lib/cp1255.h

   1 /*
   2  * Copyright (C) 1999-2001, 2004, 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  * CP1255
  22  */
  23
  24 #include "flushwc.h"
  25
  26 /* Combining characters used in Hebrew encoding CP1255. */
  27
  28 /* Relevant combining characters:
  29    0x05b4, 0x05b7, 0x05b8, 0x05b9, 0x05bc, 0x05bf, 0x05c1, 0x05c2. */
  30
  31 /* Composition tables for each of the relevant combining characters. */
  32 static const struct { unsigned short base; unsigned short composed; } cp1255_comp_table_data[] = {
  33 #define cp1255_comp_table05b4_idx 0
  34 #define cp1255_comp_table05b4_len 1
  35   { 0x05D9, 0xFB1D },
  36 #define cp1255_comp_table05b7_idx (cp1255_comp_table05b4_idx+cp1255_comp_table05b4_len)
  37 #define cp1255_comp_table05b7_len 2
  38   { 0x05D0, 0xFB2E },
  39   { 0x05F2, 0xFB1F },
  40 #define cp1255_comp_table05b8_idx (cp1255_comp_table05b7_idx+cp1255_comp_table05b7_len)
  41 #define cp1255_comp_table05b8_len 1
  42   { 0x05D0, 0xFB2F },
  43 #define cp1255_comp_table05b9_idx (cp1255_comp_table05b8_idx+cp1255_comp_table05b8_len)
  44 #define cp1255_comp_table05b9_len 1
  45   { 0x05D5, 0xFB4B },
  46 #define cp1255_comp_table05bc_idx (cp1255_comp_table05b9_idx+cp1255_comp_table05b9_len)
  47 #define cp1255_comp_table05bc_len 24
  48   { 0x05D0, 0xFB30 },
  49   { 0x05D1, 0xFB31 },
  50   { 0x05D2, 0xFB32 },
  51   { 0x05D3, 0xFB33 },
  52   { 0x05D4, 0xFB34 },
  53   { 0x05D5, 0xFB35 },
  54   { 0x05D6, 0xFB36 },
  55   { 0x05D8, 0xFB38 },
  56   { 0x05D9, 0xFB39 },
  57   { 0x05DA, 0xFB3A },
  58   { 0x05DB, 0xFB3B },
  59   { 0x05DC, 0xFB3C },
  60   { 0x05DE, 0xFB3E },
  61   { 0x05E0, 0xFB40 },
  62   { 0x05E1, 0xFB41 },
  63   { 0x05E3, 0xFB43 },
  64   { 0x05E4, 0xFB44 },
  65   { 0x05E6, 0xFB46 },
  66   { 0x05E7, 0xFB47 },
  67   { 0x05E8, 0xFB48 },
  68   { 0x05E9, 0xFB49 },
  69   { 0x05EA, 0xFB4A },
  70   { 0xFB2A, 0xFB2C },
  71   { 0xFB2B, 0xFB2D },
  72 #define cp1255_comp_table05bf_idx (cp1255_comp_table05bc_idx+cp1255_comp_table05bc_len)
  73 #define cp1255_comp_table05bf_len 3
  74   { 0x05D1, 0xFB4C },
  75   { 0x05DB, 0xFB4D },
  76   { 0x05E4, 0xFB4E },
  77 #define cp1255_comp_table05c1_idx (cp1255_comp_table05bf_idx+cp1255_comp_table05bf_len)
  78 #define cp1255_comp_table05c1_len 2
  79   { 0x05E9, 0xFB2A },
  80   { 0xFB49, 0xFB2C },
  81 #define cp1255_comp_table05c2_idx (cp1255_comp_table05c1_idx+cp1255_comp_table05c1_len)
  82 #define cp1255_comp_table05c2_len 2
  83   { 0x05E9, 0xFB2B },
  84   { 0xFB49, 0xFB2D },
  85 };
  86 static const struct { unsigned int len; unsigned int idx; } cp1255_comp_table[] = {
  87   { cp1255_comp_table05b4_len, cp1255_comp_table05b4_idx },
  88   { cp1255_comp_table05b7_len, cp1255_comp_table05b7_idx },
  89   { cp1255_comp_table05b8_len, cp1255_comp_table05b8_idx },
  90   { cp1255_comp_table05b9_len, cp1255_comp_table05b9_idx },
