lib/gbk.h

   1 /*
   2  * Copyright (C) 1999-2001, 2005, 2008, 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  * GBK
  22  */
  23
  24 /*
  25  * GBK, as described in Ken Lunde's book, is an extension of GB 2312-1980
  26  * (shifted by adding 0x8080 to the range 0xA1A1..0xFEFE, as used in EUC-CN).
  27  * It adds the following ranges:
  28  *
  29  * (part of GBK/1)  0xA2A1-0xA2AA  Small Roman numerals
  30  * GBK/3   0x{81-A0}{40-7E,80-FE}  6080 new characters, all in Unicode
  31  * GBK/4   0x{AA-FE}{40-7E,80-A0}  8160 new characters, 8080 in Unicode
  32  * GBK/5   0x{A8-A9}{40-7E,80-A0}  166 new characters, 153 in Unicode
  33  *
  34  * Furthermore, all four tables I have looked at
  35  *   - the CP936 table by Microsoft, found on ftp.unicode.org in 1999,
  36  *   - the GBK table by Sun, investigated on a Solaris 2.7 machine,
  37  *   - the GBK tables by CWEX, found in the Big5+ package,
  38  *   - the GB18030 standard (second printing),
  39  * agree in the following extensions. (Ken Lunde must have overlooked these
  40  * differences between GB2312 and GBK. Also, the CWEX tables have additional
  41  * differences.)
  42  *
  43  * 1. Some characters in the GB2312 range are defined differently:
  44  *
  45  *     code    GB2312                         GBK
  46  *    0xA1A4   0x30FB # KATAKANA MIDDLE DOT   0x00B7 # MIDDLE DOT
  47  *    0xA1AA   0x2015 # HORIZONTAL BAR        0x2014 # EM DASH
  48  *
  49  * 2. 19 characters added in the range 0xA6E0-0xA6F5.
  50  *
  51  * 3. 4 characters added in the range 0xA8BB-0xA8C0.
  52  *
  53  * CP936 as of 1999 was identical to GBK. However, since 1999, Microsoft has
  54  * added new mappings to CP936...
  55  */
  56
  57 #include "gbkext1.h"
  58 #include "gbkext2.h"
  59 #include "gbkext_inv.h"
  60 #include "cp936ext.h"
  61
  62 static int
  63 gbk_mbtowc (conv_t conv, ucs4_t *pwc, const unsigned char *s, size_t n)
  64 {
  65   unsigned char c = *s;
  66
  67   if (c >= 0x81 && c < 0xff) {
  68     if (n < 2)
  69       return RET_TOOFEW(0);
  70     if (c >= 0xa1 && c <= 0xf7) {
  71       unsigned char c2 = s[1];
  72       if (c == 0xa1) {
  73         if (c2 == 0xa4) {
  74           *pwc = 0x00b7;
  75           return 2;
  76         }
  77         if (c2 == 0xaa) {
  78           *pwc = 0x2014;
  79           return 2;
  80         }
  81       }
  82       if (c2 >= 0xa1 && c2 < 0xff) {
  83         unsigned char buf[2];
  84         int ret;
  85         buf[0] = c-0x80; buf[1] = c2-0x80;
  86         ret = gb2312_mbtowc(conv,pwc,buf,2);
  87         if (ret != RET_ILSEQ)
  88           return ret;
  89         buf[0] = c; buf[1] = c2;
  90         ret = cp936ext_mbtowc(conv,pwc,buf,2);
  91         if (ret != RET_ILSEQ)
  92           return ret;
  93       }
  94     }
  95     if (c >= 0x81 && c <= 0xa0)
  96       return gbkext1_mbtowc(conv,pwc,s,2);
  97     if (c >= 0xa8 && c <= 0xfe)
  98       return gbkext2_mbtowc(conv,pwc,s,2);
  99     if (c == 0xa2) {
 100       unsigned char c2 = s[1];
 101       if (c2 >= 0xa1 && c2 <= 0xaa) {
 102         *pwc = 0x2170+(c2-0xa1);
 103         return 2;
 104       }
 105     }
 106   }
 107   return RET_ILSEQ;
 108 }
 109
 110 static int
 111 gbk_wctomb (conv_t conv, unsigned char *r, ucs4_t wc, size_t n)
 112 {
 113   unsigned char buf[2];
 114   int ret;
 115
 116   if (wc != 0x30fb && wc != 0x2015) {
 117     ret = gb2312_wctomb(conv,buf,wc,2);
 118     if (ret != RET_ILUNI) {
 119       if (ret != 2) abort();
 120       if (n < 2)
 121         return RET_TOOSMALL;
 122       r[0] = buf[0]+0x80;
 123       r[1] = buf[1]+0x80;
 124       return 2;
 125     }
 126   }
 127   ret = gbkext_inv_wctomb(conv,buf,wc,2);
 128   if (ret != RET_ILUNI) {
 129     if (ret != 2) abort();
 130     if (n < 2)
 131       return RET_TOOSMALL;
 132     r[0] = buf[0];
 133     r[1] = buf[1];
 134     return 2;
 135   }
 136   if (wc >= 0x2170 && wc <= 0x2179) {
 137     if (n < 2)
 138       return RET_TOOSMALL;
 139     r[0] = 0xa2;
 140     r[1] = 0xa1 + (wc-0x2170);
 141     return 2;
 142   }
 143   ret = cp936ext_wctomb(conv,buf,wc,2);
 144   if (ret != RET_ILUNI) {
 145     if (ret != 2) abort();
 146     if (n < 2)
 147       return RET_TOOSMALL;
 148     r[0] = buf[0];
 149     r[1] = buf[1];
 150     return 2;
 151   }
 152   if (wc == 0x00b7) {
 153     if (n < 2)
 154       return RET_TOOSMALL;
 155     r[0] = 0xa1;
 156     r[1] = 0xa4;
 157     return 2;
 158   }
 159   if (wc == 0x2014) {
 160     if (n < 2)
 161       return RET_TOOSMALL;
 162     r[0] = 0xa1;
 163     r[1] = 0xaa;
 164     return 2;
 165   }
 166
 167   return RET_ILUNI;
 168 }