enc/windows_1253.c

   1 /**********************************************************************
   2   windows_1253.c -  Oniguruma (regular expression library)
   3 **********************************************************************/
   4 /*-
   5  * Copyright (c) 2002-2007  K.Kosako  <sndgk393 AT ybb DOT ne DOT jp>
   6  * All rights reserved.
   7  *
   8  * Redistribution and use in source and binary forms, with or without
   9  * modification, are permitted provided that the following conditions
  10  * are met:
  11  * 1. Redistributions of source code must retain the above copyright
  12  *    notice, this list of conditions and the following disclaimer.
  13  * 2. Redistributions in binary form must reproduce the above copyright
  14  *    notice, this list of conditions and the following disclaimer in the
  15  *    documentation and/or other materials provided with the distribution.
  16  *
  17  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
  18  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
  19  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
  20  * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
  21  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
  22  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
  23  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
  24  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
  25  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  26  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  27  * SUCH DAMAGE.
  28  */
  29
  30 /*
  31  * Name: windows-1253
  32  * MIBenum: 2253
  33  * Link: http://www.iana.org/assignments/character-sets
  34  * Link: http://www.microsoft.com/globaldev/reference/sbcs/1253.mspx
  35  * Link: https://en.wikipedia.org/wiki/Windows-1253
  36  */
  37
  38 #include "regenc.h"
  39
  40 #define ENC_CP1253_TO_LOWER_CASE(c) EncCP1253_ToLowerCaseTable[c]
  41 #define ENC_IS_CP1253_CTYPE(code,ctype) \
  42   ((EncCP1253_CtypeTable[code] & CTYPE_TO_BIT(ctype)) != 0)
  43
  44 static const UChar EncCP1253_ToLowerCaseTable[256] = {
  45   '\000', '\001', '\002', '\003', '\004', '\005', '\006', '\007',
  46   '\010', '\011', '\012', '\013', '\014', '\015', '\016', '\017',
  47   '\020', '\021', '\022', '\023', '\024', '\025', '\026', '\027',
  48   '\030', '\031', '\032', '\033', '\034', '\035', '\036', '\037',
  49   '\040', '\041', '\042', '\043', '\044', '\045', '\046', '\047',
  50   '\050', '\051', '\052', '\053', '\054', '\055', '\056', '\057',
  51   '\060', '\061', '\062', '\063', '\064', '\065', '\066', '\067',
  52   '\070', '\071', '\072', '\073', '\074', '\075', '\076', '\077',
  53   '\100', '\141', '\142', '\143', '\144', '\145', '\146', '\147',
  54   '\150', '\151', '\152', '\153', '\154', '\155', '\156', '\157',
  55   '\160', '\161', '\162', '\163', '\164', '\165', '\166', '\167',
  56   '\170', '\171', '\172', '\133', '\134', '\135', '\136', '\137',
  57   '\140', '\141', '\142', '\143', '\144', '\145', '\146', '\147',
  58   '\150', '\151', '\152', '\153', '\154', '\155', '\156', '\157',
  59   '\160', '\161', '\162', '\163', '\164', '\165', '\166', '\167',
  60   '\170', '\171', '\172', '\173', '\174', '\175', '\176', '\177',
  61   '\200', '\201', '\202', '\203', '\204', '\205', '\206', '\207',
  62   '\210', '\211', '\212', '\213', '\214', '\215', '\216', '\217',
  63   '\220', '\221', '\222', '\223', '\224', '\225', '\226', '\227',
  64   '\230', '\231', '\232', '\233', '\234', '\235', '\236', '\237',
  65   '\240', '\241', '\334', '\243', '\244', '\245', '\246', '\247',
  66   '\250', '\251', '\252', '\253', '\254', '\255', '\256', '\257',
  67   '\260', '\261', '\262', '\263', '\264', '\354', '\334', '\267',
  68   '\335', '\336', '\337', '\273', '\374', '\275', '\375', '\376',
  69   '\300', '\341', '\342', '\343', '\344', '\345', '\346', '\347',
  70   '\350', '\351', '\352', '\353', '\354', '\355', '\356', '\357',
  71   '\360', '\361', '\322', '\363', '\364', '\365', '\366', '\367',
  72   '\370', '\371', '\372', '\373', '\334', '\335', '\336', '\337',
