lib/iso2022_jp1.h

   1 /*
   2  * Copyright (C) 1999-2001, 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  * ISO-2022-JP-1
  22  */
  23
  24 /* Specification: RFC 2237 */
  25
  26 #define ESC 0x1b
  27
  28 /*
  29  * The state can be one of the following values.
  30  */
  31 #define STATE_ASCII          0
  32 #define STATE_JISX0201ROMAN  1
  33 #define STATE_JISX0208       2
  34 #define STATE_JISX0212       3
  35
  36 static int
  37 iso2022_jp1_mbtowc (conv_t conv, ucs4_t *pwc, const unsigned char *s, size_t n)
  38 {
  39   state_t state = conv->istate;
  40   int count = 0;
  41   unsigned char c;
  42   for (;;) {
  43     c = *s;
  44     if (c == ESC) {
  45       if (n < count+3)
  46         goto none;
  47       if (s[1] == '(') {
  48         if (s[2] == 'B') {
  49           state = STATE_ASCII;
  50           s += 3; count += 3;
  51           if (n < count+1)
  52             goto none;
  53           continue;
  54         }
  55         if (s[2] == 'J') {
  56           state = STATE_JISX0201ROMAN;
  57           s += 3; count += 3;
  58           if (n < count+1)
  59             goto none;
  60           continue;
  61         }
  62         goto ilseq;
  63       }
  64       if (s[1] == '$') {
  65         if (s[2] == '@' || s[2] == 'B') {
  66           /* We don't distinguish JIS X 0208-1978 and JIS X 0208-1983. */
  67           state = STATE_JISX0208;
  68           s += 3; count += 3;
  69           if (n < count+1)
  70             goto none;
  71           continue;
  72         }
  73         if (s[2] == '(') {
  74           if (n < count+4)
  75             goto none;
  76           if (s[3] == 'D') {
  77             state = STATE_JISX0212;
  78             s += 4; count += 4;
  79             if (n < count+1)
  80               goto none;
  81             continue;
  82           }
  83         }
  84         goto ilseq;
  85       }
  86       goto ilseq;
  87     }
  88     break;
  89   }
  90   switch (state) {
  91     case STATE_ASCII:
  92       if (c < 0x80) {
  93         int ret = ascii_mbtowc(conv,pwc,s,1);
  94         if (ret == RET_ILSEQ)
  95           goto ilseq;
  96         if (ret != 1) abort();
  97         conv->istate = state;
  98         return count+1;
  99       } else
 100         goto ilseq;
 101     case STATE_JISX0201ROMAN:
 102       if (c < 0x80) {
 103         int ret = jisx0201_mbtowc(conv,pwc,s,1);
 104         if (ret == RET_ILSEQ)
 105           goto ilseq;
 106         if (ret != 1) abort();
 107         conv->istate = state;
 108         return count+1;
 109       } else
 110         goto ilseq;
 111     case STATE_JISX0208:
 112       if (n < count+2)
 113         goto none;
 114       if (s[0] < 0x80 && s[1] < 0x80) {
 115         int ret = jisx0208_mbtowc(conv,pwc,s,2);
 116         if (ret == RET_ILSEQ)
 117           goto ilseq;
 118         if (ret != 2) abort();
 119         conv->istate = state;
 120         return count+2;
 121       } else
 122         goto ilseq;
 123     case STATE_JISX0212:
 124       if (n < count+2)
 125         goto none;
 126       if (s[0] < 0x80 && s[1] < 0x80) {
 127         int ret = jisx0212_mbtowc(conv,pwc,s,2);
 128         if (ret == RET_ILSEQ)
 129           goto ilseq;
 130         if (ret != 2) abort();
 131         conv->istate = state;
 132         return count+2;
 133       } else
 134         goto ilseq;
 135     default: abort();
