1 /* 2 * Copyright (C) 1999-2001, 2004 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., 51 Franklin Street, 18 * Fifth Floor, Boston, MA 02110-1301, USA. 19 */ 20 21 /* 22 * CP1255 23 */ 24 25 #include "flushwc.h" 26 27 /* Combining characters used in Hebrew encoding CP1255. */ 28 29 /* Relevant combining characters: 30 0x05b4, 0x05b7, 0x05b8, 0x05b9, 0x05bc, 0x05bf, 0x05c1, 0x05c2. */ 31 32 /* Composition tables for each of the relevant combining characters. */ 33 static const struct { unsigned short base; unsigned short composed; } cp1255_comp_table_data[] = { 34 #define cp1255_comp_table05b4_idx 0 35 #define cp1255_comp_table05b4_len 1 36 { 0x05D9, 0xFB1D }, 37 #define cp1255_comp_table05b7_idx (cp1255_comp_table05b4_idx+cp1255_comp_table05b4_len) 38 #define cp1255_comp_table05b7_len 2 39 { 0x05D0, 0xFB2E }, 40 { 0x05F2, 0xFB1F }, 41 #define cp1255_comp_table05b8_idx (cp1255_comp_table05b7_idx+cp1255_comp_table05b7_len) 42 #define cp1255_comp_table05b8_len 1 43 { 0x05D0, 0xFB2F }, 44 #define cp1255_comp_table05b9_idx (cp1255_comp_table05b8_idx+cp1255_comp_table05b8_len) 45 #define cp1255_comp_table05b9_len 1 46 { 0x05D5, 0xFB4B }, 47 #define cp1255_comp_table05bc_idx (cp1255_comp_table05b9_idx+cp1255_comp_table05b9_len) 48 #define cp1255_comp_table05bc_len 24 49 { 0x05D0, 0xFB30 }, 50 { 0x05D1, 0xFB31 }, 51 { 0x05D2, 0xFB32 }, 52 { 0x05D3, 0xFB33 }, 53 { 0x05D4, 0xFB34 }, 54 { 0x05D5, 0xFB35 }, 55 { 0x05D6, 0xFB36 }, 56 { 0x05D8, 0xFB38 }, 57 { 0x05D9, 0xFB39 }, 58 { 0x05DA, 0xFB3A }, 59 { 0x05DB, 0xFB3B }, 60 { 0x05DC, 0xFB3C }, 61 { 0x05DE, 0xFB3E }, 62 { 0x05E0, 0xFB40 }, 63 { 0x05E1, 0xFB41 }, 64 { 0x05E3, 0xFB43 }, 65 { 0x05E4, 0xFB44 }, 66 { 0x05E6, 0xFB46 }, 67 { 0x05E7, 0xFB47 }, 68 { 0x05E8, 0xFB48 }, 69 { 0x05E9, 0xFB49 }, 70 { 0x05EA, 0xFB4A }, 71 { 0xFB2A, 0xFB2C }, 72 { 0xFB2B, 0xFB2D }, 73 #define cp1255_comp_table05bf_idx (cp1255_comp_table05bc_idx+cp1255_comp_table05bc_len) 74 #define cp1255_comp_table05bf_len 3 75 { 0x05D1, 0xFB4C }, 76 { 0x05DB, 0xFB4D }, 77 { 0x05E4, 0xFB4E }, 78 #define cp1255_comp_table05c1_idx (cp1255_comp_table05bf_idx+cp1255_comp_table05bf_len) 79 #define cp1255_comp_table05c1_len 2 80 { 0x05E9, 0xFB2A }, 81 { 0xFB49, 0xFB2C }, 82 #define cp1255_comp_table05c2_idx (cp1255_comp_table05c1_idx+cp1255_comp_table05c1_len) 83 #define cp1255_comp_table05c2_len 2 84 { 0x05E9, 0xFB2B }, 85 { 0xFB49, 0xFB2D }, 86 }; 87 static const struct { unsigned int len; unsigned int idx; } cp1255_comp_table[] = { 88 { cp1255_comp_table05b4_len, cp1255_comp_table05b4_idx }, 89 { cp1255_comp_table05b7_len, cp1255_comp_table05b7_idx }, 90 { cp1255_comp_table05b8_len, cp1255_comp_table05b8_idx }, 91 { cp1255_comp_table05b9_len, cp1255_comp_table05b9_idx }, 92 { cp1255_comp_table05bc_len, cp1255_comp_table05bc_idx }, 93 { cp1255_comp_table05bf_len, cp1255_comp_table05bf_idx }, 94 { cp1255_comp_table05c1_len, cp1255_comp_table05c1_idx }, 95 { cp1255_comp_table05c2_len, cp1255_comp_table05c2_idx }, 96 }; 97 98 /* Decomposition table for the relevant Unicode characters. */ 99 struct cp1255_decomp { unsigned short composed; unsigned short base; int comb1 : 8; signed int comb2 : 8; }; 100 static const struct cp1255_decomp cp1255_decomp_table[] = { 101 { 0xFB1D, 0x05D9, 0, -1 }, 102 { 0xFB1F, 0x05F2, 1, -1 }, 103 { 0xFB2A, 0x05E9, 6, -1 }, 104 { 0xFB2B, 0x05E9, 7, -1 }, 105 { 0xFB2C, 0x05E9, 4, 6 }, 106 { 0xFB2D, 0x05E9, 4, 7 }, 107 { 0xFB2E, 0x05D0, 1, -1 }, 108 { 0xFB2F, 0x05D0, 2, -1 }, 109 { 0xFB30, 0x05D0, 4, -1 }, 110 { 0xFB31, 0x05D1, 4, -1 }, 111 { 0xFB32, 0x05D2, 4, -1 }, 112 { 0xFB33, 0x05D3, 4, -1 }, 113 { 0xFB34, 0x05D4, 4, -1 }, 114 { 0xFB35, 0x05D5, 4, -1 }, 115 { 0xFB36, 0x05D6, 4, -1 }, 116 { 0xFB38, 0x05D8, 4, -1 }, 117 { 0xFB39, 0x05D9, 4, -1 }, 118 { 0xFB3A, 0x05DA, 4, -1 }, 119 { 0xFB3B, 0x05DB, 4, -1 }, 120 { 0xFB3C, 0x05DC, 4, -1 }, 121 { 0xFB3E, 0x05DE, 4, -1 }, 122 { 0xFB40, 0x05E0, 4, -1 }, 123 { 0xFB41, 0x05E1, 4, -1 }, 124 { 0xFB43, 0x05E3, 4, -1 }, 125 { 0xFB44, 0x05E4, 4, -1 }, 126 { 0xFB46, 0x05E6, 4, -1 }, 127 { 0xFB47, 0x05E7, 4, -1 }, 128 { 0xFB48, 0x05E8, 4, -1 }, 129 { 0xFB49, 0x05E9, 4, -1 }, 130 { 0xFB4A, 0x05EA, 4, -1 }, 131 { 0xFB4B, 0x05D5, 3, -1 }, 132 { 0xFB4C, 0x05D1, 5, -1 }, 133 { 0xFB4D, 0x05DB, 5, -1 }, 134 { 0xFB4E, 0x05E4, 5, -1 }, 135 }; 136 137 static const unsigned char cp1255_comb_table[] = { 138 0xc4, 0xc7, 0xc8, 0xc9, 0xcc, 0xcf, 0xd1, 0xd2, 139 }; 140 141 static const unsigned short cp1255_2uni[128] = { 142 /* 0x80 */ 143 0x20ac, 0xfffd, 0x201a, 0x0192, 0x201e, 0x2026, 0x2020, 0x2021, 144 0x02c6, 0x2030, 0xfffd, 0x2039, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 145 /* 0x90 */ 146 0xfffd, 0x2018, 0x2019, 0x201c, 0x201d, 0x2022, 0x2013, 0x2014, 147 0x02dc, 0x2122, 0xfffd, 0x203a, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 148 /* 0xa0 */ 149 0x00a0, 0x00a1, 0x00a2, 0x00a3, 0x20aa, 0x00a5, 0x00a6, 0x00a7, 150 0x00a8, 0x00a9, 0x00d7, 0x00ab, 0x00ac, 0x00ad, 0x00ae, 0x00af, 151 /* 0xb0 */ 152 0x00b0, 0x00b1, 0x00b2, 0x00b3, 0x00b4, 0x00b5, 0x00b6, 0x00b7, 153 0x00b8, 0x00b9, 0x00f7, 0x00bb, 