  73   '\340', '\341', '\342', '\343', '\344', '\345', '\346', '\347',
  74   '\350', '\351', '\352', '\353', '\354', '\355', '\356', '\357',
  75   '\360', '\361', '\362', '\363', '\364', '\365', '\366', '\367',
  76   '\370', '\371', '\372', '\373', '\374', '\375', '\376', '\377'
  77 };
  78
  79 static const unsigned short EncCP1253_CtypeTable[256] = {
  80   0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008,
  81   0x4008, 0x420c, 0x4209, 0x4208, 0x4208, 0x4208, 0x4008, 0x4008,
  82   0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008,
  83   0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008,
  84   0x4284, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0,
  85   0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0,
  86   0x78b0, 0x78b0, 0x78b0, 0x78b0, 0x78b0, 0x78b0, 0x78b0, 0x78b0,
  87   0x78b0, 0x78b0, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0,
  88   0x41a0, 0x7ca2, 0x7ca2, 0x7ca2, 0x7ca2, 0x7ca2, 0x7ca2, 0x74a2,
  89   0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2,
  90   0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2,
  91   0x74a2, 0x74a2, 0x74a2, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x51a0,
  92   0x41a0, 0x78e2, 0x78e2, 0x78e2, 0x78e2, 0x78e2, 0x78e2, 0x70e2,
  93   0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2,
  94   0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2,
  95   0x70e2, 0x70e2, 0x70e2, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x4008,
  96   0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008,
  97   0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008,
  98   0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008,
  99   0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008,
 100   0x0284, 0x01a0, 0x34a2, 0x00a0, 0x0000, 0x0000, 0x00a0, 0x00a0,
 101   0x00a0, 0x00a0, 0x0000, 0x01a0, 0x00a0, 0x01a0, 0x0000, 0x01a0,
 102   0x00a0, 0x00a0, 0x10a0, 0x10a0, 0x00a0, 0x30e2, 0x34a2, 0x01a0,
 103   0x34a2, 0x34a2, 0x34a2, 0x01a0, 0x34a2, 0x10a0, 0x34a2, 0x34a2,
 104   0x30e2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2,
 105   0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2,
 106   0x34a2, 0x34a2, 0x0000, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2,
 107   0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x30e2, 0x30e2, 0x30e2, 0x30e2,
 108   0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2,
 109   0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2,
 110   0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2,
 111   0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x0000
 112 };
 113
 114 static int
 115 mbc_case_fold(OnigCaseFoldType flag,
 116               const UChar** pp, const UChar* end ARG_UNUSED, UChar* lower,
 117               OnigEncoding enc ARG_UNUSED)
 118 {
 119   const UChar* p = *pp;
 120
 121   *lower = ENC_CP1253_TO_LOWER_CASE(*p);
 122   (*pp)++;
 123   return 1;
 124 }
 125
 126 #if 0
 127 static int
 128 is_mbc_ambiguous(OnigCaseFoldType flag,
 129                  const UChar** pp, const UChar* end)
 130 {
 131   int v;
 132   const UChar* p = *pp;
 133
 134   (*pp)++;
 135   v = (EncCP1253_CtypeTable[*p] & (BIT_CTYPE_UPPER | BIT_CTYPE_LOWER));
 136   if ((v | BIT_CTYPE_LOWER) != 0) {
 137     if (*p == 0xc0 || *p == 0xe0)
 138       return FALSE;
 139     else
 140       return TRUE;
 141   }
 142
 143   return (v != 0 ? TRUE : FALSE);
 144 }
 145 #endif
 146
 147 static int
 148 is_code_ctype(OnigCodePoint code, unsigned int ctype, OnigEncoding enc ARG_UNUSED)
 149 {
 150   if (code < 256)
 151     return ENC_IS_CP1253_CTYPE(code, ctype);
 152   else
 153     return FALSE;
 154 }
 155
 156 static const OnigPairCaseFoldCodes CaseFoldMap[] = {
 157  { 0xb6, 0xdc },
 158  { 0xb8, 0xdd },
 159  { 0xb9, 0xde },
 160  { 0xba, 0xdf },
 161  { 0xbc, 0xfc },
 162  { 0xbe, 0xfd },
 163  { 0xbf, 0xfe },
 164
 165  { 0xc1, 0xe1 },
 166  { 0xc2, 0xe2 },
 167  { 0xc3, 0xe3 },
 168  { 0xc4, 0xe4 },
 169  { 0xc5, 0xe5 },
 170  { 0xc6, 0xe6 },
 171  { 0xc7, 0xe7 },
 172  { 0xc8, 0xe8 },
 173  { 0xc9, 0xe9 },
