_'
 118
 119 void conv_unreadable_eucjp(gchar *str)
 120 {
 121         register guchar *p = str;
 122
 123         while (*p != '\0') {
 124                 if (isascii(*p)) {
 125                         /* convert CR+LF -> LF */
 126                         if (*p == '\r' && *(p + 1) == '\n')
 127                                 memmove(p, p + 1, strlen(p));
 128                         /* printable 7 bit code */
 129                         p++;
 130                 } else if (iskanji(*p)) {
 131                         if (iskanji(*(p + 1)) && !isunprintablekanji(*p))
 132                                 /* printable euc-jp code */
 133                                 p += 2;
 134                         else {
 135                                 /* substitute unprintable code */
 136                                 *p++ = SUBST_CHAR;
 137                                 if (*p != '\0') {
 138                                         if (isascii(*p))
 139                                                 p++;
 140                                         else
 141                                                 *p++ = SUBST_CHAR;
 142                                 }
 143                         }
 144                 } else if (iseucss(*p)) {
 145                         if ((*(p + 1) & 0x80) != 0)
 146                                 /* euc-jp hankaku kana */
 147                                 p += 2;
 148                         else
 149                                 *p++ = SUBST_CHAR;
 150                 } else
 151                         /* substitute unprintable 1 byte code */
 152                         *p++ = SUBST_CHAR;
 153         }
 154 }
 155
 156 void conv_unreadable_8bit(gchar *str)
 157 {
 158         register guchar *p = str;
 159
 160         while (*p != '\0') {
 161                 /* convert CR+LF -> LF */
 162                 if (*p == '\r' && *(p + 1) == '\n')
 163                         memmove(p, p + 1, strlen(p));
 164                 else if (!isascii(*p)) *p = SUBST_CHAR;
 165                 p++;
 166         }
 167 }
 168
 169 void conv_unreadable_latin(gchar *str)
 170 {
 171         register guchar *p = str;
 172
 173         while (*p != '\0') {
 174                 /* convert CR+LF -> LF */
 175                 if (*p == '\r' && *(p + 1) == '\n')
 176                         memmove(p, p + 1, strlen(p));
 177                 else if ((*p & 0xff) >= 0x80 && (*p & 0xff) <= 0x9f)
 178                         *p = SUBST_CHAR;
 179                 p++;
 180         }
 181 }
 182
 183 #define NCV     '\0'
 184
 185 void conv_mb_alnum(gchar *str)
 186 {
 187         static guchar char_tbl[] = {
 188                 /* 0xa0 - 0xaf */
 189                 NCV, ' ', NCV, NCV, ',', '.', NCV, ':',
 190                 ';', '?', '!', NCV, NCV, NCV, NCV, NCV,
 191                 /* 0xb0 - 0xbf */
 192                 NCV, NCV, NCV, NCV, NCV, NCV, NCV, NCV,
 193                 NCV, NCV, NCV, NCV, NCV, NCV, NCV, NCV,
 194                 /* 0xc0 - 0xcf */
 195                 NCV, NCV, NCV, NCV, NCV, NCV, NCV, NCV,
 196                 NCV, NCV, '(', ')', NCV, NCV, '[', ']',
 197                 /* 0xd0 - 0xdf */
 198                 '{', '}', NCV, NCV, NCV, NCV, NCV, NCV,
 199                 NCV, NCV, NCV, NCV, '+', '-', NCV, NCV,
 200                 /* 0xe0 - 0xef */
 201                 NCV, '=', NCV, '<', '>', NCV, NCV, NCV,
 202                 NCV, NCV, NCV, NCV, NCV, NCV, NCV, NCV
 203         };
 204
 205         register guchar *p = str;
 206         register gint len;
 207
 208         len = strlen(str);
 209
 210         while (len > 1) {
 211                 if (*p == 0xa3) {
 212                         register guchar ch = *(p + 1);
 213
 214                         if (ch >= 0xb0 && ch <= 0xfa) {
 215                                 /* [a-zA-Z] */
 216                                 *p = ch & 0x7f;
 217                                 p++;
 218                                 len--;
 219                                 memmove(p, p + 1, len);
 220                                 len--;
 221                         } else  {
 222                                 p += 2;
 223                                 len -= 2;
 224                         }
 225                 } else if (*p == 0xa1) {
 226                         register guchar ch = *(p + 1);
 227
 228                         if (ch >= 0xa0 && ch <= 0xef &&
 229                             NCV != char_tbl[ch - 0xa0]) {
 230                                 *p = char_tbl[ch - 0xa0];
 231                                 p++;
 232                                 len--;
 233                                 memmove(p, p + 1, len);
 234                                 len--;
 235                         } else {
 236                                 p += 2;
 237                                 len -= 2;
 238                         }
 239                 } else if (iskanji(*p)) {
 240                         p += 2;
 241                         len -= 2;
 242                 } else {
 243                         p++;
 244                         len--;
 245                 }
 246         }
 247 }
 248
 249 void conv_jistodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
 250 {
 251         conv_jistoeuc(outbuf, outlen, inbuf);
 252         conv_unreadable_eucjp(outbuf);
 253 }
 254
 255 void conv_sjistodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
