lib/tcvn.h

   1 /*
   2  * Copyright (C) 1999-2002, 2004, 2016, 2019 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  * TCVN-5712
  22  */
  23
  24 #include "flushwc.h"
  25 #include "vietcomb.h"
  26
  27 static const unsigned char tcvn_comb_table[] = {
  28   0xb0, 0xb3, 0xb2, 0xb1, 0xb4,
  29 };
  30
  31 /* The possible bases in viet_comp_table_data:
  32    0x0041..0x0045, 0x0047..0x0049, 0x004B..0x0050, 0x0052..0x0057,
  33    0x0059..0x005A, 0x0061..0x0065, 0x0067..0x0069, 0x006B..0x0070,
  34    0x0072..0x0077, 0x0079..0x007A, 0x00A5, 0x00C2, 0x00CA, 0x00D3..0x00D6,
  35    0x00DA, 0x00E2, 0x00EA, 0x00F3..0x00F6, 0x00FA, 0x0102..0x0103,
  36    0x0168..0x0169, 0x01A0..0x01A1, 0x01AF..0x01B0. */
  37 static const unsigned int tcvn_comp_bases[] = {
  38   0x06fdfbbe, 0x06fdfbbe, 0x00000000, 0x00000020, 0x04780404, 0x04780404,
  39   0x0000000c, 0x00000000, 0x00000000, 0x00000300, 0x00000000, 0x00018003
  40 };
  41
  42 static const unsigned short tcvn_2uni_1[24] = {
  43   /* 0x00 */
  44   0x0000, 0x00da, 0x1ee4, 0x0003, 0x1eea, 0x1eec, 0x1eee, 0x0007,
  45   0x0008, 0x0009, 0x000a, 0x000b, 0x000c, 0x000d, 0x000e, 0x000f,
  46   /* 0x10 */
  47   0x0010, 0x1ee8, 0x1ef0, 0x1ef2, 0x1ef6, 0x1ef8, 0x00dd, 0x1ef4,
  48 };
  49 static const unsigned short tcvn_2uni_2[128] = {
  50   /* 0x80 */
  51   0x00c0, 0x1ea2, 0x00c3, 0x00c1, 0x1ea0, 0x1eb6, 0x1eac, 0x00c8,
  52   0x1eba, 0x1ebc, 0x00c9, 0x1eb8, 0x1ec6, 0x00cc, 0x1ec8, 0x0128,
  53   /* 0x90 */
  54   0x00cd, 0x1eca, 0x00d2, 0x1ece, 0x00d5, 0x00d3, 0x1ecc, 0x1ed8,
  55   0x1edc, 0x1ede, 0x1ee0, 0x1eda, 0x1ee2, 0x00d9, 0x1ee6, 0x0168,
  56   /* 0xa0 */
  57   0x00a0, 0x0102, 0x00c2, 0x00ca, 0x00d4, 0x01a0, 0x01af, 0x0110,
  58   0x0103, 0x00e2, 0x00ea, 0x00f4, 0x01a1, 0x01b0, 0x0111, 0x1eb0,
  59   /* 0xb0 */
  60   0x0300, 0x0309, 0x0303, 0x0301, 0x0323, 0x00e0, 0x1ea3, 0x00e3,
  61   0x00e1, 0x1ea1, 0x1eb2, 0x1eb1, 0x1eb3, 0x1eb5, 0x1eaf, 0x1eb4,
  62   /* 0xc0 */
  63   0x1eae, 0x1ea6, 0x1ea8, 0x1eaa, 0x1ea4, 0x1ec0, 0x1eb7, 0x1ea7,
  64   0x1ea9, 0x1eab, 0x1ea5, 0x1ead, 0x00e8, 0x1ec2, 0x1ebb, 0x1ebd,
  65   /* 0xd0 */
  66   0x00e9, 0x1eb9, 0x1ec1, 0x1ec3, 0x1ec5, 0x1ebf, 0x1ec7, 0x00ec,
  67   0x1ec9, 0x1ec4, 0x1ebe, 0x1ed2, 0x0129, 0x00ed, 0x1ecb, 0x00f2,
  68   /* 0xe0 */
  69   0x1ed4, 0x1ecf, 0x00f5, 0x00f3, 0x1ecd, 0x1ed3, 0x1ed5, 0x1ed7,
  70   0x1ed1, 0x1ed9, 0x1edd, 0x1edf, 0x1ee1, 0x1edb, 0x1ee3, 0x00f9,
  71   /* 0xf0 */
  72   0x1ed6, 0x1ee7, 0x0169, 0x00fa, 0x1ee5, 0x1eeb, 0x1eed, 0x1eef,
  73   0x1ee9, 0x1ef1, 0x1ef3, 0x1ef7, 0x1ef9, 0x00fd, 0x1ef5, 0x1ed0,
  74 };
  75
  76 /* In the TCVN to Unicode direction, the state contains a buffered
  77    character, or 0 if none. */
  78
  79 static int
  80 tcvn_mbtowc (conv_t conv, ucs4_t *pwc, const unsigned char *s, size_t n)
  81 {
  82   unsigned char c = *s;
  83   unsigned short wc;
  84   unsigned short last_wc;
  85   if (c < 0x18)
  86     wc = tcvn_2uni_1[c];
  87   else if (c < 0x80)