0x00bc, 0x00bd, 0x00be, 0x00bf, 154 /* 0xc0 */ 155 0x05b0, 0x05b1, 0x05b2, 0x05b3, 0x05b4, 0x05b5, 0x05b6, 0x05b7, 156 0x05b8, 0x05b9, 0xfffd, 0x05bb, 0x05bc, 0x05bd, 0x05be, 0x05bf, 157 /* 0xd0 */ 158 0x05c0, 0x05c1, 0x05c2, 0x05c3, 0x05f0, 0x05f1, 0x05f2, 0x05f3, 159 0x05f4, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 160 /* 0xe0 */ 161 0x05d0, 0x05d1, 0x05d2, 0x05d3, 0x05d4, 0x05d5, 0x05d6, 0x05d7, 162 0x05d8, 0x05d9, 0x05da, 0x05db, 0x05dc, 0x05dd, 0x05de, 0x05df, 163 /* 0xf0 */ 164 0x05e0, 0x05e1, 0x05e2, 0x05e3, 0x05e4, 0x05e5, 0x05e6, 0x05e7, 165 0x05e8, 0x05e9, 0x05ea, 0xfffd, 0xfffd, 0x200e, 0x200f, 0xfffd, 166 }; 167 168 /* In the CP1255 to Unicode direction, the state contains a buffered 169 character, or 0 if none. */ 170 171 static int 172 cp1255_mbtowc (conv_t conv, ucs4_t *pwc, const unsigned char *s, int n) 173 { 174 unsigned char c = *s; 175 unsigned short wc; 176 unsigned short last_wc; 177 if (c < 0x80) { 178 wc = c; 179 } else { 180 wc = cp1255_2uni[c-0x80]; 181 if (wc == 0xfffd) 182 return RET_ILSEQ; 183 } 184 last_wc = conv->istate; 185 if (last_wc) { 186 if (wc >= 0x05b0 && wc < 0x05c5) { 187 /* See whether last_wc and wc can be combined. */ 188 unsigned int k; 189 unsigned int i1, i2; 190 switch (wc) { 191 case 0x05b4: k = 0; break; 192 case 0x05b7: k = 1; break; 193 case 0x05b8: k = 2; break; 194 case 0x05b9: k = 3; break; 195 case 0x05bc: k = 4; break; 196 case 0x05bf: k = 5; break; 197 case 0x05c1: k = 6; break; 198 case 0x05c2: k = 7; break; 199 default: goto not_combining; 200 } 201 i1 = cp1255_comp_table[k].idx; 202 i2 = i1 + cp1255_comp_table[k].len-1; 203 if (last_wc >= cp1255_comp_table_data[i1].base 204 && last_wc <= cp1255_comp_table_data[i2].base) { 205 unsigned int i; 206 for (;;) { 207 i = (i1+i2)>>1; 208 if (last_wc == cp1255_comp_table_data[i].base) 209 break; 210 if (last_wc < cp1255_comp_table_data[i].base) { 211 if (i1 == i) 212 goto not_combining; 213 i2 = i; 214 } else { 215 if (i1 != i) 216 i1 = i; 217 else { 218 i = i2; 219 if (last_wc == cp1255_comp_table_data[i].base) 220 break; 221 goto not_combining; 222 } 223 } 224 } 225 last_wc = cp1255_comp_table_data[i].composed; 226 if (last_wc == 0xfb2a || last_wc == 0xfb2b || last_wc == 0xfb49) { 227 /* Buffer the combined character. */ 228 conv->istate = last_wc; 229 return RET_TOOFEW(1); 230 } else { 231 /* Output the combined character. */ 232 conv->istate = 0; 233 *pwc = (ucs4_t) last_wc; 234 return 1; 235 } 236 } 237 } 238 not_combining: 239 /* Output the buffered character. */ 240 conv->istate = 0; 241 *pwc = (ucs4_t) last_wc; 242 return 0; /* Don't advance the input pointer. */ 243 } 244 if ((wc >= 0x05d0 && wc <= 0x05ea && ((0x07db5f7f >> (wc - 0x05d0)) & 1)) 245 || wc == 0x05f2) { 246 /* wc is a possible match in cp1255_comp_table_data. Buffer it. */ 247 conv->istate = wc; 248 return RET_TOOFEW(1); 249 } else { 250 /* Output wc immediately. */ 251 *pwc = (ucs4_t) wc; 252 return 1; 253 } 254 } 255 256 #define cp1255_flushwc normal_flushwc 257 258 static const unsigned char cp1255_page00[88] = { 259 0xa0, 0xa1, 0xa2, 0xa3, 0x00, 0xa5, 0xa6, 0xa7, /* 0xa0-0xa7 */ 260 0xa8, 0xa9, 0x00, 0xab, 0xac, 0xad, 0xae, 0xaf, /* 0xa8-0xaf */ 261 0xb0, 0xb1, 0xb2, 0xb3, 0xb4, 0xb5, 0xb6, 0xb7, /* 0xb0-0xb7 */ 262 0xb8, 0xb9, 0x00, 0xbb, 0xbc, 0xbd, 0xbe, 0xbf, /* 0xb8-0xbf */ 263 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0xc0-0xc7 */ 264 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0xc8-0xcf */ 265 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0xaa, /* 0xd0-0xd7 */ 266 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0xd8-0xdf */ 267 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0xe0-0xe7 */ 268 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0xe8-0xef */ 269 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0xba, /* 0xf0-0xf7 */ 270 }; 271 static const unsigned char cp1255_page02[32] = { 272 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x88, 0x00, /* 0xc0-0xc7 */ 273 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0xc8-0xcf */ 274 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0xd0-0xd7 */ 275 0x00, 0x00, 0x00, 0x00, 0x98, 0x00, 0x00, 0x00, /* 0xd8-0xdf */ 276 }; 277 static const unsigned char cp1255_page05[72] = { 278 0xc0, 0xc1, 0xc2, 0xc3, 0xc4, 0xc5, 0xc6, 0xc7, /* 0xb0-0xb7 */ 279 0xc8, 0xc9, 0x00, 0xcb, 0xcc, 0xcd, 0xce, 0xcf, /* 0xb8-0xbf */ 280 0xd0, 0xd1, 0xd2, 0xd3, 0x00, 0x00, 0x00, 0x00, /* 0xc0-0xc7 */ 281 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0xc8-0xcf */ 282 0xe0, 0xe1, 0xe2, 0xe3, 0xe4, 0xe5, 0xe6, 0xe7, /* 0xd0-0xd7 */ 283 0xe8, 0xe9, 0xea, 0xeb, 0xec, 0xed, 0xee, 0xef, /* 0xd8-0xdf */ 284 0xf0, 0xf1, 0xf2, 0xf3, 0xf4, 0xf5, 0xf6, 0xf7, /* 0xe0-0xe7 */ 285 0xf8, 0xf9, 0xfa, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0xe8-0xef */ 286 0xd4, 0xd5, 0xd6, 0xd7, 0xd8, 0x00, 0x00, 0x00, /* 0xf0-0xf7 */ 287 }; 288 static const unsigned char cp1255_page20[56] = { 289 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0xfd, 0xfe, /* 0x08-0x0f */ 290 0x00, 0x00, 0x00, 0x96, 0x97, 0x00, 0x00, 0x00, /* 