 256 {
 257         conv_sjistoeuc(outbuf, outlen, inbuf);
 258         conv_unreadable_eucjp(outbuf);
 259 }
 260
 261 void conv_euctodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
 262 {
 263         strncpy2(outbuf, inbuf, outlen);
 264         conv_unreadable_eucjp(outbuf);
 265 }
 266
 267 void conv_ustodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
 268 {
 269         strncpy2(outbuf, inbuf, outlen);
 270         conv_unreadable_8bit(outbuf);
 271 }
 272
 273 void conv_latintodisp(gchar *outbuf, gint outlen, const gchar *inbuf)
 274 {
 275         strncpy2(outbuf, inbuf, outlen);
 276         conv_unreadable_latin(outbuf);
 277 }
 278
 279 void conv_noconv(gchar *outbuf, gint outlen, const gchar *inbuf)
 280 {
 281         strncpy2(outbuf, inbuf, outlen);
 282 }
 283
 284 CodeConverter *conv_code_converter_new(const gchar *charset)
 285 {
 286         CodeConverter *conv;
 287
 288         conv = g_new0(CodeConverter, 1);
 289 #if !HAVE_LIBJCONV
 290         conv->code_conv_func = conv_get_code_conv_func(charset);
 291 #endif
 292         conv->charset_str = g_strdup(charset);
 293         conv->charset = conv_get_charset_from_str(charset);
 294
 295         return conv;
 296 }
 297
 298 void conv_code_converter_destroy(CodeConverter *conv)
 299 {
 300         g_free(conv->charset_str);
 301         g_free(conv);
 302 }
 303
 304 gint conv_convert(CodeConverter *conv, gchar *outbuf, gint outlen,
 305                   const gchar *inbuf)
 306 {
 307 #if HAVE_LIBJCONV
 308         gchar *str;
 309
 310         str = conv_codeset_strdup(inbuf, conv->charset_str, NULL);
 311         if (!str)
 312                 return -1;
 313         else {
 314                 strncpy2(outbuf, str, outlen);
 315                 g_free(str);
 316         }
 317 #else /* !HAVE_LIBJCONV */
 318         conv->code_conv_func(outbuf, outlen, inbuf);
 319 #endif
 320
 321         return 0;
 322 }
 323
 324 gchar *conv_codeset_strdup(const gchar *inbuf,
 325                            const gchar *src_codeset, const gchar *dest_codeset)
 326 {
 327         gchar *buf;
 328         size_t len;
 329 #if HAVE_LIBJCONV
 330         gint actual_codeset;
 331         const gchar *const *codesets;
 332         gint n_codesets;
 333 #else /* !HAVE_LIBJCONV */
 334         CharSet src_charset = C_AUTO, dest_charset = C_AUTO;
 335 #endif
 336
 337         if (!dest_codeset) {
 338                 CodeConvFunc func;
 339
 340                 func = conv_get_code_conv_func(src_codeset);
 341                 if (func != conv_noconv) {
 342                         if (func == conv_jistodisp || func == conv_sjistodisp)
 343                                 len = strlen(inbuf) * 2 + 1;
 344                         else
 345                                 len = strlen(inbuf) + 1;
 346                         buf = g_malloc(len);
 347                         if (!buf) return NULL;
 348                         func(buf, len, inbuf);
 349                         buf = g_realloc(buf, strlen(buf) + 1);
 350                         return buf;
 351                 }
 352         }
 353
 354         /* don't convert if src and dest codeset are identical */
 355         if (src_codeset && dest_codeset &&
 356             !strcasecmp(src_codeset, dest_codeset))
 357                 return g_strdup(inbuf);
 358
 359 #if HAVE_LIBJCONV
 360         if (src_codeset) {
 361                 codesets = &src_codeset;
 362                 n_codesets = 1;
 363         } else
 364                 codesets = jconv_info_get_pref_codesets(&n_codesets);
 365         if (!dest_codeset) {
 366                 dest_codeset = conv_get_current_charset_str();
 367                 /* don't convert if current codeset is US-ASCII */
 368                 if (!strcasecmp(dest_codeset, CS_US_ASCII))
 369                         return g_strdup(inbuf);
 370         }
 371
 372         if (jconv_alloc_conv(inbuf, strlen(inbuf), &buf, &len,
 373                              codesets, n_codesets,
 374                              &actual_codeset, dest_codeset)
 375             == 0)
 376                 return buf;
 377         else
 378                 return NULL;
 379 #else /* !HAVE_LIBJCONV */
 380         if (src_codeset) {
 381                 if (!strcasecmp(src_codeset, CS_EUC_JP) ||
 382                     !strcasecmp(src_codeset, "EUCJP"))
 383                         src_charset = C_EUC_JP;
 384                 else if (!strcasecmp(src_codeset, CS_SHIFT_JIS) ||
 385                          !strcasecmp(src_codeset, "SHIFT-JIS") ||
 386                          !strcasecmp(src_codeset, "SJIS"))
 387                         src_charset = C_SHIFT_JIS;
 388                 if (dest_codeset && !strcasecmp(dest_codeset, CS_ISO_2022_JP))
 389                         dest_charset = C_ISO_2022_JP;
 390         }
 391
 392         if ((src_charset == C_EUC_JP || src_charset == C_SHIFT_JIS) &&
 393             dest_charset == C_ISO_2022_JP) {
 394                 len = (strlen(inbuf) + 1) * 3;
 395                 buf = g_malloc(len);
 396                 if (buf) {
 397                         if (src_charset == C_EUC_JP)
 398                                 conv_euctojis(buf, len, inbuf);
 399                         else
 400                                 conv_anytojis(buf, len, inbuf);
 401                         buf = g_realloc(buf, strlen(buf) + 1);
 402                 }