 174  { 0xca, 0xea },
 175  { 0xcb, 0xeb },
 176  { 0xcc, 0xec },
 177  { 0xcd, 0xed },
 178  { 0xce, 0xee },
 179  { 0xcf, 0xef },
 180
 181  { 0xd0, 0xf0 },
 182  { 0xd1, 0xf1 },
 183  { 0xd2, 0xf2 },
 184  { 0xd3, 0xf3 },
 185  { 0xd4, 0xf4 },
 186  { 0xd5, 0xf5 },
 187  { 0xd6, 0xf6 },
 188  { 0xd7, 0xf7 },
 189  { 0xd8, 0xf8 },
 190  { 0xd9, 0xf9 },
 191  { 0xda, 0xfa },
 192  { 0xdb, 0xfb }
 193 };
 194
 195 static int
 196 apply_all_case_fold(OnigCaseFoldType flag,
 197                     OnigApplyAllCaseFoldFunc f, void* arg,
 198                     OnigEncoding enc ARG_UNUSED)
 199 {
 200   return onigenc_apply_all_case_fold_with_map(
 201              numberof(CaseFoldMap), CaseFoldMap, 0,
 202              flag, f, arg);
 203 }
 204
 205 static int
 206 get_case_fold_codes_by_str(OnigCaseFoldType flag,
 207                            const OnigUChar* p, const OnigUChar* end,
 208                            OnigCaseFoldCodeItem items[],
 209                            OnigEncoding enc ARG_UNUSED)
 210 {
 211   return onigenc_get_case_fold_codes_by_str_with_map(
 212              numberof(CaseFoldMap), CaseFoldMap, 0,
 213              flag, p, end, items);
 214 }
 215
 216 static int
 217 case_map(OnigCaseFoldType* flagP, const OnigUChar** pp,
 218          const OnigUChar* end, OnigUChar* to, OnigUChar* to_end,
 219          const struct OnigEncodingTypeST* enc)
 220 {
 221   OnigCodePoint code;
 222   OnigUChar *to_start = to;
 223   OnigCaseFoldType flags = *flagP;
 224
 225   while (*pp < end && to < to_end) {
 226     code = *(*pp)++;
 227     if (code == 0xF2) {
 228       if (flags & ONIGENC_CASE_UPCASE) {
 229         flags |= ONIGENC_CASE_MODIFIED;
 230         code = 0xD3;
 231       }
 232       else if (flags & ONIGENC_CASE_FOLD) {
 233         flags |= ONIGENC_CASE_MODIFIED;
 234         code = 0xF3;
 235       }
 236     }
 237     else if (code == 0xB5) {
 238       if (flags & ONIGENC_CASE_UPCASE) {
 239         flags |= ONIGENC_CASE_MODIFIED;
 240         code = 0xCC;
 241       }
 242       else if (flags & ONIGENC_CASE_FOLD) {
 243         flags |= ONIGENC_CASE_MODIFIED;
 244         code = 0xEC;
 245       }
 246     }
 247     else if (code == 0xC0 || code == 0xE0 || code == 0xB6)
 248       ;
 249     else if ((EncCP1253_CtypeTable[code] & BIT_CTYPE_UPPER)
 250              && (flags & (ONIGENC_CASE_DOWNCASE | ONIGENC_CASE_FOLD))) {
 251       flags |= ONIGENC_CASE_MODIFIED;
 252       code = ENC_CP1253_TO_LOWER_CASE(code);
 253     }
 254     else if ((EncCP1253_CtypeTable[code] & BIT_CTYPE_LOWER)
 255              && (flags & ONIGENC_CASE_UPCASE)) {
 256       flags |= ONIGENC_CASE_MODIFIED;
 257       if (code == 0xDC)
 258         code = 0xA2;
 259       else if (code >= 0xDD && code <= 0xDF)
 260         code -= 0x25;
 261       else if (code == 0xFC)
 262         code = 0xBC;
 263       else if (code == 0xFD || code == 0xFE)
 264         code -= 0x3F;
 265       else
 266         code -= 0x20;
 267     }
 268     *to++ = code;
 269     if (flags & ONIGENC_CASE_TITLECASE)  /* switch from titlecase to lowercase for capitalize */
 270       flags ^= (ONIGENC_CASE_UPCASE | ONIGENC_CASE_DOWNCASE | ONIGENC_CASE_TITLECASE);
 271   }
 272   *flagP = flags;
 273   return (int )(to - to_start);
 274 }
 275
 276 OnigEncodingDefine(windows_1253, Windows_1253) = {
 277   onigenc_single_byte_mbc_enc_len,
 278   "Windows-1253",  /* name */
 279   1,             /* max enc length */
 280   1,             /* min enc length */
 281   onigenc_is_mbc_newline_0x0a,
 282   onigenc_single_byte_mbc_to_code,
 283   onigenc_single_byte_code_to_mbclen,
 284   onigenc_single_byte_code_to_mbc,
 285   mbc_case_fold,
 286   apply_all_case_fold,
 287   get_case_fold_codes_by_str,
 288   onigenc_minimum_property_name_to_ctype,
 289   is_code_ctype,
 290   onigenc_not_support_get_ctype_code_range,
 291   onigenc_single_byte_left_adjust_char_head,
 292   onigenc_always_true_is_allowed_reverse_match,
 293   case_map,
 294   0,
 295   ONIGENC_FLAG_NONE,
 296 };
 297 ENC_ALIAS("CP1253", "Windows-1253")