  88     wc = c;
  89   else
  90     wc = tcvn_2uni_2[c-0x80];
  91   last_wc = conv->istate;
  92   if (last_wc) {
  93     if (wc >= 0x0300 && wc < 0x0340) {
  94       /* See whether last_wc and wc can be combined. */
  95       unsigned int k;
  96       unsigned int i1, i2;
  97       switch (wc) {
  98         case 0x0300: k = 0; break;
  99         case 0x0301: k = 1; break;
 100         case 0x0303: k = 2; break;
 101         case 0x0309: k = 3; break;
 102         case 0x0323: k = 4; break;
 103         default: abort();
 104       }
 105       i1 = viet_comp_table[k].idx;
 106       i2 = i1 + viet_comp_table[k].len-1;
 107       if (last_wc >= viet_comp_table_data[i1].base
 108           && last_wc <= viet_comp_table_data[i2].base) {
 109         unsigned int i;
 110         for (;;) {
 111           i = (i1+i2)>>1;
 112           if (last_wc == viet_comp_table_data[i].base)
 113             break;
 114           if (last_wc < viet_comp_table_data[i].base) {
 115             if (i1 == i)
 116               goto not_combining;
 117             i2 = i;
 118           } else {
 119             if (i1 != i)
 120               i1 = i;
 121             else {
 122               i = i2;
 123               if (last_wc == viet_comp_table_data[i].base)
 124                 break;
 125               goto not_combining;
 126             }
 127           }
 128         }
 129         last_wc = viet_comp_table_data[i].composed;
 130         /* Output the combined character. */
 131         conv->istate = 0;
 132         *pwc = (ucs4_t) last_wc;
 133         return 1;
 134       }
 135     }
 136   not_combining:
 137     /* Output the buffered character. */
 138     conv->istate = 0;
 139     *pwc = (ucs4_t) last_wc;
 140     return 0; /* Don't advance the input pointer. */
 141   }
 142   if (wc >= 0x0041 && wc <= 0x01b0
 143       && ((tcvn_comp_bases[(wc - 0x0040) >> 5] >> (wc & 0x1f)) & 1)) {
 144     /* wc is a possible match in viet_comp_table_data. Buffer it. */
 145     conv->istate = wc;
 146     return RET_TOOFEW(1);
 147   } else {
 148     /* Output wc immediately. */
 149     *pwc = (ucs4_t) wc;
 150     return 1;
 151   }
 152 }
 153
 154 #define tcvn_flushwc normal_flushwc
 155
 156 static const unsigned char tcvn_page00[96+184] = {
 157   0xa0, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0xa0-0xa7 */
 158   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0xa8-0xaf */
 159   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0xb0-0xb7 */
 160   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0xb8-0xbf */
 161   0x80, 0x83, 0xa2, 0x82, 0x00, 0x00, 0x00, 0x00, /* 0xc0-0xc7 */
 162   0x87, 0x8a, 0xa3, 0x00, 0x8d, 0x90, 0x00, 0x00, /* 0xc8-0xcf */
 163   0x00, 0x00, 0x92, 0x95, 0xa4, 0x94, 0x00, 0x00, /* 0xd0-0xd7 */
 164   0x00, 0x9d, 0x01, 0x00, 0x00, 0x16, 0x00, 0x00, /* 0xd8-0xdf */
 165   0xb5, 0xb8, 0xa9, 0xb7, 0x00, 0x00, 0x00, 0x00, /* 0xe0-0xe7 */
