lib/big5hkscs2004.h

   1 /*
   2  * Copyright (C) 1999-2002, 2006, 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  * BIG5-HKSCS:2004
  22  */
  23
  24 /*
  25  * BIG5-HKSCS:2004 can be downloaded from
  26  *   http://www.info.gov.hk/digital21/eng/hkscs/download.html
  27  *   http://www.info.gov.hk/digital21/eng/hkscs/index.html
  28  *
  29  * It extends BIG5-HKSCS:2001 through 123 characters.
  30  *
  31  * It extends BIG5 (without the rows 0xC6..0xC7) through the ranges
  32  *
  33  *   0x{87..8D}{40..7E,A1..FE}      880 characters
  34  *   0x{8E..A0}{40..7E,A1..FE}     2898 characters
  35  *   0x{C6..C8}{40..7E,A1..FE}      359 characters
  36  *   0xF9{D6..FE}                    41 characters
  37  *   0x{FA..FE}{40..7E,A1..FE}      763 characters
  38  *
  39  * Note that some HKSCS characters are not contained in Unicode 3.2
  40  * and are therefore best represented as sequences of Unicode characters:
  41  *   0x8862  U+00CA U+0304  LATIN CAPITAL LETTER E WITH CIRCUMFLEX AND MACRON
  42  *   0x8864  U+00CA U+030C  LATIN CAPITAL LETTER E WITH CIRCUMFLEX AND CARON
  43  *   0x88A3  U+00EA U+0304  LATIN SMALL LETTER E WITH CIRCUMFLEX AND MACRON
  44  *   0x88A5  U+00EA U+030C  LATIN SMALL LETTER E WITH CIRCUMFLEX AND CARON
  45  */
  46
  47 #include "hkscs2004.h"
  48 #include "flushwc.h"
  49
  50 static int
  51 big5hkscs2004_mbtowc (conv_t conv, ucs4_t *pwc, const unsigned char *s, size_t n)
  52 {
  53   ucs4_t last_wc = conv->istate;
  54   if (last_wc) {
  55     /* Output the buffered character. */
  56     conv->istate = 0;
  57     *pwc = last_wc;
  58     return 0; /* Don't advance the input pointer. */
  59   } else {
  60     unsigned char c = *s;
  61     /* Code set 0 (ASCII) */
  62     if (c < 0x80)
  63       return ascii_mbtowc(conv,pwc,s,n);
  64     /* Code set 1 (BIG5 extended) */
  65     if (c >= 0xa1 && c < 0xff) {
  66       if (n < 2)
  67         return RET_TOOFEW(0);
  68       {
  69         unsigned char c2 = s[1];
  70         if ((c2 >= 0x40 && c2 < 0x7f) || (c2 >= 0xa1 && c2 < 0xff)) {
  71           if (!((c == 0xc6 && c2 >= 0xa1) || c == 0xc7)) {
  72             int ret = big5_mbtowc(conv,pwc,s,2);
  73             if (ret != RET_ILSEQ)
  74               return ret;
  75           }
  76         }
  77       }
  78     }
  79     {
  80       int ret = hkscs1999_mbtowc(conv,pwc,s,n);
  81       if (ret != RET_ILSEQ)
  82         return ret;
  83     }
  84     {
  85       int ret = hkscs2001_mbtowc(conv,pwc,s,n);
  86       if (ret != RET_ILSEQ)
  87         return ret;
  88     }
  89     {
  90       int ret = hkscs2004_mbtowc(conv,pwc,s,n);
  91       if (ret != RET_ILSEQ)
  92         return ret;
  93     }
  94     if (c == 0x88) {
  95       if (n < 2)
  96         return RET_TOOFEW(0);
  97       {
  98         unsigned char c2 = s[1];
  99         if (c2 == 0x62 || c2 == 0x64 || c2 == 0xa3 || c2 == 0xa5) {
 100           /* It's a composed character. */
 101           ucs4_t wc1 = ((c2 >> 3) << 2) + 0x009a; /* = 0x00ca or 0x00ea */
 102           ucs4_t wc2 = ((c2 & 6) << 2) + 0x02fc; /* = 0x0304 or 0x030c */
 103           /* We cannot output two Unicode characters at once. So,
 104              output the first character and buffer the second one. */
 105           *pwc = wc1;
 106           conv->istate = wc2;
 107           return 2;
 108         }
 109       }
 110     }
