lib/cp1255.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  * CP1255
  23  */
  24
  25 /* Combining characters used in Hebrew encoding CP1255. */
  26
  27 #ifdef notusedyet
  28
  29 /* Relevant combining characters. */
  30 static const unsigned short cp1255_comb_table_uni[] = {
  31   0x05b4, 0x05b7, 0x05b8, 0x05b9, 0x05bc, 0x05bf, 0x05c1, 0x05c2,
  32 };
  33
  34 /* Composition tables for each of the relevant combining characters. */
  35 static const unsigned short cp1255_comp_table_data[][2] = {
  36 #define cp1255_comp_table05b4_idx 0
  37 #define cp1255_comp_table05b4_len 1
  38   { 0x05D9, 0xFB1D },
  39 #define cp1255_comp_table05b7_idx (cp1255_comp_table05b4_idx+cp1255_comp_table05b4_len)
  40 #define cp1255_comp_table05b7_len 2
  41   { 0x05D0, 0xFB2E },
  42   { 0x05F2, 0xFB1F },
  43 #define cp1255_comp_table05b8_idx (cp1255_comp_table05b7_idx+cp1255_comp_table05b7_len)
  44 #define cp1255_comp_table05b8_len 1
  45   { 0x05D0, 0xFB2F },
  46 #define cp1255_comp_table05b9_idx (cp1255_comp_table05b8_idx+cp1255_comp_table05b8_len)
  47 #define cp1255_comp_table05b9_len 1
  48   { 0x05D5, 0xFB4B },
  49 #define cp1255_comp_table05bc_idx (cp1255_comp_table05b9_idx+cp1255_comp_table05b9_len)
  50 #define cp1255_comp_table05bc_len 24
  51   { 0x05D0, 0xFB30 },
  52   { 0x05D1, 0xFB31 },
  53   { 0x05D2, 0xFB32 },
  54   { 0x05D3, 0xFB33 },
  55   { 0x05D4, 0xFB34 },
  56   { 0x05D5, 0xFB35 },
  57   { 0x05D6, 0xFB36 },
  58   { 0x05D8, 0xFB38 },
  59   { 0x05D9, 0xFB39 },
  60   { 0x05DA, 0xFB3A },
  61   { 0x05DB, 0xFB3B },
  62   { 0x05DC, 0xFB3C },
  63   { 0x05DE, 0xFB3E },
  64   { 0x05E0, 0xFB40 },
  65   { 0x05E1, 0xFB41 },
  66   { 0x05E3, 0xFB43 },
  67   { 0x05E4, 0xFB44 },
  68   { 0x05E6, 0xFB46 },
  69   { 0x05E7, 0xFB47 },
  70   { 0x05E8, 0xFB48 },
  71   { 0x05E9, 0xFB49 },
  72   { 0x05EA, 0xFB4A },
  73   { 0xFB2A, 0xFB2C },
  74   { 0xFB2B, 0xFB2D },
  75 #define cp1255_comp_table05bf_idx (cp1255_comp_table05bc_idx+cp1255_comp_table05bc_len)
  76 #define cp1255_comp_table05bf_len 3
  77   { 0x05D1, 0xFB4C },
  78   { 0x05DB, 0xFB4D },
  79   { 0x05E4, 0xFB4E },
  80 #define cp1255_comp_table05c1_idx (cp1255_comp_table05bf_idx+cp1255_comp_table05bf_len)
  81 #define cp1255_comp_table05c1_len 2
  82   { 0x05E9, 0xFB2A },
  83   { 0xFB49, 0xFB2C },
  84 #define cp1255_comp_table05c2_idx (cp1255_comp_table05c1_idx+cp1255_comp_table05c1_len)
  85 #define cp1255_comp_table05c2_len 2
  86   { 0x05E9, 0xFB2B },
  87   { 0xFB49, 0xFB2D },
  88 };
  89 static const struct { unsigned int len; unsigned int offset; } cp1255_comp_table[] = {
  90   { cp1255_comp_table05b4_len, cp1255_comp_table05b4_idx },
  91   { cp1255_comp_table05b7_len, cp1255_comp_table05b7_idx },
  92   { cp1255_comp_table05b8_len, cp1255_comp_table05b8_idx },
  93   { cp1255_comp_table05b9_len, cp1255_comp_table05b9_idx },
  94   { cp1255_comp_table05bc_len, cp1255_comp_table05bc_idx },
  95   { cp1255_comp_table05bf_len, cp1255_comp_table05bf_idx },
