enc/iso_8859_1.c

   1 /**********************************************************************
   2   iso8859_1.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 #include "regenc.h"
  31 #include "iso_8859.h"
  32
  33 #define ENC_IS_ISO_8859_1_CTYPE(code,ctype) \
  34   ((EncISO_8859_1_CtypeTable[code] & CTYPE_TO_BIT(ctype)) != 0)
  35
  36 static const unsigned short EncISO_8859_1_CtypeTable[256] = {
  37   0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008,
  38   0x4008, 0x420c, 0x4209, 0x4208, 0x4208, 0x4208, 0x4008, 0x4008,
  39   0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008,
  40   0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008,
  41   0x4284, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0,
  42   0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0,
  43   0x78b0, 0x78b0, 0x78b0, 0x78b0, 0x78b0, 0x78b0, 0x78b0, 0x78b0,
  44   0x78b0, 0x78b0, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0,
  45   0x41a0, 0x7ca2, 0x7ca2, 0x7ca2, 0x7ca2, 0x7ca2, 0x7ca2, 0x74a2,
  46   0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2,
  47   0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2,
  48   0x74a2, 0x74a2, 0x74a2, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x51a0,
  49   0x41a0, 0x78e2, 0x78e2, 0x78e2, 0x78e2, 0x78e2, 0x78e2, 0x70e2,
  50   0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2,
  51   0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2,
  52   0x70e2, 0x70e2, 0x70e2, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x4008,
  53   0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008,
  54   0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008,
  55   0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008,
  56   0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008,
  57   0x0284, 0x01a0, 0x00a0, 0x00a0, 0x00a0, 0x00a0, 0x00a0, 0x00a0,
  58   0x00a0, 0x00a0, 0x30e2, 0x01a0, 0x00a0, 0x01a0, 0x00a0, 0x00a0,
  59   0x00a0, 0x00a0, 0x10a0, 0x10a0, 0x00a0, 0x30e2, 0x00a0, 0x01a0,
  60   0x00a0, 0x10a0, 0x30e2, 0x01a0, 0x10a0, 0x10a0, 0x10a0, 0x01a0,
  61   0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2,
  62   0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2,
  63   0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x00a0,
  64   0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x30e2,
  65   0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2,
  66   0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2,
  67   0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x00a0,
  68   0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2
  69 };
  70
  71 static const OnigPairCaseFoldCodes CaseFoldMap[] = {
  72   { 0xc0, 0xe0 },
  73   { 0xc1, 0xe1 },
  74   { 0xc2, 0xe2 },
  75   { 0xc3, 0xe3 },
  76   { 0xc4, 0xe4 },
  77   { 0xc5, 0xe5 },
  78   { 0xc6, 0xe6 },
  79   { 0xc7, 0xe7 },
  80   { 0xc8, 0xe8 },
  81   { 0xc9, 0xe9 },
  82   { 0xca, 0xea },
  83   { 0xcb, 0xeb },
  84   { 0xcc, 0xec },
  85   { 0xcd, 0xed },
  86   { 0xce, 0xee },
  87   { 0xcf, 0xef },
  88
  89   { 0xd0, 0xf0 },
  90   { 0xd1, 0xf1 },
  91   { 0xd2, 0xf2 },
  92   { 0xd3, 0xf3 },
  93   { 0xd4, 0xf4 },
  94   { 0xd5, 0xf5 },
  95   { 0xd6, 0xf6 },
  96   { 0xd8, 0xf8 },
  97   { 0xd9, 0xf9 },
  98   { 0xda, 0xfa },
  99   { 0xdb, 0xfb },
 100   { 0xdc, 0xfc },
 101   { 0xdd, 0xfd },
 102   { 0xde, 0xfe }
 103 };
 104
 105 static int
 106 apply_all_case_fold(OnigCaseFoldType flag,