 403         } else
 404                 buf = g_strdup(inbuf);
 405
 406         return buf;
 407 #endif /* !HAVE_LIBJCONV */
 408 }
 409
 410 CodeConvFunc conv_get_code_conv_func(const gchar *charset)
 411 {
 412         CodeConvFunc code_conv;
 413
 414         if (!charset) {
 415                 if (conv_get_outgoing_charset() == C_ISO_2022_JP)
 416                         return conv_jistodisp;
 417                 else
 418                         return conv_noconv;
 419         }
 420
 421         if (!strcasecmp(charset, CS_ISO_2022_JP) ||
 422             !strcasecmp(charset, CS_ISO_2022_JP_2))
 423                 code_conv = conv_jistodisp;
 424         else if (!strcasecmp(charset, CS_US_ASCII))
 425                 code_conv = conv_ustodisp;
 426         else if (!strncasecmp(charset, CS_ISO_8859_1, 10))
 427                 code_conv = conv_latintodisp;
 428 #if !HAVE_LIBJCONV
 429         else if (!strncasecmp(charset, "ISO-8859-", 9))
 430                 code_conv = conv_latintodisp;
 431 #endif
 432         else if (!strcasecmp(charset, CS_SHIFT_JIS) ||
 433                  !strcasecmp(charset, "SHIFT-JIS")  ||
 434                  !strcasecmp(charset, "SJIS")       ||
 435                  !strcasecmp(charset, "X-SJIS"))
 436                 code_conv = conv_sjistodisp;
 437         else if (!strcasecmp(charset, CS_EUC_JP) ||
 438                  !strcasecmp(charset, "EUCJP"))
 439                 code_conv = conv_euctodisp;
 440         else
 441                 code_conv = conv_noconv;
 442
 443         return code_conv;
 444 }
 445
 446 static const struct {
 447         CharSet charset;
 448         gchar *const name;
 449 } charsets[] = {
 450         {C_US_ASCII,            CS_US_ASCII},
 451         {C_US_ASCII,            CS_ANSI_X3_4_1968},
 452         {C_UTF_8,               CS_UTF_8},
 453         {C_ISO_8859_1,          CS_ISO_8859_1},
 454         {C_ISO_8859_2,          CS_ISO_8859_2},
 455         {C_ISO_8859_4,          CS_ISO_8859_4},
 456         {C_ISO_8859_5,          CS_ISO_8859_5},
 457         {C_ISO_8859_7,          CS_ISO_8859_7},
 458         {C_ISO_8859_8,          CS_ISO_8859_8},
 459         {C_ISO_8859_9,          CS_ISO_8859_9},
 460         {C_ISO_8859_13,         CS_ISO_8859_13},
 461         {C_ISO_8859_15,         CS_ISO_8859_15},
 462         {C_BALTIC,              CS_BALTIC},
 463         {C_CP1251,              CS_CP1251},
 464         {C_WINDOWS_1251,        CS_WINDOWS_1251},
 465         {C_KOI8_R,              CS_KOI8_R},
 466         {C_KOI8_U,              CS_KOI8_U},
 467         {C_ISO_2022_JP,         CS_ISO_2022_JP},
 468         {C_ISO_2022_JP_2,       CS_ISO_2022_JP_2},
 469         {C_EUC_JP,              CS_EUC_JP},
 470         {C_SHIFT_JIS,           CS_SHIFT_JIS},
 471         {C_ISO_2022_KR,         CS_ISO_2022_KR},
 472         {C_EUC_KR,              CS_EUC_KR},
 473         {C_ISO_2022_CN,         CS_ISO_2022_CN},
 474         {C_EUC_CN,              CS_EUC_CN},
 475         {C_GB2312,              CS_GB2312},
 476         {C_EUC_TW,              CS_EUC_TW},
 477         {C_BIG5,                CS_BIG5},
 478 };
 479
 480 #if !HAVE_LIBJCONV
 481 static const struct {
 482         gchar *const locale;
 483         CharSet charset;
 484 } locale_table[] = {
 485         {"ja_JP.eucJP"  , C_EUC_JP},
 486         {"ja_JP.ujis"   , C_EUC_JP},
 487         {"ja_JP.EUC"    , C_EUC_JP},
 488         {"ja_JP.SJIS"   , C_SHIFT_JIS},
 489         {"ja_JP.JIS"    , C_ISO_2022_JP},
 490         {"ja_JP"        , C_EUC_JP},
 491         {"ko_KR"        , C_EUC_KR},
 492         {"zh_CN.GB2312" , C_GB2312},
 493         {"zh_CN"        , C_GB2312},
 494         {"zh_TW.eucTW"  , C_EUC_TW},
 495         {"zh_TW.Big5"   , C_BIG5},
 496         {"zh_TW"        , C_BIG5},
 497
 498         {"ru_RU.KOI8-R" , C_KOI8_R},
 499         {"ru_RU.CP1251" , C_CP1251},
 500
 501         {"en_US"        , C_ISO_8859_1},
 502         {"ca_ES"        , C_ISO_8859_1},
 503         {"da_DK"        , C_ISO_8859_1},
 504         {"de_DE"        , C_ISO_8859_1},
 505         {"nl_NL"        , C_ISO_8859_1},
 506         {"et_EE"        , C_ISO_8859_1},
 507         {"fi_FI"        , C_ISO_8859_1},
 508         {"fr_FR"        , C_ISO_8859_1},
 509         {"is_IS"        , C_ISO_8859_1},
 510         {"it_IT"        , C_ISO_8859_1},
 511         {"no_NO"        , C_ISO_8859_1},
 512         {"pt_PT"        , C_ISO_8859_1},
 513         {"pt_BR"        , C_ISO_8859_1},
 514         {"es_ES"        , C_ISO_8859_1},
 515         {"sv_SE"        , C_ISO_8859_1},
 516
 517         {"hr_HR"        , C_ISO_8859_2},
 518         {"hu_HU"        , C_ISO_8859_2},
 519         {"pl_PL"        , C_ISO_8859_2},
 520         {"ro_RO"        , C_ISO_8859_2},
 521         {"sk_SK"        , C_ISO_8859_2},
 522         {"sl_SI"        , C_ISO_8859_2},
 523         {"ru_RU"        , C_ISO_8859_5},
 524         {"el_GR"        , C_ISO_8859_7},
 525         {"iw_IL"        , C_ISO_8859_8},
 526         {"tr_TR"        , C_ISO_8859_9},
 527
 528         {"lt_LT.iso88594"       , C_ISO_8859_4},
 529         {"lt_LT.ISO8859-4"      , C_ISO_8859_4},
 530         {"lt_LT.ISO_8859-4"     , C_ISO_8859_4},
 531         {"lt_LT"                , C_ISO_8859_13},
 532         {"lv_LV"                , C_ISO_8859_13},
 533
 534         {"C"                    , C_US_ASCII},
 535         {"POSIX"                , C_US_ASCII},