  91   { cp1255_comp_table05bc_len, cp1255_comp_table05bc_idx },
  92   { cp1255_comp_table05bf_len, cp1255_comp_table05bf_idx },
  93   { cp1255_comp_table05c1_len, cp1255_comp_table05c1_idx },
  94   { cp1255_comp_table05c2_len, cp1255_comp_table05c2_idx },
  95 };
  96
  97 /* Decomposition table for the relevant Unicode characters. */
  98 struct cp1255_decomp { unsigned short composed; unsigned short base; int comb1 : 8; signed int comb2 : 8; };
  99 static const struct cp1255_decomp cp1255_decomp_table[] = {
 100   { 0xFB1D, 0x05D9, 0, -1 },
 101   { 0xFB1F, 0x05F2, 1, -1 },
 102   { 0xFB2A, 0x05E9, 6, -1 },
 103   { 0xFB2B, 0x05E9, 7, -1 },
 104   { 0xFB2C, 0x05E9, 4, 6 },
 105   { 0xFB2D, 0x05E9, 4, 7 },
 106   { 0xFB2E, 0x05D0, 1, -1 },
 107   { 0xFB2F, 0x05D0, 2, -1 },
 108   { 0xFB30, 0x05D0, 4, -1 },
 109   { 0xFB31, 0x05D1, 4, -1 },
 110   { 0xFB32, 0x05D2, 4, -1 },
 111   { 0xFB33, 0x05D3, 4, -1 },
 112   { 0xFB34, 0x05D4, 4, -1 },
 113   { 0xFB35, 0x05D5, 4, -1 },
 114   { 0xFB36, 0x05D6, 4, -1 },
 115   { 0xFB38, 0x05D8, 4, -1 },
 116   { 0xFB39, 0x05D9, 4, -1 },
 117   { 0xFB3A, 0x05DA, 4, -1 },
 118   { 0xFB3B, 0x05DB, 4, -1 },
 119   { 0xFB3C, 0x05DC, 4, -1 },
 120   { 0xFB3E, 0x05DE, 4, -1 },
 121   { 0xFB40, 0x05E0, 4, -1 },
 122   { 0xFB41, 0x05E1, 4, -1 },
 123   { 0xFB43, 0x05E3, 4, -1 },
 124   { 0xFB44, 0x05E4, 4, -1 },
 125   { 0xFB46, 0x05E6, 4, -1 },
 126   { 0xFB47, 0x05E7, 4, -1 },
 127   { 0xFB48, 0x05E8, 4, -1 },
 128   { 0xFB49, 0x05E9, 4, -1 },
 129   { 0xFB4A, 0x05EA, 4, -1 },
 130   { 0xFB4B, 0x05D5, 3, -1 },
 131   { 0xFB4C, 0x05D1, 5, -1 },
 132   { 0xFB4D, 0x05DB, 5, -1 },
 133   { 0xFB4E, 0x05E4, 5, -1 },
 134 };
 135
 136 static const unsigned char cp1255_comb_table[] = {
 137   0xc4, 0xc7, 0xc8, 0xc9, 0xcc, 0xcf, 0xd1, 0xd2,
 138 };
 139
 140 static const unsigned short cp1255_2uni[128] = {
 141   /* 0x80 */
 142   0x20ac, 0xfffd, 0x201a, 0x0192, 0x201e, 0x2026, 0x2020, 0x2021,
 143   0x02c6, 0x2030, 0xfffd, 0x2039, 0xfffd, 0xfffd, 0xfffd, 0xfffd,
 144   /* 0x90 */
 145   0xfffd, 0x2018, 0x2019, 0x201c, 0x201d, 0x2022, 0x2013, 0x2014,
 146   0x02dc, 0x2122, 0xfffd, 0x203a, 0xfffd, 0xfffd, 0xfffd, 0xfffd,
 147   /* 0xa0 */
 148   0x00a0, 0x00a1, 0x00a2, 0x00a3, 0x20aa, 0x00a5, 0x00a6, 0x00a7,
 149   0x00a8, 0x00a9, 0x00d7, 0x00ab, 0x00ac, 0x00ad, 0x00ae, 0x00af,
 150   /* 0xb0 */
 151   0x00b0, 0x00b1, 0x00b2, 0x00b3, 0x00b4, 0x00b5, 0x00b6, 0x00b7,
 152   0x00b8, 0x00b9, 0x00f7, 0x00bb, 0x00bc, 0x00bd, 0x00be, 0x00bf,
 153   /* 0xc0 */
 154   0x05b0, 0x05b1, 0x05b2, 0x05b3, 0x05b4, 0x05b5, 0x05b6, 0x05b7,
 155   0x05b8, 0x05b9, 0x05ba, 0x05bb, 0x05bc, 0x05bd, 0x05be, 0x05bf,
 156   /* 0xd0 */
 157   0x05c0, 0x05c1, 0x05c2, 0x05c3, 0x05f0, 0x05f1, 0x05f2, 0x05f3,
 158   0x05f4, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd,
 159   /* 0xe0 */