  96   { cp1255_comp_table05c1_len, cp1255_comp_table05c1_idx },
  97   { cp1255_comp_table05c2_len, cp1255_comp_table05c2_idx },
  98 };
  99
 100 #endif
 101
 102 /* Decomposition table for the relevant Unicode characters. */
 103 struct cp1255_decomp { unsigned short composed; unsigned short base; int comb1 : 8; int comb2 : 8; };
 104 static const struct cp1255_decomp cp1255_decomp_table[] = {
 105   { 0xFB1D, 0x05D9, 0, -1 },
 106   { 0xFB1F, 0x05F2, 1, -1 },
 107   { 0xFB2A, 0x05E9, 6, -1 },
 108   { 0xFB2B, 0x05E9, 7, -1 },
 109   { 0xFB2C, 0x05E9, 4, 6 },
 110   { 0xFB2D, 0x05E9, 4, 7 },
 111   { 0xFB2E, 0x05D0, 1, -1 },
 112   { 0xFB2F, 0x05D0, 2, -1 },
 113   { 0xFB30, 0x05D0, 4, -1 },
 114   { 0xFB31, 0x05D1, 4, -1 },
 115   { 0xFB32, 0x05D2, 4, -1 },
 116   { 0xFB33, 0x05D3, 4, -1 },
 117   { 0xFB34, 0x05D4, 4, -1 },
 118   { 0xFB35, 0x05D5, 4, -1 },
 119   { 0xFB36, 0x05D6, 4, -1 },
 120   { 0xFB38, 0x05D8, 4, -1 },
 121   { 0xFB39, 0x05D9, 4, -1 },
 122   { 0xFB3A, 0x05DA, 4, -1 },
 123   { 0xFB3B, 0x05DB, 4, -1 },
 124   { 0xFB3C, 0x05DC, 4, -1 },
 125   { 0xFB3E, 0x05DE, 4, -1 },
 126   { 0xFB40, 0x05E0, 4, -1 },
 127   { 0xFB41, 0x05E1, 4, -1 },
 128   { 0xFB43, 0x05E3, 4, -1 },
 129   { 0xFB44, 0x05E4, 4, -1 },
 130   { 0xFB46, 0x05E6, 4, -1 },
 131   { 0xFB47, 0x05E7, 4, -1 },
 132   { 0xFB48, 0x05E8, 4, -1 },
 133   { 0xFB49, 0x05E9, 4, -1 },
 134   { 0xFB4A, 0x05EA, 4, -1 },
 135   { 0xFB4B, 0x05D5, 3, -1 },
 136   { 0xFB4C, 0x05D1, 5, -1 },
 137   { 0xFB4D, 0x05DB, 5, -1 },
 138   { 0xFB4E, 0x05E4, 5, -1 },
 139 };
 140
 141 static const unsigned char cp1255_comb_table[] = {
 142   0xc4, 0xc7, 0xc8, 0xc9, 0xcc, 0xcf, 0xd1, 0xd2,
 143 };
 144
 145 static const unsigned short cp1255_2uni[128] = {
 146   /* 0x80 */
 147   0x20ac, 0xfffd, 0x201a, 0x0192, 0x201e, 0x2026, 0x2020, 0x2021,
 148   0x02c6, 0x2030, 0xfffd, 0x2039, 0xfffd, 0xfffd, 0xfffd, 0xfffd,
 149   /* 0x90 */
 150   0xfffd, 0x2018, 0x2019, 0x201c, 0x201d, 0x2022, 0x2013, 0x2014,
 151   0x02dc, 0x2122, 0xfffd, 0x203a, 0xfffd, 0xfffd, 0xfffd, 0xfffd,
 152   /* 0xa0 */
 153   0x00a0, 0x00a1, 0x00a2, 0x00a3, 0x20aa, 0x00a5, 0x00a6, 0x00a7,
 154   0x00a8, 0x00a9, 0x00d7, 0x00ab, 0x00ac, 0x00ad, 0x00ae, 0x00af,
 155   /* 0xb0 */
 156   0x00b0, 0x00b1, 0x00b2, 0x00b3, 0x00b4, 0x00b5, 0x00b6, 0x00b7,
 157   0x00b8, 0x00b9, 0x00f7, 0x00bb, 0x00bc, 0x00bd, 0x00be, 0x00bf,
 158   /* 0xc0 */
 159   0x05b0, 0x05b1, 0x05b2, 0x05b3, 0x05b4, 0x05b5, 0x05b6, 0x05b7,
 160   0x05b8, 0x05b9, 0xfffd, 0x05bb, 0x05bc, 0x05bd, 0x05be, 0x05bf,
 161   /* 0xd0 */
 162   0x05c0, 0x05c1, 0x05c2, 0x05c3, 0x05f0, 0x05f1, 0x05f2, 0x05f3,
 163   0x05f4, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd,
 164   /* 0xe0 */
 165   0x05d0, 0x05d1, 0x05d2, 0x05d3, 0x05d4, 0x05d5, 0x05d6, 0x05d7,
 166   0x05d8, 0x05d9, 0x05da, 0x05db, 0x05dc, 0x05dd, 0x05de, 0x05df,
 167   /* 0xf0 */
 168   0x05e0, 0x05e1, 0x05e2, 0x05e3, 0x05e4, 0x05e5, 0x05e6, 0x05e7,
 169   0x05e8, 0x05e9, 0x05ea, 0xfffd, 0xfffd, 0x200e, 0x200f, 0xfffd,