 536         {"ANSI_X3.4-1968"       , C_US_ASCII},
 537 };
 538 #endif /* !HAVE_LIBJCONV */
 539
 540 const gchar *conv_get_charset_str(CharSet charset)
 541 {
 542         gint i;
 543
 544         for (i = 0; i < sizeof(charsets) / sizeof(charsets[0]); i++) {
 545                 if (charsets[i].charset == charset)
 546                         return charsets[i].name;
 547         }
 548
 549         return NULL;
 550 }
 551
 552 CharSet conv_get_charset_from_str(const gchar *charset)
 553 {
 554         gint i;
 555
 556         if (!charset) return C_AUTO;
 557
 558         for (i = 0; i < sizeof(charsets) / sizeof(charsets[0]); i++) {
 559                 if (!strcasecmp(charsets[i].name, charset))
 560                         return charsets[i].charset;
 561         }
 562
 563         return C_AUTO;
 564 }
 565
 566 CharSet conv_get_current_charset(void)
 567 {
 568         static CharSet cur_charset = -1;
 569         gint i;
 570
 571 #if HAVE_LIBJCONV
 572         const gchar *cur_codeset;
 573 #else
 574         const gchar *cur_locale;
 575 #endif
 576
 577         if (cur_charset != -1)
 578                 return cur_charset;
 579
 580 #if HAVE_LIBJCONV
 581         cur_codeset = jconv_info_get_current_codeset();
 582         for (i = 0; i < sizeof(charsets) / sizeof(charsets[0]); i++) {
 583                 if (!strcasecmp(cur_codeset, charsets[i].name)) {
 584                         cur_charset = charsets[i].charset;
 585                         return cur_charset;
 586                 }
 587         }
 588 #else
 589         cur_locale = g_getenv("LC_ALL");
 590         if (!cur_locale) cur_locale = g_getenv("LC_CTYPE");
 591         if (!cur_locale) cur_locale = g_getenv("LANG");
 592         if (!cur_locale) cur_locale = setlocale(LC_CTYPE, NULL);
 593
 594         debug_print("current locale: %s\n",
 595                     cur_locale ? cur_locale : "(none)");
 596
 597         if (!cur_locale) {
 598                 cur_charset = C_US_ASCII;
 599                 return cur_charset;
 600         }
 601
 602         if (strcasestr(cur_locale, "UTF-8")) {
 603                 cur_charset = C_UTF_8;
 604                 return cur_charset;
 605         }
 606
 607         for (i = 0; i < sizeof(locale_table) / sizeof(locale_table[0]); i++) {
 608                 const gchar *p;
 609
 610                 /* "ja_JP.EUC" matches with "ja_JP.eucJP" and "ja_JP.EUC" */
 611                 /* "ja_JP" matches with "ja_JP.xxxx" and "ja" */
 612                 if (!strncasecmp(cur_locale, locale_table[i].locale,
 613                                  strlen(locale_table[i].locale))) {
 614                         cur_charset = locale_table[i].charset;
 615                         return cur_charset;
 616                 } else if ((p = strchr(locale_table[i].locale, '_')) &&
 617                          !strchr(p + 1, '.')) {
 618                         if (strlen(cur_locale) == 2 &&
 619                             !strncasecmp(cur_locale, locale_table[i].locale, 2)) {
 620                                 cur_charset = locale_table[i].charset;
 621                                 return cur_charset;
 622                         }
 623                 }
 624         }
 625 #endif
 626
 627         cur_charset = C_AUTO;
 628         return cur_charset;
 629 }
 630
 631 const gchar *conv_get_current_charset_str(void)
 632 {
 633         static const gchar *codeset = NULL;
 634
 635         if (!codeset)
 636                 codeset = conv_get_charset_str(conv_get_current_charset());
 637
 638         return codeset ? codeset : "US-ASCII";
 639 }
 640
 641 CharSet conv_get_outgoing_charset(void)
 642 {
 643         static CharSet out_charset = -1;
 644
 645 #if HAVE_LIBJCONV
 646         gint i, j, n_pref_codesets;
 647         const gchar *const *pref_codesets;
 648 #else
 649         CharSet cur_charset;
 650 #endif
 651
 652         if (out_charset != -1)
 653                 return out_charset;
 654
 655 #if HAVE_LIBJCONV
 656         /* skip US-ASCII and UTF-8 */
 657         pref_codesets = jconv_info_get_pref_codesets(&n_pref_codesets);
 658         for (i = 0; i < n_pref_codesets; i++) {
 659                 for (j = 3; j < sizeof(charsets) / sizeof(charsets[0]); j++) {
 660                         if (!strcasecmp(pref_codesets[i], charsets[j].name)) {
 661                                 out_charset = charsets[j].charset;
 662                                 return out_charset;
 663                         }
 664                 }
 665         }
 666
 667         for (i = 0; i < n_pref_codesets; i++) {
 668                 if (!strcasecmp(pref_codesets[i], "UTF-8")) {
 669                         out_charset = C_UTF_8;
 670                         return out_charset;
 671                 }
 672         }
 673
 674         out_charset = C_AUTO;
 675 #else
 676         cur_charset = conv_get_current_charset();
 677         switch (cur_charset) {
 678         case C_EUC_JP:
 679         case C_SHIFT_JIS:
 680                 out_charset = C_ISO_2022_JP;
 681                 break;
 682         default:
 683                 out_charset = cur_charset;
 684         }
 685 #endif
 686
 687         return out_charset;
 688 }
 689
 690 const gchar *conv_get_outgoing_charset_str(void)
 691 {
 692         CharSet out_charset;
 693         const gchar *str;
 694
 695         if (prefs_common.outgoing_charset) {