 107                     OnigApplyAllCaseFoldFunc f, void* arg,
 108                     OnigEncoding enc ARG_UNUSED)
 109 {
 110   return onigenc_apply_all_case_fold_with_map(
 111             numberof(CaseFoldMap), CaseFoldMap, 1,
 112             flag, f, arg);
 113 }
 114
 115 static int
 116 get_case_fold_codes_by_str(OnigCaseFoldType flag,
 117                            const OnigUChar* p, const OnigUChar* end,
 118                            OnigCaseFoldCodeItem items[],
 119                            OnigEncoding enc ARG_UNUSED)
 120 {
 121   if (0x41 <= *p && *p <= 0x5a) {
 122     items[0].byte_len = 1;
 123     items[0].code_len = 1;
 124     items[0].code[0] = (OnigCodePoint )(*p + 0x20);
 125     if (*p == 0x53 && end > p + 1
 126         && (*(p+1) == 0x53 || *(p+1) == 0x73)) { /* SS */
 127       items[1].byte_len = 2;
 128       items[1].code_len = 1;
 129       items[1].code[0] = (OnigCodePoint )SHARP_s;
 130       return 2;
 131     }
 132     else
 133       return 1;
 134   }
 135   else if (0x61 <= *p && *p <= 0x7a) {
 136     items[0].byte_len = 1;
 137     items[0].code_len = 1;
 138     items[0].code[0] = (OnigCodePoint )(*p - 0x20);
 139     if (*p == 0x73 && end > p + 1
 140         && (*(p+1) == 0x73 || *(p+1) == 0x53)) { /* ss */
 141       items[1].byte_len = 2;
 142       items[1].code_len = 1;
 143       items[1].code[0] = (OnigCodePoint )SHARP_s;
 144       return 2;
 145     }
 146     else
 147       return 1;
 148   }
 149   else if (0xc0 <= *p && *p <= 0xcf) {
 150     items[0].byte_len = 1;
 151     items[0].code_len = 1;
 152     items[0].code[0] = (OnigCodePoint )(*p + 0x20);
 153     return 1;
 154   }
 155   else if (0xd0 <= *p && *p <= 0xdf) {
 156     if (*p == SHARP_s) {
 157       items[0].byte_len = 1;
 158       items[0].code_len = 2;
 159       items[0].code[0] = (OnigCodePoint )'s';
 160       items[0].code[1] = (OnigCodePoint )'s';
 161
 162       items[1].byte_len = 1;
 163       items[1].code_len = 2;
 164       items[1].code[0] = (OnigCodePoint )'S';
 165       items[1].code[1] = (OnigCodePoint )'S';
 166
 167       items[2].byte_len = 1;
 168       items[2].code_len = 2;
 169       items[2].code[0] = (OnigCodePoint )'s';
 170       items[2].code[1] = (OnigCodePoint )'S';
 171
 172       items[3].byte_len = 1;
 173       items[3].code_len = 2;
 174       items[3].code[0] = (OnigCodePoint )'S';
 175       items[3].code[1] = (OnigCodePoint )'s';
 176
 177       return 4;
 178     }
 179     else if (*p != 0xd7) {
 180       items[0].byte_len = 1;
 181       items[0].code_len = 1;
 182       items[0].code[0] = (OnigCodePoint )(*p + 0x20);
 183       return 1;
 184     }
 185   }
 186   else if (0xe0 <= *p && *p <= 0xef) {
 187     items[0].byte_len = 1;
 188     items[0].code_len = 1;
 189     items[0].code[0] = (OnigCodePoint )(*p - 0x20);
 190     return 1;
 191   }
 192   else if (0xf0 <= *p && *p <= 0xfe) {
 193     if (*p != 0xf7) {
 194       items[0].byte_len = 1;
 195       items[0].code_len = 1;
 196       items[0].code[0] = (OnigCodePoint )(*p - 0x20);
 197       return 1;
 198     }
 199   }
 200
 201   return 0;
 202 }
 203
 204 static int
 205 mbc_case_fold(OnigCaseFoldType flag, const UChar** pp, const UChar* end ARG_UNUSED,
 206               UChar* lower, OnigEncoding enc ARG_UNUSED)
 207 {
 208   const UChar* p = *pp;
 209
 210   if (*p == SHARP_s && (flag & INTERNAL_ONIGENC_CASE_FOLD_MULTI_CHAR) != 0) {
 211     *lower++ = 's';
 212     *lower   = 's';
