lib/gbk.h

   1 /*
   2  * Copyright (C) 1999-2001 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 Library General Public
   7  * License as published by the Free Software Foundation; either version 2
   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  * Library General Public License for more details.
  14  *
  15  * You should have received a copy of the GNU Library General Public
  16  * License along with the GNU LIBICONV Library; see the file COPYING.LIB.
  17  * If not, write to the Free Software Foundation, Inc., 59 Temple Place -
  18  * Suite 330, Boston, MA 02111-1307, USA.
  19  */
  20
  21 /*
  22  * GBK
  23  */
  24
  25 /*
  26  * GBK, as described in Ken Lunde's book, is an extension of GB 2312-1980
  27  * (shifted by adding 0x8080 to the range 0xA1A1..0xFEFE, as used in EUC-CN).
  28  * It adds the following ranges:
  29  *
  30  * (part of GBK/1)  0xA2A1-0xA2AA  Small Roman numerals
  31  * GBK/3   0x{81-A0}{40-7E,80-FE}  6080 new characters, all in Unicode
  32  * GBK/4   0x{AA-FE}{40-7E,80-A0}  8160 new characters, 8080 in Unicode
  33  * GBK/5   0x{A8-A9}{40-7E,80-A0}  166 new characters, 153 in Unicode
  34  */
  35
  36 /*
  37  * CP936 is nearly identical to GBK. It differs as follows:
  38  *
  39  * 1. Some characters in the GB2312 range are defined differently:
  40  *
  41  *     code    GB2312                         CP936.TXT
  42  *    0xA1A4   0x30FB # KATAKANA MIDDLE DOT   0x00B7 # MIDDLE DOT
  43  *    0xA1AA   0x2015 # HORIZONTAL BAR        0x2014 # EM DASH
  44  *
  45  * 2. 19 characters added in the range 0xA6E0-0xA6F5.
  46  *
  47  * 3. 4 characters added in the range 0xA8BB-0xA8C0.
  48  */
  49
  50 /*
  51  * Since all three tables I have looked at
  52  *   - the CP936 table by Microsoft, found on ftp.unicode.org,
  53  *   - the GBK table by Sun, investigated on a Solaris 2.7 machine,
  54  *   - the GBK tables by CWEX, found in the Big5+ package,
  55  * all include these CP936 extensions (the CWEX tables have additional
  56  * differences), I conclude that either Ken Lunde has overlooked some of
  57  * the differences between GB2312 and GBK, or he is right but the major
  58  * vendors don't care about it. In either case, CP936 is the de facto
  59  * standard under the name "GBK", and we should better support it.
  60  *
  61  * So in what follows, when we write "GBK" we always mean "CP936".
  62  */
  63
  64 #include "gbkext1.h"
  65 #include "gbkext2.h"
  66 #include "gbkext_inv.h"
  67 #include "cp936ext.h"
  68
  69 static int
  70 gbk_mbtowc (conv_t conv, ucs4_t *pwc, const unsigned char *s, int n)
  71 {
  72   unsigned char c = *s;
  73
  74   if (c >= 0x81 && c < 0xff) {
  75     if (n < 2)
  76       return RET_TOOFEW(0);
  77     if (c >= 0xa1 && c <= 0xf7) {
  78       unsigned char c2 = s[1];
  79       if (c == 0xa1) {
  80         if (c2 == 0xa4) {
  81           *pwc = 0x00b7;
  82           return 2;
  83         }
  84         if (c2 == 0xaa) {
  85           *pwc = 0x2014;
  86           return 2;
  87         }
  88       }
  89       if (c2 >= 0xa1 && c2 < 0xff) {
  90         unsigned char buf[2];
  91         int ret;
  92         buf[0] = c-0x80; buf[1] = c2-0x80;
  93         ret = gb2312_mbtowc(conv,pwc,buf,2);
  94         if (ret != RET_ILSEQ)
  95           return ret;
  96         buf[0] = c; buf[1] = c2;
  97         ret = cp936ext_mbtowc(conv,pwc,buf,2);
  98         if (ret != RET_ILSEQ)
  99           return ret;
 100       }
 101     }
 102     if (c >= 0x81 && c <= 0xa0)
 103       return gbkext1_mbtowc(conv,pwc,s,2);
 104     if (c >= 0xa8 && c <= 0xfe)
 105       return gbkext2_mbtowc(conv,pwc,s,2);
 106     if (c == 0xa2) {
 107       unsigned char c2 = s[1];
 108       if (c2 >= 0xa1 && c2 <= 0xaa) {
 109         *pwc = 0x2170+(c2-0xa1);
 110         return 2;
 111       }
 112     }
 113   }
 114   return RET_ILSEQ;
 115 }
 116
 117 static int
 118 gbk_wctomb (conv_t conv, unsigned char *r, ucs4_t wc, int n)
 119 {
 120   unsigned char buf[2];
 121   int ret;
 122
 123   if (wc != 0x30fb && wc != 0x2015) {
 124     ret = gb2312_wctomb(conv,buf,wc,2);
 125     if (ret != RET_ILUNI) {
 126       if (ret != 2) abort();
 127       if (n < 2)
 128         return RET_TOOSMALL;
 129       r[0] = buf[0]+0x80;
 130       r[1] = buf[1]+0x80;
 131       return 2;
 132     }
 133   }
 134   ret = gbkext_inv_wctomb(conv,buf,wc,2);
 135   if (ret != RET_ILUNI) {
 136     if (ret != 2) abort();
 137     if (n < 2)
 138       return RET_TOOSMALL;
 139     r[0] = buf[0];
 140     r[1] = buf[1];
 141     return 2;
 142   }
 143   if (wc >= 0x2170 && wc <= 0x2179) {
 144     r[0] = 0xa2;
 145     r[1] = 0xa1 + (wc-0x2170);
 146     return 2;
 147   }
 148   ret = cp936ext_wctomb(conv,buf,wc,2);
 149   if (ret != RET_ILUNI) {
 150     if (ret != 2) abort();
 151     if (n < 2)
 152       return RET_TOOSMALL;
 153     r[0] = buf[0];
 154     r[1] = buf[1];
 155     return 2;
 156   }
 157   if (wc == 0x00b7) {
 158     if (n < 2)
 159       return RET_TOOSMALL;
 160     r[0] = 0xa1;
 161     r[1] = 0xa4;
 162     return 2;
 163   }
 164   if (wc == 0x2014) {
 165     if (n < 2)
 166       return RET_TOOSMALL;
 167     r[0] = 0xa1;
 168     r[1] = 0xaa;
 169     return 2;
 170   }
 171
 172   return RET_ILUNI;
 173 }