 160   0x05d0, 0x05d1, 0x05d2, 0x05d3, 0x05d4, 0x05d5, 0x05d6, 0x05d7,
 161   0x05d8, 0x05d9, 0x05da, 0x05db, 0x05dc, 0x05dd, 0x05de, 0x05df,
 162   /* 0xf0 */
 163   0x05e0, 0x05e1, 0x05e2, 0x05e3, 0x05e4, 0x05e5, 0x05e6, 0x05e7,
 164   0x05e8, 0x05e9, 0x05ea, 0xfffd, 0xfffd, 0x200e, 0x200f, 0xfffd,
 165 };
 166
 167 /* In the CP1255 to Unicode direction, the state contains a buffered
 168    character, or 0 if none. */
 169
 170 static int
 171 cp1255_mbtowc (conv_t conv, ucs4_t *pwc, const unsigned char *s, size_t n)
 172 {
 173   unsigned char c = *s;
 174   unsigned short wc;
 175   unsigned short last_wc;
 176   if (c < 0x80) {
 177     wc = c;
 178   } else {
 179     wc = cp1255_2uni[c-0x80];
 180     if (wc == 0xfffd)
 181       return RET_ILSEQ;
 182   }
 183   last_wc = conv->istate;
 184   if (last_wc) {
 185     if (wc >= 0x05b0 && wc < 0x05c5) {
 186       /* See whether last_wc and wc can be combined. */
 187       unsigned int k;
 188       unsigned int i1, i2;
 189       switch (wc) {
 190         case 0x05b4: k = 0; break;
 191         case 0x05b7: k = 1; break;
 192         case 0x05b8: k = 2; break;
 193         case 0x05b9: k = 3; break;
 194         case 0x05bc: k = 4; break;
 195         case 0x05bf: k = 5; break;
 196         case 0x05c1: k = 6; break;
 197         case 0x05c2: k = 7; break;
 198         default: goto not_combining;
 199       }
 200       i1 = cp1255_comp_table[k].idx;
 201       i2 = i1 + cp1255_comp_table[k].len-1;
 202       if (last_wc >= cp1255_comp_table_data[i1].base
 203           && last_wc <= cp1255_comp_table_data[i2].base) {
 204         unsigned int i;
 205         for (;;) {
 206           i = (i1+i2)>>1;
 207           if (last_wc == cp1255_comp_table_data[i].base)
 208             break;
 209           if (last_wc < cp1255_comp_table_data[i].base) {
 210             if (i1 == i)
 211               goto not_combining;
 212             i2 = i;
 213           } else {
 214             if (i1 != i)
 215               i1 = i;
 216             else {
 217               i = i2;
 218               if (last_wc == cp1255_comp_table_data[i].base)
 219                 break;
 220               goto not_combining;
 221             }
 222           }
 223         }
 224         last_wc = cp1255_comp_table_data[i].composed;
 225         if (last_wc == 0xfb2a || last_wc == 0xfb2b || last_wc == 0xfb49) {
 226           /* Buffer the combined character. */
 227           conv->istate = last_wc;
 228           return RET_TOOFEW(1);
 229         } else {
 230           /* Output the combined character. */
 231           conv->istate = 0;
 232           *pwc = (ucs4_t) last_wc;
 233           return 1;
 234         }
 235       }
 236     }
 237   not_combining:
 238     /* Output the buffered character. */
 239     conv->istate = 0;
 240     *pwc = (ucs4_t) last_wc;
 241     return 0; /* Don't advance the input pointer. */