 213     (*pp)++;
 214     return 2;
 215   }
 216
 217   *lower = ONIGENC_ISO_8859_1_TO_LOWER_CASE(*p);
 218   (*pp)++;
 219   return 1;
 220 }
 221
 222 #if 0
 223 static int
 224 is_mbc_ambiguous(OnigCaseFoldType flag,
 225                  const UChar** pp, const UChar* end)
 226 {
 227   int v;
 228   const UChar* p = *pp;
 229
 230   if (*p == SHARP_s && (flag & INTERNAL_ONIGENC_CASE_FOLD_MULTI_CHAR) != 0) {
 231     (*pp)++;
 232     return TRUE;
 233   }
 234
 235   (*pp)++;
 236   v = (EncISO_8859_1_CtypeTable[*p] & (BIT_CTYPE_UPPER | BIT_CTYPE_LOWER));
 237   if ((v | BIT_CTYPE_LOWER) != 0) {
 238     /* 0xdf, 0xaa, 0xb5, 0xba are lower case letter, but can't convert. */
 239     if (*p >= 0xaa && *p <= 0xba)
 240       return FALSE;
 241     else
 242       return TRUE;
 243   }
 244
 245   return (v != 0 ? TRUE : FALSE);
 246 }
 247 #endif
 248
 249 static int
 250 is_code_ctype(OnigCodePoint code, unsigned int ctype, OnigEncoding enc ARG_UNUSED)
 251 {
 252   if (code < 256)
 253     return ENC_IS_ISO_8859_1_CTYPE(code, ctype);
 254   else
 255     return FALSE;
 256 }
 257
 258 static int
 259 case_map(OnigCaseFoldType* flagP, const OnigUChar** pp,
 260          const OnigUChar* end, OnigUChar* to, OnigUChar* to_end,
 261          const struct OnigEncodingTypeST* enc)
 262 {
 263   OnigCodePoint code;
 264   OnigUChar *to_start = to;
 265   OnigCaseFoldType flags = *flagP;
 266
 267   while (*pp < end && to < to_end) {
 268     code = *(*pp)++;
 269     if (code == SHARP_s) {
 270       if (flags & ONIGENC_CASE_UPCASE) {
 271         flags |= ONIGENC_CASE_MODIFIED;
 272         *to++ = 'S';
 273         code = (flags & ONIGENC_CASE_TITLECASE) ? 's' : 'S';
 274       }
 275       else if (flags & ONIGENC_CASE_FOLD) {
 276         flags |= ONIGENC_CASE_MODIFIED;
 277         *to++ = 's';
 278         code = 's';
 279       }
 280     }
 281     else if ((EncISO_8859_1_CtypeTable[code] & BIT_CTYPE_UPPER)
 282              && (flags & (ONIGENC_CASE_DOWNCASE | ONIGENC_CASE_FOLD))) {
 283       flags |= ONIGENC_CASE_MODIFIED;
 284       code += 0x20;
 285     }
 286     else if (code == 0xAA || code == 0xBA || code == 0xB5 || code == 0xFF)
 287       ;
 288     else if ((EncISO_8859_1_CtypeTable[code] & BIT_CTYPE_LOWER)
 289              && (flags & ONIGENC_CASE_UPCASE)) {
 290       flags |= ONIGENC_CASE_MODIFIED;
 291       code -= 0x20;
 292     }
 293     *to++ = code;
 294     if (flags & ONIGENC_CASE_TITLECASE)  /* switch from titlecase to lowercase for capitalize */
 295       flags ^= (ONIGENC_CASE_UPCASE | ONIGENC_CASE_DOWNCASE | ONIGENC_CASE_TITLECASE);
 296   }
 297   *flagP = flags;
 298   return (int )(to - to_start);
 299 }
 300
 301 OnigEncodingDefine(iso_8859_1, ISO_8859_1) = {
 302   onigenc_single_byte_mbc_enc_len,
 303   "ISO-8859-1",  /* name */
 304   1,             /* max enc length */
 305   1,             /* min enc length */
 306   onigenc_is_mbc_newline_0x0a,
 307   onigenc_single_byte_mbc_to_code,
 308   onigenc_single_byte_code_to_mbclen,
 309   onigenc_single_byte_code_to_mbc,
 310   mbc_case_fold,
 311   apply_all_case_fold,
 312   get_case_fold_codes_by_str,
 313   onigenc_minimum_property_name_to_ctype,
 314   is_code_ctype,
 315   onigenc_not_support_get_ctype_code_range,
 316   onigenc_single_byte_left_adjust_char_head,
 317   onigenc_always_true_is_allowed_reverse_match,
 318   case_map,
 319   0,
 320   ONIGENC_FLAG_NONE,
 321 };
 322 ENC_ALIAS("ISO8859-1", "ISO-8859-1")