lib/johab.h

   1 /*
   2  * Copyright (C) 1999-2001, 2007, 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  * JOHAB
  22  */
  23
  24 /*
  25    Conversion between JOHAB codes (s1,s2) and KSX1001 codes (c1,c2):
  26    Example. (s1,s2) = 0xD931, (c1,c2) = 0x2121.
  27             (s1,s2) = 0xDEF1, (c1,c2) = 0x2C71.
  28             (s1,s2) = 0xE031, (c1,c2) = 0x4A21.
  29             (s1,s2) = 0xF9FE, (c1,c2) = 0x7D7E.
  30    0xD9 <= s1 <= 0xDE || 0xE0 <= s1 <= 0xF9,
  31    0x31 <= s2 <= 0x7E || 0x91 <= s2 <= 0xFE,
  32    0x21 <= c1 <= 0x2C || 0x4A <= c1 <= 0x7D,
  33    0x21 <= c2 <= 0x7E.
  34    Invariant:
  35      94*(s1 < 0xE0 ? 2*s1-0x1B2 : 2*s1-0x197) + (s2 < 0x91 ? s2-0x31 : s2-0x43)
  36      = 94*(c1-0x21)+(c2-0x21)
  37    Conversion (s1,s2) -> (c1,c2):
  38      t1 := (s1 < 0xE0 ? 2*s1-0x1B2 : 2*s1-0x197)
  39      t2 := (s2 < 0x91 ? s2-0x31 : s2-0x43)
  40      c1 := t1 + (t2 < 0x5E ? 0 : 1) + 0x21
  41      c2 := (t2 < 0x5E ? t2 : t2-0x5E) + 0x21
  42    Conversion (c1,c2) -> (s1,s2):
  43      t := (c1 < 0x4A ? (c1-0x21+0x1B2) : (c1-0x21+0x197))
  44      s1 := t >> 1
  45      t2 := (t & 1) * 0x5E + (c2 - 0x21)
  46      s2 := (t2 < 0x4E ? t2+0x31 : t2+0x43)
  47  */
  48
  49 static int
  50 johab_mbtowc (conv_t conv, ucs4_t *pwc, const unsigned char *s, size_t n)
  51 {
  52   unsigned char c = *s;
  53   if (c < 0x80) {
  54     if (c == 0x5c)
  55       *pwc = (ucs4_t) 0x20a9;
  56     else
  57       *pwc = (ucs4_t) c;
  58     return 1;
  59   } else if (c < 0xd8) {
  60     return johab_hangul_mbtowc(conv,pwc,s,n);
  61   } else {
  62     unsigned char s1, s2;
  63     s1 = c;
  64     if ((s1 >= 0xd9 && s1 <= 0xde) || (s1 >= 0xe0 && s1 <= 0xf9)) {
  65       if (n < 2)
  66         return RET_TOOFEW(0);
  67       s2 = s[1];
  68       if ((s2 >= 0x31 && s2 <= 0x7e) || (s2 >= 0x91 && s2 <= 0xfe)) {
  69         /* In KSC 5601, now KS X 1001, the region s1 = 0xDA, 0xA1 <= s2 <= 0xD3
  70            contains the 51 Jamo (Hangul letters). But in the Johab encoding,
  71            they have been moved to the Hangul section, see
  72            johab_hangul_page31. */
  73         if (!(s1 == 0xda && (s2 >= 0xa1 && s2 <= 0xd3))) {
  74           unsigned char t1 = (s1 < 0xe0 ? 2*(s1-0xd9) : 2*s1-0x197);
  75           unsigned char t2 = (s2 < 0x91 ? s2-0x31 : s2-0x43);
  76           unsigned char buf[2];
  77           buf[0] = t1 + (t2 < 0x5e ? 0 : 1) + 0x21;
  78           buf[1] = (t2 < 0x5e ? t2 : t2-0x5e) + 0x21;
  79           return ksc5601_mbtowc(conv,pwc,buf,2);
  80         }
  81       }
  82     }
  83     return RET_ILSEQ;
  84   }
  85 }
  86
  87 static int
  88 johab_wctomb (conv_t conv, unsigned char *r, ucs4_t wc, size_t n)
  89 {
  90   unsigned char buf[2];
  91   int ret;
  92
  93   /* Try ASCII variation. */
  94   if (wc < 0x0080 && wc != 0x005c) {
  95     *r = wc;
  96     return 1;
  97   }
  98   if (wc == 0x20a9) {
  99     *r = 0x5c;
 100     return 1;
 101   }
 102
 103   /* Try JOHAB Hangul table before KSC5601 table, because the KSC5601 table
 104      contains some (2350 out of 11172) Hangul syllables (rows 0x30XX..0x48XX),
 105      and we want the search to return the JOHAB Hangul table entry. */
 106
 107   /* Try JOHAB Hangul. */
 108   ret = johab_hangul_wctomb(conv,buf,wc,2);
 109   if (ret != RET_ILUNI) {
 110     if (ret != 2) abort();
 111     if (n < 2)
 112       return RET_TOOSMALL;
 113     r[0] = buf[0];
 114     r[1] = buf[1];
 115     return 2;
 116   }
 117
 118   /* Try KSC5601, now KS X 1001. */
 119   ret = ksc5601_wctomb(conv,buf,wc,2);
 120   if (ret != RET_ILUNI) {
 121     unsigned char c1, c2;
 122     if (ret != 2) abort();
 123     if (n < 2)
 124       return RET_TOOSMALL;
 125     c1 = buf[0];
 126     c2 = buf[1];
 127     if (((c1 >= 0x21 && c1 <= 0x2c) || (c1 >= 0x4a && c1 <= 0x7d))
 128         && (c2 >= 0x21 && c2 <= 0x7e)) {
 129       unsigned int t = (c1 < 0x4A ? (c1-0x21+0x1B2) : (c1-0x21+0x197));
 130       unsigned char t2 = ((t & 1) ? 0x5e : 0) + (c2 - 0x21);
 131       r[0] = t >> 1;
 132       r[1] = (t2 < 0x4e ? t2+0x31 : t2+0x43);
 133       return 2;
 134     }
 135   }
 136
 137   return RET_ILUNI;
 138 }