lib/euc_jp.h

   1 /*
   2  * Copyright (C) 1999-2001, 2005, 2016 Free Software Foundation, Inc.
   3  * This file is part of the GNU LIBICONV Library.
   4  *
   5  * The GNU LIBICONV Library is free software; you can redistribute it
   6  * and/or modify it under the terms of the GNU Lesser General Public
   7  * License as published by the Free Software Foundation; either version 2.1
   8  * of the License, or (at your option) any later version.
   9  *
  10  * The GNU LIBICONV Library is distributed in the hope that it will be
  11  * useful, but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  13  * Lesser General Public License for more details.
  14  *
  15  * You should have received a copy of the GNU Lesser General Public
  16  * License along with the GNU LIBICONV Library; see the file COPYING.LIB.
  17  * If not, see <https://www.gnu.org/licenses/>.
  18  */
  19
  20 /*
  21  * EUC-JP
  22  */
  23
  24 static int
  25 euc_jp_mbtowc (conv_t conv, ucs4_t *pwc, const unsigned char *s, size_t n)
  26 {
  27   unsigned char c = *s;
  28   /* Code set 0 (ASCII or JIS X 0201-1976 Roman) */
  29   if (c < 0x80)
  30     return ascii_mbtowc(conv,pwc,s,n);
  31   /* Code set 1 (JIS X 0208) */
  32   if (c >= 0xa1 && c < 0xff) {
  33     if (n < 2)
  34       return RET_TOOFEW(0);
  35     if (c < 0xf5) {
  36       unsigned char c2 = s[1];
  37       if (c2 >= 0xa1 && c2 < 0xff) {
  38         unsigned char buf[2];
  39         buf[0] = c-0x80; buf[1] = c2-0x80;
  40         return jisx0208_mbtowc(conv,pwc,buf,2);
  41       } else
  42         return RET_ILSEQ;
  43     } else {
  44       /* User-defined range. See
  45        * Ken Lunde's "CJKV Information Processing", table 4-66, p. 206. */
  46       unsigned char c2 = s[1];
  47       if (c2 >= 0xa1 && c2 < 0xff) {
  48         *pwc = 0xe000 + 94*(c-0xf5) + (c2-0xa1);
  49         return 2;
  50       } else
  51         return RET_ILSEQ;
  52     }
  53   }
  54   /* Code set 2 (half-width katakana) */
  55   if (c == 0x8e) {
  56     if (n < 2)
  57       return RET_TOOFEW(0);
  58     {
  59       unsigned char c2 = s[1];
  60       if (c2 >= 0xa1 && c2 < 0xe0) {
  61         int ret = jisx0201_mbtowc(conv,pwc,s+1,n-1);
  62         if (ret == RET_ILSEQ)
  63           return RET_ILSEQ;
  64         if (ret != 1) abort();
  65         return 2;
  66       } else
  67         return RET_ILSEQ;
  68     }
  69   }
  70   /* Code set 3 (JIS X 0212-1990) */
  71   if (c == 0x8f) {
  72     if (n < 2)
  73       return RET_TOOFEW(0);
  74     {
  75       unsigned char c2 = s[1];
  76       if (c2 >= 0xa1 && c2 < 0xff) {
  77         if (n < 3)
  78           return RET_TOOFEW(0);
  79         if (c2 < 0xf5) {
  80           unsigned char c3 = s[2];
  81           if (c3 >= 0xa1 && c3 < 0xff) {
  82             unsigned char buf[2];
  83             int ret;
  84             buf[0] = c2-0x80; buf[1] = c3-0x80;
  85             ret = jisx0212_mbtowc(conv,pwc,buf,2);
  86             if (ret == RET_ILSEQ)
  87               return RET_ILSEQ;
  88             if (ret != 2) abort();
  89             return 3;
  90           } else
  91             return RET_ILSEQ;
  92         } else {
  93           /* User-defined range. See
  94            * Ken Lunde's "CJKV Information Processing", table 4-66, p. 206. */
  95           unsigned char c3 = s[2];
  96           if (c3 >= 0xa1 && c3 < 0xff) {
  97             *pwc = 0xe3ac + 94*(c2-0xf5) + (c3-0xa1);
  98             return 3;
  99           } else
 100             return RET_ILSEQ;
 101         }
 102       } else
 103         return RET_ILSEQ;
 104     }
 105   }
 106   return RET_ILSEQ;
 107 }
 108
 109 static int
 110 euc_jp_wctomb (conv_t conv, unsigned char *r, ucs4_t wc, size_t n)
 111 {
 112   unsigned char buf[2];
 113   int ret;
 114
 115   /* Code set 0 (ASCII or JIS X 0201-1976 Roman) */
 116   ret = ascii_wctomb(conv,r,wc,n);
 117   if (ret != RET_ILUNI)
 118     return ret;
 119
 120   /* Code set 1 (JIS X 0208) */
 121   ret = jisx0208_wctomb(conv,buf,wc,2);
 122   if (ret != RET_ILUNI) {
 123     if (ret != 2) abort();
 124     if (n < 2)
 125       return RET_TOOSMALL;
 126     r[0] = buf[0]+0x80;
 127     r[1] = buf[1]+0x80;
 128     return 2;
 129   }
 130
 131   /* Code set 2 (half-width katakana) */
 132   ret = jisx0201_wctomb(conv,buf,wc,1);
 133   if (ret != RET_ILUNI && buf[0] >= 0x80) {
 134     if (ret != 1) abort();
 135     if (n < 2)
 136       return RET_TOOSMALL;
 137     r[0] = 0x8e;
 138     r[1] = buf[0];
 139     return 2;
 140   }
 141
 142   /* Code set 3 (JIS X 0212-1990) */
 143   ret = jisx0212_wctomb(conv,buf,wc,2);
 144   if (ret != RET_ILUNI) {
 145     if (ret != 2) abort();
 146     if (n < 3)
 147       return RET_TOOSMALL;
 148     r[0] = 0x8f;
 149     r[1] = buf[0]+0x80;
 150     r[2] = buf[1]+0x80;
 151     return 3;
 152   }
 153
 154   /* Extra compatibility with Shift_JIS.  */
 155   if (wc == 0x00a5) {
 156     r[0] = 0x5c;
 157     return 1;
 158   }
 159   if (wc == 0x203e) {
 160     r[0] = 0x7e;
 161     return 1;
 162   }
 163
 164   /* User-defined range. See
 165    * Ken Lunde's "CJKV Information Processing", table 4-66, p. 206. */
 166   if (wc >= 0xe000 && wc < 0xe758) {
 167     if (wc < 0xe3ac) {
 168       unsigned char c1, c2;
 169       if (n < 2)
 170         return RET_TOOSMALL;
 171       c1 = (unsigned int) (wc - 0xe000) / 94;
 172       c2 = (unsigned int) (wc - 0xe000) % 94;
 173       r[0] = c1+0xf5;
 174       r[1] = c2+0xa1;
 175       return 2;
 176     } else {
 177       unsigned char c1, c2;
 178       if (n < 3)
 179         return RET_TOOSMALL;
 180       c1 = (unsigned int) (wc - 0xe3ac) / 94;
 181       c2 = (unsigned int) (wc - 0xe3ac) % 94;
 182       r[0] = 0x8f;
 183       r[1] = c1+0xf5;
 184       r[2] = c2+0xa1;
 185       return 3;
 186     }
 187   }
 188
 189   return RET_ILUNI;
 190 }