0x10-0x17 */ 291 0x91, 0x92, 0x82, 0x00, 0x93, 0x94, 0x84, 0x00, /* 0x18-0x1f */ 292 0x86, 0x87, 0x95, 0x00, 0x00, 0x00, 0x85, 0x00, /* 0x20-0x27 */ 293 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x28-0x2f */ 294 0x89, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x30-0x37 */ 295 0x00, 0x8b, 0x9b, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x38-0x3f */ 296 }; 297 298 static int 299 cp1255_wctomb (conv_t conv, unsigned char *r, ucs4_t wc, int n) 300 { 301 unsigned char c = 0; 302 if (wc < 0x0080) { 303 *r = wc; 304 return 1; 305 } 306 else if (wc >= 0x00a0 && wc < 0x00f8) 307 c = cp1255_page00[wc-0x00a0]; 308 else if (wc == 0x0192) 309 c = 0x83; 310 else if (wc >= 0x02c0 && wc < 0x02e0) 311 c = cp1255_page02[wc-0x02c0]; 312 else if (wc >= 0x05b0 && wc < 0x05f8) 313 c = cp1255_page05[wc-0x05b0]; 314 else if (wc >= 0x2008 && wc < 0x2040) 315 c = cp1255_page20[wc-0x2008]; 316 else if (wc == 0x20aa) 317 c = 0xa4; 318 else if (wc == 0x20ac) 319 c = 0x80; 320 else if (wc == 0x2122) 321 c = 0x99; 322 if (c != 0) { 323 *r = c; 324 return 1; 325 } 326 /* Try canonical decomposition. */ 327 { 328 /* Binary search through cp1255_decomp_table. */ 329 unsigned int i1 = 0; 330 unsigned int i2 = sizeof(cp1255_decomp_table)/sizeof(cp1255_decomp_table[0])-1; 331 if (wc >= cp1255_decomp_table[i1].composed 332 && wc <= cp1255_decomp_table[i2].composed) { 333 unsigned int i; 334 for (;;) { 335 /* Here i2 - i1 > 0. */ 336 i = (i1+i2)>>1; 337 if (wc == cp1255_decomp_table[i].composed) 338 break; 339 if (wc < cp1255_decomp_table[i].composed) { 340 if (i1 == i) 341 return RET_ILUNI; 342 /* Here i1 < i < i2. */ 343 i2 = i; 344 } else { 345 /* Here i1 <= i < i2. */ 346 if (i1 != i) 347 i1 = i; 348 else { 349 /* Here i2 - i1 = 1. */ 350 i = i2; 351 if (wc == cp1255_decomp_table[i].composed) 352 break; 353 else 354 return RET_ILUNI; 355 } 356 } 357 } 358 /* Found a canonical decomposition. */ 359 wc = cp1255_decomp_table[i].base; 360 /* wc is one of 0x05d0..0x05d6, 0x05d8..0x05dc, 0x05de, 0x05e0..0x05e1, 361 0x05e3..0x05e4, 0x05e6..0x05ea, 0x05f2. */ 362 c = cp1255_page05[wc-0x05b0]; 363 if (cp1255_decomp_table[i].comb2 < 0) { 364 if (n < 2) 365 return RET_TOOSMALL; 366 r[0] = c; 367 r[1] = cp1255_comb_table[cp1255_decomp_table[i].comb1]; 368 return 2; 369 } else { 370 if (n < 3) 371 return RET_TOOSMALL; 372 r[0] = c; 373 r[1] = cp1255_comb_table[cp1255_decomp_table[i].comb1]; 374 r[2] = cp1255_comb_table[cp1255_decomp_table[i].comb2]; 375 return 3; 376 } 377 } 378 } 379 return RET_ILUNI; 380 } 381