 170 };
 171
 172 /* CP1255 as a stateless encoding. Suitable for locales, but it has
 173    the drawback that it can produce Unicode strings which are not
 174    in Normalization Form C and therefore not suitable for interchange.
 175    FIXME: It should produce Normalization Form C instead. */
 176
 177 static int
 178 cp1255_mbtowc (conv_t conv, ucs4_t *pwc, const unsigned char *s, int n)
 179 {
 180   unsigned char c = *s;
 181   if (c < 0x80) {
 182     *pwc = (ucs4_t) c;
 183     return 1;
 184   }
 185   else {
 186     unsigned short wc = cp1255_2uni[c-0x80];
 187     if (wc != 0xfffd) {
 188       *pwc = (ucs4_t) wc;
 189       return 1;
 190     }
 191   }
 192   return RET_ILSEQ;
 193 }
 194
 195 static const unsigned char cp1255_page00[88] = {
 196   0xa0, 0xa1, 0xa2, 0xa3, 0x00, 0xa5, 0xa6, 0xa7, /* 0xa0-0xa7 */
 197   0xa8, 0xa9, 0x00, 0xab, 0xac, 0xad, 0xae, 0xaf, /* 0xa8-0xaf */
 198   0xb0, 0xb1, 0xb2, 0xb3, 0xb4, 0xb5, 0xb6, 0xb7, /* 0xb0-0xb7 */
 199   0xb8, 0xb9, 0x00, 0xbb, 0xbc, 0xbd, 0xbe, 0xbf, /* 0xb8-0xbf */
 200   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0xc0-0xc7 */
 201   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0xc8-0xcf */
 202   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0xaa, /* 0xd0-0xd7 */
 203   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0xd8-0xdf */
 204   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0xe0-0xe7 */
 205   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0xe8-0xef */
 206   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0xba, /* 0xf0-0xf7 */
 207 };
 208 static const unsigned char cp1255_page02[32] = {
 209   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x88, 0x00, /* 0xc0-0xc7 */
 210   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0xc8-0xcf */
 211   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0xd0-0xd7 */
 212   0x00, 0x00, 0x00, 0x00, 0x98, 0x00, 0x00, 0x00, /* 0xd8-0xdf */
 213 };
 214 static const unsigned char cp1255_page05[72] = {
 215   0xc0, 0xc1, 0xc2, 0xc3, 0xc4, 0xc5, 0xc6, 0xc7, /* 0xb0-0xb7 */
 216   0xc8, 0xc9, 0x00, 0xcb, 0xcc, 0xcd, 0xce, 0xcf, /* 0xb8-0xbf */
 217   0xd0, 0xd1, 0xd2, 0xd3, 0x00, 0x00, 0x00, 0x00, /* 0xc0-0xc7 */
 218   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0xc8-0xcf */
 219   0xe0, 0xe1, 0xe2, 0xe3, 0xe4, 0xe5, 0xe6, 0xe7, /* 0xd0-0xd7 */
 220   0xe8, 0xe9, 0xea, 0xeb, 0xec, 0xed, 0xee, 0xef, /* 0xd8-0xdf */
 221   0xf0, 0xf1, 0xf2, 0xf3, 0xf4, 0xf5, 0xf6, 0xf7, /* 0xe0-0xe7 */
 222   0xf8, 0xf9, 0xfa, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0xe8-0xef */
 223   0xd4, 0xd5, 0xd6, 0xd7, 0xd8, 0x00, 0x00, 0x00, /* 0xf0-0xf7 */
 224 };
 225 static const unsigned char cp1255_page20[56] = {
 226   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0xfd, 0xfe, /* 0x08-0x0f */
 227   0x00, 0x00, 0x00, 0x96, 0x97, 0x00, 0x00, 0x00, /* 0x10-0x17 */
 228   0x91, 0x92, 0x82, 0x00, 0x93, 0x94, 0x84, 0x00, /* 0x18-0x1f */
 229   0x86, 0x87, 0x95, 0x00, 0x00, 0x00, 0x85, 0x00, /* 0x20-0x27 */