 111     return RET_ILSEQ;
 112   }
 113 }
 114
 115 #define big5hkscs2004_flushwc normal_flushwc
 116
 117 static int
 118 big5hkscs2004_wctomb (conv_t conv, unsigned char *r, ucs4_t wc, size_t n)
 119 {
 120   int count = 0;
 121   unsigned char last = conv->ostate;
 122
 123   if (last) {
 124     /* last is = 0x66 or = 0xa7. */
 125     if (wc == 0x0304 || wc == 0x030c) {
 126       /* Output the combined character. */
 127       if (n >= 2) {
 128         r[0] = 0x88;
 129         r[1] = last + ((wc & 24) >> 2) - 4; /* = 0x62 or 0x64 or 0xa3 or 0xa5 */
 130         conv->ostate = 0;
 131         return 2;
 132       } else
 133         return RET_TOOSMALL;
 134     }
 135
 136     /* Output the buffered character. */
 137     if (n < 2)
 138       return RET_TOOSMALL;
 139     r[0] = 0x88;
 140     r[1] = last;
 141     r += 2;
 142     count = 2;
 143   }
 144
 145   /* Code set 0 (ASCII) */
 146   if (wc < 0x0080) {
 147     /* Plain ASCII character. */
 148     if (n > count) {
 149       r[0] = (unsigned char) wc;
 150       conv->ostate = 0;
 151       return count+1;
 152     } else
 153       return RET_TOOSMALL;
 154   } else {
 155     unsigned char buf[2];
 156     int ret;
 157
 158     /* Code set 1 (BIG5 extended) */
 159     ret = big5_wctomb(conv,buf,wc,2);
 160     if (ret != RET_ILUNI) {
 161       if (ret != 2) abort();
 162       if (!((buf[0] == 0xc6 && buf[1] >= 0xa1) || buf[0] == 0xc7)) {
 163         if (n >= count+2) {
 164           r[0] = buf[0];
 165           r[1] = buf[1];
 166           conv->ostate = 0;
 167           return count+2;
 168         } else
 169           return RET_TOOSMALL;
 170       }
 171     }
 172     ret = hkscs1999_wctomb(conv,buf,wc,2);
 173     if (ret != RET_ILUNI) {
 174       if (ret != 2) abort();
 175       if ((wc & ~0x0020) == 0x00ca) {
 176         /* A possible first character of a multi-character sequence. We have to
 177            buffer it. */
 178         if (!(buf[0] == 0x88 && (buf[1] == 0x66 || buf[1] == 0xa7))) abort();
 179         conv->ostate = buf[1]; /* = 0x66 or = 0xa7 */
 180         return count+0;
 181       }
 182       if (n >= count+2) {
 183         r[0] = buf[0];
 184         r[1] = buf[1];
 185         conv->ostate = 0;
 186         return count+2;
 187       } else
 188         return RET_TOOSMALL;
 189     }
 190     ret = hkscs2001_wctomb(conv,buf,wc,2);
 191     if (ret != RET_ILUNI) {
 192       if (ret != 2) abort();
 193       if (n >= count+2) {
 194         r[0] = buf[0];
 195         r[1] = buf[1];
 196         conv->ostate = 0;
 197         return count+2;
 198       } else
 199         return RET_TOOSMALL;
 200     }
 201     ret = hkscs2004_wctomb(conv,buf,wc,2);
 202     if (ret != RET_ILUNI) {
 203       if (ret != 2) abort();
 204       if (n >= count+2) {
 205         r[0] = buf[0];
 206         r[1] = buf[1];
 207         conv->ostate = 0;
 208         return count+2;
 209       } else
 210         return RET_TOOSMALL;
 211     }
 212     return RET_ILUNI;
 213   }
 214 }
 215
 216 static int
 217 big5hkscs2004_reset (conv_t conv, unsigned char *r, size_t n)
 218 {
 219   unsigned char last = conv->ostate;
 220
 221   if (last) {
 222     if (n < 2)
 223       return RET_TOOSMALL;
 224     r[0] = 0x88;
 225     r[1] = last;
 226     /* conv->ostate = 0; will be done by the caller */
 227     return 2;
 228   } else
 229     return 0;
 230 }