 696                 if (!isalpha(prefs_common.outgoing_charset[0])) {
 697                         g_free(prefs_common.outgoing_charset);
 698                         prefs_common.outgoing_charset = g_strdup(CS_AUTO);
 699                 } else if (strcmp(prefs_common.outgoing_charset, CS_AUTO) != 0)
 700                         return prefs_common.outgoing_charset;
 701         }
 702
 703         out_charset = conv_get_outgoing_charset();
 704         str = conv_get_charset_str(out_charset);
 705
 706         return str ? str : "US-ASCII";
 707 }
 708
 709 const gchar *conv_get_current_locale(void)
 710 {
 711         gchar *cur_locale;
 712
 713         cur_locale = g_getenv("LC_ALL");
 714         if (!cur_locale) cur_locale = g_getenv("LANG");
 715         if (!cur_locale) cur_locale = setlocale(LC_CTYPE, NULL);
 716
 717         debug_print("current locale: %s\n",
 718                     cur_locale ? cur_locale : "(none)");
 719
 720         return cur_locale;
 721 }
 722
 723 void conv_unmime_header_overwrite(gchar *str)
 724 {
 725         gchar *buf;
 726         gint outlen;
 727         CharSet cur_charset;
 728
 729         cur_charset = conv_get_current_charset();
 730
 731 #if HAVE_LIBJCONV
 732         Xstrdup_a(buf, str, return);
 733         outlen = strlen(str) + 1;
 734         UnMimeHeaderConv(buf, str, outlen);
 735         if (cur_charset == C_EUC_JP) {
 736                 gchar *tmp;
 737                 gint len;
 738
 739                 len = strlen(str) * 2 + 1;
 740                 Xalloca(tmp, len, return);
 741                 conv_jistodisp(tmp, len, str);
 742                 strncpy2(str, tmp, outlen);
 743         }
 744 #else
 745         if (cur_charset == C_EUC_JP) {
 746                 gchar *tmp;
 747                 gint len;
 748
 749                 Xstrdup_a(buf, str, return);
 750                 outlen = strlen(str) + 1;
 751                 UnMimeHeader(buf);
 752                 len = strlen(buf) * 2 + 1;
 753                 Xalloca(tmp, len, {strncpy2(str, buf, outlen); return;});
 754                 conv_jistodisp(tmp, len, buf);
 755                 strncpy2(str, tmp, outlen);
 756         } else
 757                 UnMimeHeader(str);
 758 #endif
 759 }
 760
 761 void conv_unmime_header(gchar *outbuf, gint outlen, const gchar *str,
 762                         const gchar *charset)
 763 {
 764         gchar *buf;
 765         CharSet cur_charset;
 766
 767         cur_charset = conv_get_current_charset();
 768         Xstrdup_a(buf, str, return);
 769
 770 #if HAVE_LIBJCONV
 771         UnMimeHeaderConv(buf, outbuf, outlen);
 772 #else
 773         UnMimeHeader(buf);
 774         strncpy2(outbuf, buf, outlen);
 775 #endif
 776         if (cur_charset == C_EUC_JP) {
 777                 gint len;
 778
 779                 len = strlen(outbuf) * 2 + 1;
 780                 Xalloca(buf, len, return);
 781                 conv_jistodisp(buf, len, outbuf);
 782                 strncpy2(outbuf, buf, outlen);
 783         }
 784 }
 785
 786 #define MAX_ENCLEN      75
 787 #define MAX_LINELEN     76
 788
 789 #define B64LEN(len)     ((len) / 3 * 4 + ((len) % 3 ? 4 : 0))
 790
 791 #if HAVE_LIBJCONV
 792 void conv_encode_header(gchar *dest, gint len, const gchar *src,
 793                         gint header_len)
 794 {
 795         wchar_t *wsrc;
 796         wchar_t *wsrcp;
 797         gchar *destp;
 798         size_t line_len, mimehdr_len, mimehdr_begin_len;
 799         gchar *mimehdr_init = "=?";
 800         gchar *mimehdr_end = "?=";
 801         gchar *mimehdr_enctype = "?B?";
 802         const gchar *mimehdr_charset;
 803
 804         /* g_print("src = %s\n", src); */
 805         mimehdr_charset = conv_get_outgoing_charset_str();
 806
 807         /* convert to wide-character string */
 808         wsrcp = wsrc = strdup_mbstowcs(src);
 809
 810         mimehdr_len = strlen(mimehdr_init) + strlen(mimehdr_end) +
 811                 strlen(mimehdr_charset) + strlen(mimehdr_enctype);
 812         mimehdr_begin_len = strlen(mimehdr_init) +
 813                 strlen(mimehdr_charset) + strlen(mimehdr_enctype);
 814         /* line_len = 1; */
 815         line_len = header_len;
 816         destp = dest;
 817         *dest = '\0';
 818
 819         g_return_if_fail(wsrc != NULL);
 820
 821         while (*wsrcp) {
 822                 wchar_t *wp, *wtmp, *wtmpp;
 823                 gint nspc = 0;
 824
 825                 /* irresponsible buffer overrun check */
 826                 if ((len - (destp - dest)) < (MAX_LINELEN + 1) * 2) break;
 827
 828                 /* encode string including space
 829                    if non-ASCII string follows */
 830                 if (is_next_nonascii(wsrcp)) {
 831                         wp = wsrcp;
 832                         while ((wp = find_wspace(wp)) != NULL)
 833                                 if (!is_next_nonascii(wp)) break;
 834                 } else
 835                         wp = find_wspace(wsrcp);
 836
 837                 if (wp != NULL) {
 838                         wtmp = wcsndup(wsrcp, wp - wsrcp);
 839                         wsrcp = wp + 1;
 840                         while (iswspace(wsrcp[nspc])) nspc++;
 841                 } else {
 842                         wtmp = wcsdup(wsrcp);
 843                         wsrcp += wcslen(wsrcp);
 844                 }
 845
 846                 wtmpp = wtmp;
 847
 848                 do {