 242   }
 243   if ((wc >= 0x05d0 && wc <= 0x05ea && ((0x07db5f7f >> (wc - 0x05d0)) & 1))
 244       || wc == 0x05f2) {
 245     /* wc is a possible match in cp1255_comp_table_data. Buffer it. */
 246     conv->istate = wc;
 247     return RET_TOOFEW(1);
 248   } else {
 249     /* Output wc immediately. */
 250     *pwc = (ucs4_t) wc;
 251     return 1;
 252   }
 253 }
 254
 255 #define cp1255_flushwc normal_flushwc
 256
 257 static const unsigned char cp1255_page00[88] = {
 258   0xa0, 0xa1, 0xa2, 0xa3, 0x00, 0xa5, 0xa6, 0xa7, /* 0xa0-0xa7 */
 259   0xa8, 0xa9, 0x00, 0xab, 0xac, 0xad, 0xae, 0xaf, /* 0xa8-0xaf */
 260   0xb0, 0xb1, 0xb2, 0xb3, 0xb4, 0xb5, 0xb6, 0xb7, /* 0xb0-0xb7 */
 261   0xb8, 0xb9, 0x00, 0xbb, 0xbc, 0xbd, 0xbe, 0xbf, /* 0xb8-0xbf */
 262   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0xc0-0xc7 */
 263   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0xc8-0xcf */
 264   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0xaa, /* 0xd0-0xd7 */
 265   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0xd8-0xdf */
 266   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0xe0-0xe7 */
 267   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0xe8-0xef */
 268   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0xba, /* 0xf0-0xf7 */
 269 };
 270 static const unsigned char cp1255_page02[32] = {
 271   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x88, 0x00, /* 0xc0-0xc7 */
 272   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0xc8-0xcf */
 273   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0xd0-0xd7 */
 274   0x00, 0x00, 0x00, 0x00, 0x98, 0x00, 0x00, 0x00, /* 0xd8-0xdf */
 275 };
 276 static const unsigned char cp1255_page05[72] = {
 277   0xc0, 0xc1, 0xc2, 0xc3, 0xc4, 0xc5, 0xc6, 0xc7, /* 0xb0-0xb7 */
 278   0xc8, 0xc9, 0xca, 0xcb, 0xcc, 0xcd, 0xce, 0xcf, /* 0xb8-0xbf */
 279   0xd0, 0xd1, 0xd2, 0xd3, 0x00, 0x00, 0x00, 0x00, /* 0xc0-0xc7 */
 280   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0xc8-0xcf */
 281   0xe0, 0xe1, 0xe2, 0xe3, 0xe4, 0xe5, 0xe6, 0xe7, /* 0xd0-0xd7 */
 282   0xe8, 0xe9, 0xea, 0xeb, 0xec, 0xed, 0xee, 0xef, /* 0xd8-0xdf */
 283   0xf0, 0xf1, 0xf2, 0xf3, 0xf4, 0xf5, 0xf6, 0xf7, /* 0xe0-0xe7 */
 284   0xf8, 0xf9, 0xfa, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0xe8-0xef */
 285   0xd4, 0xd5, 0xd6, 0xd7, 0xd8, 0x00, 0x00, 0x00, /* 0xf0-0xf7 */
 286 };
 287 static const unsigned char cp1255_page20[56] = {
 288   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0xfd, 0xfe, /* 0x08-0x0f */
 289   0x00, 0x00, 0x00, 0x96, 0x97, 0x00, 0x00, 0x00, /* 0x10-0x17 */
 290   0x91, 0x92, 0x82, 0x00, 0x93, 0x94, 0x84, 0x00, /* 0x18-0x1f */
 291   0x86, 0x87, 0x95, 0x00, 0x00, 0x00, 0x85, 0x00, /* 0x20-0x27 */
 292   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x28-0x2f */