 136   }
 137
 138 none:
 139   conv->istate = state;
 140   return RET_TOOFEW(count);
 141
 142 ilseq:
 143   conv->istate = state;
 144   return RET_SHIFT_ILSEQ(count);
 145 }
 146
 147 static int
 148 iso2022_jp1_wctomb (conv_t conv, unsigned char *r, ucs4_t wc, size_t n)
 149 {
 150   state_t state = conv->ostate;
 151   unsigned char buf[2];
 152   int ret;
 153
 154   /* Try ASCII. */
 155   ret = ascii_wctomb(conv,buf,wc,1);
 156   if (ret != RET_ILUNI) {
 157     if (ret != 1) abort();
 158     if (buf[0] < 0x80) {
 159       int count = (state == STATE_ASCII ? 1 : 4);
 160       if (n < count)
 161         return RET_TOOSMALL;
 162       if (state != STATE_ASCII) {
 163         r[0] = ESC;
 164         r[1] = '(';
 165         r[2] = 'B';
 166         r += 3;
 167         state = STATE_ASCII;
 168       }
 169       r[0] = buf[0];
 170       conv->ostate = state;
 171       return count;
 172     }
 173   }
 174
 175   /* Try JIS X 0201-1976 Roman. */
 176   ret = jisx0201_wctomb(conv,buf,wc,1);
 177   if (ret != RET_ILUNI) {
 178     if (ret != 1) abort();
 179     if (buf[0] < 0x80) {
 180       int count = (state == STATE_JISX0201ROMAN ? 1 : 4);
 181       if (n < count)
 182         return RET_TOOSMALL;
 183       if (state != STATE_JISX0201ROMAN) {
 184         r[0] = ESC;
 185         r[1] = '(';
 186         r[2] = 'J';
 187         r += 3;
 188         state = STATE_JISX0201ROMAN;
 189       }
 190       r[0] = buf[0];
 191       conv->ostate = state;
 192       return count;
 193     }
 194   }
 195
 196   /* Try JIS X 0208-1990 in place of JIS X 0208-1978 and JIS X 0208-1983. */
 197   ret = jisx0208_wctomb(conv,buf,wc,2);
 198   if (ret != RET_ILUNI) {
 199     if (ret != 2) abort();
 200     if (buf[0] < 0x80 && buf[1] < 0x80) {
 201       int count = (state == STATE_JISX0208 ? 2 : 5);
 202       if (n < count)
 203         return RET_TOOSMALL;
 204       if (state != STATE_JISX0208) {
 205         r[0] = ESC;
 206         r[1] = '$';
 207         r[2] = 'B';
 208         r += 3;
 209         state = STATE_JISX0208;
 210       }
 211       r[0] = buf[0];
 212       r[1] = buf[1];
 213       conv->ostate = state;
 214       return count;
 215     }
 216   }
 217
 218   /* Try JIS X 0212-1990. */
 219   ret = jisx0212_wctomb(conv,buf,wc,2);
 220   if (ret != RET_ILUNI) {
 221     if (ret != 2) abort();
 222     if (buf[0] < 0x80 && buf[1] < 0x80) {
 223       int count = (state == STATE_JISX0212 ? 2 : 6);
 224       if (n < count)
 225         return RET_TOOSMALL;
 226       if (state != STATE_JISX0212) {
 227         r[0] = ESC;
 228         r[1] = '$';
 229         r[2] = '(';
 230         r[3] = 'D';
 231         r += 4;
 232         state = STATE_JISX0212;
 233       }
 234       r[0] = buf[0];
 235       r[1] = buf[1];
 236       conv->ostate = state;
 237       return count;
 238     }
 239   }
 240
 241   return RET_ILUNI;
 242 }
 243
 244 static int
 245 iso2022_jp1_reset (conv_t conv, unsigned char *r, size_t n)
 246 {
 247   state_t state = conv->ostate;
 248   if (state != STATE_ASCII) {
 249     if (n < 3)
 250       return RET_TOOSMALL;
 251     r[0] = ESC;
 252     r[1] = '(';
 253     r[2] = 'B';
 254     /* conv->ostate = 0; will be done by the caller */
 255     return 3;
 256   } else
 257     return 0;
 258 }
 259
 260 #undef STATE_JISX0212
 261 #undef STATE_JISX0208
 262 #undef STATE_JISX0201ROMAN
 263 #undef STATE_ASCII