 166   0xcc, 0xd0, 0xaa, 0x00, 0xd7, 0xdd, 0x00, 0x00, /* 0xe8-0xef */
 167   0x00, 0x00, 0xdf, 0xe3, 0xab, 0xe2, 0x00, 0x00, /* 0xf0-0xf7 */
 168   0x00, 0xef, 0xf3, 0x00, 0x00, 0xfd, 0x00, 0x00, /* 0xf8-0xff */
 169   /* 0x0100 */
 170   0x00, 0x00, 0xa1, 0xa8, 0x00, 0x00, 0x00, 0x00, /* 0x00-0x07 */
 171   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x08-0x0f */
 172   0xa7, 0xae, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x10-0x17 */
 173   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x18-0x1f */
 174   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x20-0x27 */
 175   0x8f, 0xdc, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x28-0x2f */
 176   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x30-0x37 */
 177   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x38-0x3f */
 178   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x40-0x47 */
 179   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x48-0x4f */
 180   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x50-0x57 */
 181   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x58-0x5f */
 182   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x60-0x67 */
 183   0x9f, 0xf2, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x68-0x6f */
 184   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x70-0x77 */
 185   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x78-0x7f */
 186   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x80-0x87 */
 187   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x88-0x8f */
 188   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x90-0x97 */
 189   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x98-0x9f */
 190   0xa5, 0xac, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0xa0-0xa7 */
 191   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0xa6, /* 0xa8-0xaf */
 192   0xad, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0xb0-0xb7 */
 193 };
 194 static const unsigned char tcvn_page03[40] = {
 195   0xb0, 0xb3, 0x00, 0xb2, 0x00, 0x00, 0x00, 0x00, /* 0x00-0x07 */
 196   0x00, 0xb1, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x08-0x0f */
 197   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x10-0x17 */
 198   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x18-0x1f */
 199   0x00, 0x00, 0x00, 0xb4, 0x00, 0x00, 0x00, 0x00, /* 0x20-0x27 */
 200 };
 201 static const unsigned char tcvn_page1e[96] = {
 202   0x84, 0xb9, 0x81, 0xb6, 0xc4, 0xca, 0xc1, 0xc7, /* 0xa0-0xa7 */
 203   0xc2, 0xc8, 0xc3, 0xc9, 0x86, 0xcb, 0xc0, 0xbe, /* 0xa8-0xaf */
 204   0xaf, 0xbb, 0xba, 0xbc, 0xbf, 0xbd, 0x85, 0xc6, /* 0xb0-0xb7 */
 205   0x8b, 0xd1, 0x88, 0xce, 0x89, 0xcf, 0xda, 0xd5, /* 0xb8-0xbf */
 206   0xc5, 0xd2, 0xcd, 0xd3, 0xd9, 0xd4, 0x8c, 0xd6, /* 0xc0-0xc7 */
 207   0x8e, 0xd8, 0x91, 0xde, 0x96, 0xe4, 0x93, 0xe1, /* 0xc8-0xcf */
 208   0xff, 0xe8, 0xdb, 0xe5, 0xe0, 0xe6, 0xf0, 0xe7, /* 0xd0-0xd7 */
 209   0x97, 0xe9, 0x9b, 0xed, 0x98, 0xea, 0x99, 0xeb, /* 0xd8-0xdf */