 849                         gint tlen = 0, str_ascii = 1;
 850                         gchar *tmp; /* internal codeset */
 851                         gchar *raw; /* converted, but not base64 encoded */
 852                         register gchar *tmpp;
 853                         gint raw_len;
 854
 855                         tmpp = tmp = g_malloc(wcslen(wtmpp) * MB_CUR_MAX + 1);
 856                         *tmp = '\0';
 857                         raw = g_strdup("");
 858                         raw_len = 0;
 859
 860                         while (*wtmpp != (wchar_t)0) {
 861                                 gint mbl;
 862                                 gint dummy;
 863                                 gchar *raw_new = NULL;
 864                                 int raw_new_len = 0;
 865                                 const gchar *src_codeset;
 866
 867                                 if (*wtmpp < 32 || *wtmpp >= 127)
 868                                         str_ascii = 0;
 869                                 mbl = wctomb(tmpp, *wtmpp);
 870                                 if (mbl == -1) {
 871                                         g_warning("invalid wide character\n");
 872                                         wtmpp++;
 873                                         continue;
 874                                 }
 875                                 /* g_free(raw); */
 876                                 src_codeset = conv_get_current_charset_str();
 877                                 /* printf ("tmp = %s, tlen = %d, mbl\n",
 878                                         tmp, tlen, mbl); */
 879                                 if (jconv_alloc_conv(tmp, tlen + mbl,
 880                                                      &raw_new, &raw_new_len,
 881                                                      &src_codeset, 1,
 882                                                      &dummy, mimehdr_charset)
 883                                     != 0) {
 884                                         g_warning("can't convert\n");
 885                                         tmpp[0] = '\0';
 886                                         continue;
 887                                 }
 888                                 if (!str_ascii) {
 889                                         gint dlen = mimehdr_len +
 890                                                 B64LEN(raw_len);
 891                                         if ((line_len + dlen +
 892                                              (*(wtmpp + 1) ? 0 : nspc) +
 893                                              (line_len > 1 ? 1 : 0))
 894                                             > MAX_LINELEN) {
 895                                                 g_free(raw_new);
 896                                                 if (tlen == 0) {
 897                                                         *destp++ = '\n';
 898                                                         *destp++ = ' ';
 899                                                         line_len = 1;
 900                                                         str_ascii = 1;
 901                                                         continue;
 902                                                 } else {
 903                                                         *tmpp = '\0';
 904                                                         break;
 905                                                 }
 906                                         }
 907                                 } else if ((line_len + tlen + mbl +
 908                                             (*(wtmpp + 1) ? 0 : nspc) +
 909                                             (line_len > 1 ? 1 : 0))
 910                                            > MAX_LINELEN) {
 911                                         g_free(raw_new);
 912                                         if (1 + tlen + mbl +
 913                                             (*(wtmpp + 1) ? 0 : nspc)
 914                                             >= MAX_LINELEN) {
 915                                                 *tmpp = '\0';
 916                                                 break;
 917                                         }
 918                                         *destp++ = '\n';
 919                                         *destp++ = ' ';
 920                                         line_len = 1;
 921                                         continue;
 922                                 }
 923
 924                                 tmpp += mbl;
 925                                 *tmpp = '\0';
 926
 927                                 tlen += mbl;
 928
 929                                 g_free(raw);
 930                                 raw = raw_new;
 931                                 raw_len = raw_new_len;
 932
 933                                 wtmpp++;
 934                         }
 935                         /*
 936                           g_print("tmp = %s, tlen = %d, mb_seqlen = %d\n",
 937                                 tmp, tlen, mb_seqlen);
 938                         */
 939
 940                         if (tlen == 0 || raw_len == 0) {
 941                                 g_free(tmp);
 942                                 g_free(raw);
 943                                 continue;
 944                         }
 945
