lib/big5hkscs2008.h

   1 /*
   2  * Copyright (C) 1999-2002, 2006, 2010 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:2008
  22  */
  23
  24 /*
  25  * BIG5-HKSCS:2008 can be downloaded from
  26  *   http://www.ogcio.gov.hk/ccli/eng/hkscs/download.html
  27  *   http://www.ogcio.gov.hk/ccli/eng/hkscs/introduction.html
  28  *
  29  * It extends BIG5-HKSCS:2004 through 68 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 "hkscs2008.h"
  48 #include "flushwc.h"
  49
  50 static int
  51 big5hkscs2008_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     {
  95       int ret = hkscs2008_mbtowc(conv,pwc,s,n);
  96       if (ret != RET_ILSEQ)
  97         return ret;
  98     }
  99     if (c == 0x88) {
 100       if (n < 2)
 101         return RET_TOOFEW(0);
 102       {
 103         unsigned char c2 = s[1];
 104         if (c2 == 0x62 || c2 == 0x64 || c2 == 0xa3 || c2 == 0xa5) {
 105           /* It's a composed character. */
 106           ucs4_t wc1 = ((c2 >> 3) << 2) + 0x009a; /* = 0x00ca or 0x00ea */
 107           ucs4_t wc2 = ((c2 & 6) << 2) + 0x02fc; /* = 0x0304 or 0x030c */
 108           /* We cannot output two Unicode characters at once. So,
 109              output the first character and buffer the second one. */
 110           *pwc = wc1;
 111           conv->istate = wc2;
 112           return 2;
 113         }
 114       }
 115     }
 116     return RET_ILSEQ;
 117   }
 118 }
 119
 120 #define big5hkscs2008_flushwc normal_flushwc
 121
 122 static int
 123 big5hkscs2008_wctomb (conv_t conv, unsigned char *r, ucs4_t wc, size_t n)
 124 {
 125   int count = 0;
 126   unsigned char last = conv->ostate;
 127
 128   if (last) {
 129     /* last is = 0x66 or = 0xa7. */
 130     if (wc == 0x0304 || wc == 0x030c) {
 131       /* Output the combined character. */
 132       if (n >= 2) {
 133         r[0] = 0x88;
 134         r[1] = last + ((wc & 24) >> 2) - 4; /* = 0x62 or 0x64 or 0xa3 or 0xa5 */
 135         conv->ostate = 0;
 136         return 2;
 137       } else
 138         return RET_TOOSMALL;
 139     }
 140
 141     /* Output the buffered character. */
 142     if (n < 2)
 143       return RET_TOOSMALL;
 144     r[0] = 0x88;
 145     r[1] = last;
 146     r += 2;
 147     count = 2;
 148   }
 149
 150   /* Code set 0 (ASCII) */
 151   if (wc < 0x0080) {
 152     /* Plain ASCII character. */
 153     if (n > count) {
 154       r[0] = (unsigned char) wc;
 155       conv->ostate = 0;
 156       return count+1;
 157     } else
 158       return RET_TOOSMALL;
 159   } else {
 160     unsigned char buf[2];
 161     int ret;
 162
 163     /* Code set 1 (BIG5 extended) */
 164     ret = big5_wctomb(conv,buf,wc,2);
 165     if (ret != RET_ILUNI) {
 166       if (ret != 2) abort();
 167       if (!((buf[0] == 0xc6 && buf[1] >= 0xa1) || buf[0] == 0xc7)) {
 168         if (n >= count+2) {
 169           r[0] = buf[0];
 170           r[1] = buf[1];
 171           conv->ostate = 0;
 172           return count+2;
 173         } else
 174           return RET_TOOSMALL;
 175       }
 176     }
 177     ret = hkscs1999_wctomb(conv,buf,wc,2);
 178     if (ret != RET_ILUNI) {
 179       if (ret != 2) abort();
 180       if ((wc & ~0x0020) == 0x00ca) {
 181         /* A possible first character of a multi-character sequence. We have to
 182            buffer it. */
 183         if (!(buf[0] == 0x88 && (buf[1] == 0x66 || buf[1] == 0xa7))) abort();
 184         conv->ostate = buf[1]; /* = 0x66 or = 0xa7 */
 185         return count+0;
 186       }
 187       if (n >= count+2) {
 188         r[0] = buf[0];
 189         r[1] = buf[1];
 190         conv->ostate = 0;
 191         return count+2;
 192       } else
 193         return RET_TOOSMALL;
 194     }
 195     ret = hkscs2001_wctomb(conv,buf,wc,2);
 196     if (ret != RET_ILUNI) {
 197       if (ret != 2) abort();
 198       if (n >= count+2) {
 199         r[0] = buf[0];
 200         r[1] = buf[1];
 201         conv->ostate = 0;
 202         return count+2;
 203       } else
 204         return RET_TOOSMALL;
 205     }
 206     ret = hkscs2004_wctomb(conv,buf,wc,2);
 207     if (ret != RET_ILUNI) {
 208       if (ret != 2) abort();
 209       if (n >= count+2) {
 210         r[0] = buf[0];
 211         r[1] = buf[1];
 212         conv->ostate = 0;
 213         return count+2;
 214       } else
 215         return RET_TOOSMALL;
 216     }
 217     ret = hkscs2008_wctomb(conv,buf,wc,2);
 218     if (ret != RET_ILUNI) {
 219       if (ret != 2) abort();
 220       if (n >= count+2) {
 221         r[0] = buf[0];
 222         r[1] = buf[1];
 223         conv->ostate = 0;
 224         return count+2;
 225       } else
 226         return RET_TOOSMALL;
 227     }
 228     return RET_ILUNI;
 229   }
 230 }
 231
 232 static int
 233 big5hkscs2008_reset (conv_t conv, unsigned char *r, size_t n)
 234 {
 235   unsigned char last = conv->ostate;
 236
 237   if (last) {
 238     if (n < 2)
 239       return RET_TOOSMALL;
 240     r[0] = 0x88;
 241     r[1] = last;
 242     /* conv->ostate = 0; will be done by the caller */
 243     return 2;
 244   } else
 245     return 0;
 246 }