 210   0x9a, 0xec, 0x9c, 0xee, 0x02, 0xf4, 0x9e, 0xf1, /* 0xe0-0xe7 */
 211   0x11, 0xf8, 0x04, 0xf5, 0x05, 0xf6, 0x06, 0xf7, /* 0xe8-0xef */
 212   0x12, 0xf9, 0x13, 0xfa, 0x17, 0xfe, 0x14, 0xfb, /* 0xf0-0xf7 */
 213   0x15, 0xfc, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0xf8-0xff */
 214 };
 215
 216 static int
 217 tcvn_wctomb (conv_t conv, unsigned char *r, ucs4_t wc, size_t n)
 218 {
 219   unsigned char c = 0;
 220   if (wc < 0x0080 && (wc >= 0x0020 || (0x00fe0076U & (1U << wc)) == 0)) {
 221     *r = wc;
 222     return 1;
 223   }
 224   else if (wc >= 0x00a0 && wc < 0x01b8)
 225     c = tcvn_page00[wc-0x00a0];
 226   else if (wc >= 0x0300 && wc < 0x0328)
 227     c = tcvn_page03[wc-0x0300];
 228   else if (wc >= 0x0340 && wc < 0x0342) /* deprecated Vietnamese tone marks */
 229     c = tcvn_page03[wc-0x0340];
 230   else if (wc >= 0x1ea0 && wc < 0x1f00)
 231     c = tcvn_page1e[wc-0x1ea0];
 232   if (c != 0) {
 233     *r = c;
 234     return 1;
 235   }
 236   /* Try compatibility or canonical decomposition. */
 237   {
 238     /* Binary search through viet_decomp_table. */
 239     unsigned int i1 = 0;
 240     unsigned int i2 = sizeof(viet_decomp_table)/sizeof(viet_decomp_table[0])-1;
 241     if (wc >= viet_decomp_table[i1].composed
 242         && wc <= viet_decomp_table[i2].composed) {
 243       unsigned int i;
 244       for (;;) {
 245         /* Here i2 - i1 > 0. */
 246         i = (i1+i2)>>1;
 247         if (wc == viet_decomp_table[i].composed)
 248           break;
 249         if (wc < viet_decomp_table[i].composed) {
 250           if (i1 == i)
 251             return RET_ILUNI;
 252           /* Here i1 < i < i2. */
 253           i2 = i;
 254         } else {
 255           /* Here i1 <= i < i2. */
 256           if (i1 != i)
 257             i1 = i;
 258           else {
 259             /* Here i2 - i1 = 1. */
 260             i = i2;
 261             if (wc == viet_decomp_table[i].composed)
 262               break;
 263             else
 264               return RET_ILUNI;
 265           }
 266         }
 267       }
 268       /* Found a compatibility or canonical decomposition. */
 269       wc = viet_decomp_table[i].base;
 270       /* wc is one of 0x0020, 0x0041..0x005a, 0x0061..0x007a, 0x00a5, 0x00a8,
 271          0x00c2, 0x00c5..0x00c7, 0x00ca, 0x00cf, 0x00d3, 0x00d4, 0x00d6,
 272          0x00d8, 0x00da, 0x00dc, 0x00e2, 0x00e5..0x00e7, 0x00ea, 0x00ef,
 273          0x00f3, 0x00f4, 0x00f6, 0x00f8, 0x00fc, 0x0102, 0x0103, 0x01a0,
 274          0x01a1, 0x01af, 0x01b0. */
 275       if (wc < 0x0080)
 276         c = wc;
 277       else {
 278         c = tcvn_page00[wc-0x00a0];
 279         if (c == 0)
 280           return RET_ILUNI;
 281       }
 282       if (n < 2)
 283         return RET_TOOSMALL;
 284       r[0] = c;
 285       r[1] = tcvn_comb_table[viet_decomp_table[i].comb1];
 286       return 2;
 287     }
 288   }
 289   return RET_ILUNI;
 290 }