 946                         if (line_len > 1 && destp > dest) {
 947                                 *destp++ = ' ';
 948                                 *destp = '\0';
 949                                 line_len++;
 950                         }
 951
 952                         if (!str_ascii) {
 953                                 g_snprintf(destp, len - strlen(dest), "%s%s%s",
 954                                            mimehdr_init, mimehdr_charset,
 955                                            mimehdr_enctype);
 956                                 destp += mimehdr_begin_len;
 957                                 line_len += mimehdr_begin_len;
 958
 959                                 to64frombits(destp, raw, raw_len);
 960                                 line_len += strlen(destp);
 961                                 destp += strlen(destp);
 962
 963                                 strcpy(destp, mimehdr_end);
 964                                 destp += strlen(mimehdr_end);
 965                                 line_len += strlen(mimehdr_end);
 966                         } else {
 967                                 strcpy(destp, tmp);
 968                                 line_len += strlen(destp);
 969                                 destp += strlen(destp);
 970                         }
 971
 972                         g_free(tmp);
 973                         g_free(raw);
 974                         /* g_print("line_len = %d\n\n", line_len); */
 975                 } while (*wtmpp != (wchar_t)0);
 976
 977                 while (iswspace(*wsrcp)) {
 978                         gint mbl;
 979
 980                         mbl = wctomb(destp, *wsrcp++);
 981                         if (mbl != -1) {
 982                                 destp += mbl;
 983                                 line_len += mbl;
 984                         }
 985                 }
 986                 *destp = '\0';
 987
 988                 g_free(wtmp);
 989         }
 990
 991         g_free(wsrc);
 992
 993         /* g_print("dest = %s\n", dest);  */
 994 }
 995 #else /* !HAVE_LIBJCONV */
 996
 997 #define JIS_SEQLEN      3
 998
 999 void conv_encode_header(gchar *dest, gint len, const gchar *src,
1000                         gint header_len)
1001 {
1002         wchar_t *wsrc;
1003         wchar_t *wsrcp;
1004         gchar *destp;
1005         size_t line_len, mimehdr_len, mimehdr_begin_len;
1006         gchar *mimehdr_init = "=?";
1007         gchar *mimehdr_end = "?=";
1008         gchar *mimehdr_enctype = "?B?";
1009         const gchar *mimehdr_charset;
1010
1011         /* g_print("src = %s\n", src); */
1012         mimehdr_charset = conv_get_outgoing_charset_str();
1013         if (strcmp(mimehdr_charset, "ISO-2022-JP") != 0) {
1014                 /* currently only supports Japanese */
1015                 strncpy2(dest, src, len);
1016                 return;
1017         }
1018
1019         /* convert to wide-character string */
1020         wsrcp = wsrc = strdup_mbstowcs(src);
1021
1022         mimehdr_len = strlen(mimehdr_init) + strlen(mimehdr_end) +
1023                       strlen(mimehdr_charset) + strlen(mimehdr_enctype);
1024         mimehdr_begin_len = strlen(mimehdr_init) +
1025                             strlen(mimehdr_charset) + strlen(mimehdr_enctype);
1026         /* line_len = 1; */
1027         line_len = header_len;
1028         destp = dest;
1029         *dest = '\0';
1030
1031         g_return_if_fail(wsrc != NULL);
1032
1033         while (*wsrcp) {
1034                 wchar_t *wp, *wtmp, *wtmpp;
1035                 gint nspc = 0;
1036
1037                 /* irresponsible buffer overrun check */
1038                 if ((len - (destp - dest)) < (MAX_LINELEN + 1) * 2) break;
1039
1040                 /* encode string including space
1041                    if non-ASCII string follows */
1042                 if (is_next_nonascii(wsrcp)) {
1043                         wp = wsrcp;
1044                         while ((wp = find_wspace(wp)) != NULL)
1045                                 if (!is_next_nonascii(wp)) break;
1046                 } else
1047                         wp = find_wspace(wsrcp);
1048
1049                 if (wp != NULL) {
1050                         wtmp = wcsndup(wsrcp, wp - wsrcp);
1051                         wsrcp = wp + 1;
1052                         while (iswspace(wsrcp[nspc])) nspc++;
1053                 } else {
1054                         wtmp = wcsdup(wsrcp);
1055                         wsrcp += wcslen(wsrcp);
1056                 }
1057
1058                 wtmpp = wtmp;
1059
1060                 do {
1061                         gint prev_mbl = 1, tlen = 0, mb_seqlen = 0;
1062                         gchar *tmp;
1063                         register gchar *tmpp;
1064
1065                         tmpp = tmp = g_malloc(wcslen(wtmpp) * MB_CUR_MAX + 1);
1066                         *tmp = '\0';
1067
1068                         while (*wtmpp != (wchar_t)0) {
1069                                 gint mbl;
1070
1071                                 mbl = wctomb(tmpp, *wtmpp);
1072                                 if (mbl == -1) {
1073                                         g_warning("invalid wide character\n");
1074                                         wtmpp++;
1075                                         continue;
1076                                 }
1077
1078                                 /* length of KI + KO */