 230   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x28-0x2f */
 231   0x89, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x30-0x37 */
 232   0x00, 0x8b, 0x9b, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x38-0x3f */
 233 };
 234
 235 static int
 236 cp1255_wctomb (conv_t conv, unsigned char *r, ucs4_t wc, int n)
 237 {
 238   unsigned char c = 0;
 239   if (wc < 0x0080) {
 240     *r = wc;
 241     return 1;
 242   }
 243   else if (wc >= 0x00a0 && wc < 0x00f8)
 244     c = cp1255_page00[wc-0x00a0];
 245   else if (wc == 0x0192)
 246     c = 0x83;
 247   else if (wc >= 0x02c0 && wc < 0x02e0)
 248     c = cp1255_page02[wc-0x02c0];
 249   else if (wc >= 0x05b0 && wc < 0x05f8)
 250     c = cp1255_page05[wc-0x05b0];
 251   else if (wc >= 0x2008 && wc < 0x2040)
 252     c = cp1255_page20[wc-0x2008];
 253   else if (wc == 0x20aa)
 254     c = 0xa4;
 255   else if (wc == 0x20ac)
 256     c = 0x80;
 257   else if (wc == 0x2122)
 258     c = 0x99;
 259   if (c != 0) {
 260     *r = c;
 261     return 1;
 262   }
 263   /* Try canonical decomposition. */
 264   {
 265     /* Binary search through cp1255_decomp_table. */
 266     unsigned int i1 = 0;
 267     unsigned int i2 = sizeof(cp1255_decomp_table)/sizeof(cp1255_decomp_table[0])-1;
 268     if (wc >= cp1255_decomp_table[i1].composed
 269         && wc <= cp1255_decomp_table[i2].composed) {
 270       unsigned int i;
 271       for (;;) {
 272         /* Here i2 - i1 > 0. */
 273         i = (i1+i2)>>1;
 274         if (wc == cp1255_decomp_table[i].composed)
 275           break;
 276         if (wc < cp1255_decomp_table[i].composed) {
 277           if (i1 == i)
 278             return RET_ILUNI;
 279           /* Here i1 < i < i2. */
 280           i2 = i;
 281         } else {
 282           /* Here i1 <= i < i2. */
 283           if (i1 != i)
 284             i1 = i;
 285           else {
 286             /* Here i2 - i1 = 1. */
 287             i = i2;
 288             if (wc == cp1255_decomp_table[i].composed)
 289               break;
 290             else
 291               return RET_ILUNI;
 292           }
 293         }
 294       }
 295       /* Found a canonical decomposition. */
 296       wc = cp1255_decomp_table[i].base;
 297       /* wc is one of 0x05d0..0x05d6, 0x05d8..0x05dc, 0x05de, 0x05e0..0x05e1,
 298          0x05e3..0x05e4, 0x05e6..0x05ea, 0x05f2. */
 299       c = cp1255_page05[wc-0x05b0];
 300       if (cp1255_decomp_table[i].comb2 < 0) {
 301         if (n < 2)
 302           return RET_TOOSMALL;
 303         r[0] = c;
 304         r[1] = cp1255_comb_table[cp1255_decomp_table[i].comb1];
 305         return 2;
 306       } else {
 307         if (n < 3)
 308           return RET_TOOSMALL;
 309         r[0] = c;
 310         r[1] = cp1255_comb_table[cp1255_decomp_table[i].comb1];
 311         r[2] = cp1255_comb_table[cp1255_decomp_table[i].comb2];
 312         return 3;
 313       }
 314     }
 315   }
 316   return RET_ILUNI;
 317 }