 293   0x89, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x30-0x37 */
 294   0x00, 0x8b, 0x9b, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x38-0x3f */
 295 };
 296
 297 static int
 298 cp1255_wctomb (conv_t conv, unsigned char *r, ucs4_t wc, size_t n)
 299 {
 300   unsigned char c = 0;
 301   if (wc < 0x0080) {
 302     *r = wc;
 303     return 1;
 304   }
 305   else if (wc >= 0x00a0 && wc < 0x00f8)
 306     c = cp1255_page00[wc-0x00a0];
 307   else if (wc == 0x0192)
 308     c = 0x83;
 309   else if (wc >= 0x02c0 && wc < 0x02e0)
 310     c = cp1255_page02[wc-0x02c0];
 311   else if (wc >= 0x05b0 && wc < 0x05f8)
 312     c = cp1255_page05[wc-0x05b0];
 313   else if (wc >= 0x2008 && wc < 0x2040)
 314     c = cp1255_page20[wc-0x2008];
 315   else if (wc == 0x20aa)
 316     c = 0xa4;
 317   else if (wc == 0x20ac)
 318     c = 0x80;
 319   else if (wc == 0x2122)
 320     c = 0x99;
 321   if (c != 0) {
 322     *r = c;
 323     return 1;
 324   }
 325   /* Try canonical decomposition. */
 326   {
 327     /* Binary search through cp1255_decomp_table. */
 328     unsigned int i1 = 0;
 329     unsigned int i2 = sizeof(cp1255_decomp_table)/sizeof(cp1255_decomp_table[0])-1;
 330     if (wc >= cp1255_decomp_table[i1].composed
 331         && wc <= cp1255_decomp_table[i2].composed) {
 332       unsigned int i;
 333       for (;;) {
 334         /* Here i2 - i1 > 0. */
 335         i = (i1+i2)>>1;
 336         if (wc == cp1255_decomp_table[i].composed)
 337           break;
 338         if (wc < cp1255_decomp_table[i].composed) {
 339           if (i1 == i)
 340             return RET_ILUNI;
 341           /* Here i1 < i < i2. */
 342           i2 = i;
 343         } else {
 344           /* Here i1 <= i < i2. */
 345           if (i1 != i)
 346             i1 = i;
 347           else {
 348             /* Here i2 - i1 = 1. */
 349             i = i2;
 350             if (wc == cp1255_decomp_table[i].composed)
 351               break;
 352             else
 353               return RET_ILUNI;
 354           }
 355         }
 356       }
 357       /* Found a canonical decomposition. */
 358       wc = cp1255_decomp_table[i].base;
 359       /* wc is one of 0x05d0..0x05d6, 0x05d8..0x05dc, 0x05de, 0x05e0..0x05e1,
 360          0x05e3..0x05e4, 0x05e6..0x05ea, 0x05f2. */
 361       c = cp1255_page05[wc-0x05b0];
 362       if (cp1255_decomp_table[i].comb2 < 0) {
 363         if (n < 2)
 364           return RET_TOOSMALL;
 365         r[0] = c;
 366         r[1] = cp1255_comb_table[cp1255_decomp_table[i].comb1];
 367         return 2;
 368       } else {
 369         if (n < 3)
 370           return RET_TOOSMALL;
 371         r[0] = c;
 372         r[1] = cp1255_comb_table[cp1255_decomp_table[i].comb1];
 373         r[2] = cp1255_comb_table[cp1255_decomp_table[i].comb2];
 374         return 3;
 375       }
 376     }
 377   }
 378   return RET_ILUNI;
 379 }