1079                                 if (prev_mbl == 1 && mbl == 2)
1080                                         mb_seqlen += JIS_SEQLEN * 2;
1081
1082                                 if (mb_seqlen) {
1083                                         gint dlen = mimehdr_len +
1084                                                 B64LEN(tlen + mb_seqlen + mbl);
1085
1086                                         if ((line_len + dlen +
1087                                              (*(wtmpp + 1) ? 0 : nspc) +
1088                                              (line_len > 1 ? 1 : 0))
1089                                             > MAX_LINELEN) {
1090                                                 if (tlen == 0) {
1091                                                         *destp++ = '\n';
1092                                                         *destp++ = ' ';
1093                                                         line_len = 1;
1094                                                         mb_seqlen = 0;
1095                                                         continue;
1096                                                 } else {
1097                                                         *tmpp = '\0';
1098                                                         break;
1099                                                 }
1100                                         }
1101                                 } else if ((line_len + tlen + mbl +
1102                                             (*(wtmpp + 1) ? 0 : nspc) +
1103                                             (line_len > 1 ? 1 : 0))
1104                                            > MAX_LINELEN) {
1105                                         if (1 + tlen + mbl +
1106                                             (*(wtmpp + 1) ? 0 : nspc)
1107                                             >= MAX_LINELEN) {
1108                                                 *tmpp = '\0';
1109                                                 break;
1110                                         }
1111                                         *destp++ = '\n';
1112                                         *destp++ = ' ';
1113                                         line_len = 1;
1114                                         continue;
1115                                 }
1116
1117                                 tmpp += mbl;
1118                                 *tmpp = '\0';
1119
1120                                 tlen += mbl;
1121                                 prev_mbl = mbl;
1122
1123                                 wtmpp++;
1124                         }
1125                         /* g_print("tmp = %s, tlen = %d, mb_seqlen = %d\n",
1126                                 tmp, tlen, mb_seqlen);
1127                         */
1128
1129                         if (tlen == 0) {
1130                                 g_free(tmp);
1131                                 continue;
1132                         }
1133
1134                         if (line_len > 1 && destp > dest) {
1135                                 *destp++ = ' ';
1136                                 *destp = '\0';
1137                                 line_len++;
1138                         }
1139
1140                         if (mb_seqlen) {
1141                                 gchar *tmp_jis;
1142
1143                                 tmp_jis = g_new(gchar, tlen + mb_seqlen + 1);
1144                                 conv_euctojis(tmp_jis,
1145                                               tlen + mb_seqlen + 1, tmp);
1146                                 g_snprintf(destp, len - strlen(dest), "%s%s%s",
1147                                            mimehdr_init, mimehdr_charset,
1148                                            mimehdr_enctype);
1149                                 destp += mimehdr_begin_len;
1150                                 line_len += mimehdr_begin_len;
1151
1152                                 to64frombits(destp, tmp_jis, strlen(tmp_jis));
1153                                 line_len += strlen(destp);
1154                                 destp += strlen(destp);
1155
1156                                 strcpy(destp, mimehdr_end);
1157                                 destp += strlen(mimehdr_end);
1158                                 line_len += strlen(mimehdr_end);
1159
1160                                 g_free(tmp_jis);
1161                         } else {
1162                                 strcpy(destp, tmp);
1163                                 line_len += strlen(destp);
1164                                 destp += strlen(destp);
1165                         }
1166
1167                         g_free(tmp);
1168                         /* g_print("line_len = %d\n\n", line_len); */
1169                 } while (*wtmpp != (wchar_t)0);
1170
1171                 while (iswspace(*wsrcp)) {
1172                         gint mbl;
1173
1174                         mbl = wctomb(destp, *wsrcp++);
1175                         if (mbl != -1) {
1176                                 destp += mbl;
1177                                 line_len += mbl;
1178                         }
1179                 }
1180                 *destp = '\0';
1181
1182                 g_free(wtmp);
1183         }
1184
1185         g_free(wsrc);
1186
1187         /* g_print("dest = %s\n", dest); */
1188 }
1189 #endif /* HAVE_LIBJCONV */