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
  32 #define ENC_IS_ISO_8859_1_CTYPE(code,ctype) \
  33   ((EncISO_8859_1_CtypeTable[code] & CTYPE_TO_BIT(ctype)) != 0)
  34
  35 static const unsigned short EncISO_8859_1_CtypeTable[256] = {
  36   0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008,
  37   0x4008, 0x420c, 0x4209, 0x4208, 0x4208, 0x4208, 0x4008, 0x4008,
  38   0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008,
  39   0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008,
  40   0x4284, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0,
  41   0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0,
  42   0x78b0, 0x78b0, 0x78b0, 0x78b0, 0x78b0, 0x78b0, 0x78b0, 0x78b0,
  43   0x78b0, 0x78b0, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0,
  44   0x41a0, 0x7ca2, 0x7ca2, 0x7ca2, 0x7ca2, 0x7ca2, 0x7ca2, 0x74a2,
  45   0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2,
  46   0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2,
  47   0x74a2, 0x74a2, 0x74a2, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x51a0,
  48   0x41a0, 0x78e2, 0x78e2, 0x78e2, 0x78e2, 0x78e2, 0x78e2, 0x70e2,
  49   0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2,
  50   0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2,
  51   0x70e2, 0x70e2, 0x70e2, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x4008,
  52   0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008,
  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   0x0284, 0x01a0, 0x00a0, 0x00a0, 0x00a0, 0x00a0, 0x00a0, 0x00a0,
  57   0x00a0, 0x00a0, 0x30e2, 0x01a0, 0x00a0, 0x01a0, 0x00a0, 0x00a0,
  58   0x00a0, 0x00a0, 0x10a0, 0x10a0, 0x00a0, 0x30e2, 0x00a0, 0x01a0,
  59   0x00a0, 0x10a0, 0x30e2, 0x01a0, 0x10a0, 0x10a0, 0x10a0, 0x01a0,
  60   0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2,
  61   0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2,
  62   0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x00a0,
  63   0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x30e2,
  64   0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2,
  65   0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2,
  66   0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x00a0,
  67   0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2
  68 };
  69
  70 static const OnigPairCaseFoldCodes CaseFoldMap[] = {
  71   { 0xc0, 0xe0 },
  72   { 0xc1, 0xe1 },
  73   { 0xc2, 0xe2 },
  74   { 0xc3, 0xe3 },
  75   { 0xc4, 0xe4 },
  76   { 0xc5, 0xe5 },
  77   { 0xc6, 0xe6 },
  78   { 0xc7, 0xe7 },
  79   { 0xc8, 0xe8 },
  80   { 0xc9, 0xe9 },
  81   { 0xca, 0xea },
  82   { 0xcb, 0xeb },
  83   { 0xcc, 0xec },
  84   { 0xcd, 0xed },
  85   { 0xce, 0xee },
  86   { 0xcf, 0xef },
  87
  88   { 0xd0, 0xf0 },
  89   { 0xd1, 0xf1 },
  90   { 0xd2, 0xf2 },
  91   { 0xd3, 0xf3 },
  92   { 0xd4, 0xf4 },
  93   { 0xd5, 0xf5 },
  94   { 0xd6, 0xf6 },
  95   { 0xd8, 0xf8 },
  96   { 0xd9, 0xf9 },
  97   { 0xda, 0xfa },
  98   { 0xdb, 0xfb },
  99   { 0xdc, 0xfc },
 100   { 0xdd, 0xfd },
 101   { 0xde, 0xfe }
 102 };
 103
 104 static int
 105 apply_all_case_fold(OnigCaseFoldType flag,
 106                     OnigApplyAllCaseFoldFunc f, void* arg,
 107                     OnigEncoding enc ARG_UNUSED)
 108 {
 109   return onigenc_apply_all_case_fold_with_map(
 110             sizeof(CaseFoldMap)/sizeof(OnigPairCaseFoldCodes), CaseFoldMap, 1,
 111             flag, f, arg);
 112 }
 113
 114 static int
 115 get_case_fold_codes_by_str(OnigCaseFoldType flag,
 116                            const OnigUChar* p, const OnigUChar* end,
 117                            OnigCaseFoldCodeItem items[],
 118                            OnigEncoding enc ARG_UNUSED)
 119 {
 120   if (0x41 <= *p && *p <= 0x5a) {
 121     items[0].byte_len = 1;
 122     items[0].code_len = 1;
 123     items[0].code[0] = (OnigCodePoint )(*p + 0x20);
 124     if (*p == 0x53 && end > p + 1
 125         && (*(p+1) == 0x53 || *(p+1) == 0x73)) { /* SS */
 126       items[1].byte_len = 2;
 127       items[1].code_len = 1;
 128       items[1].code[0] = (OnigCodePoint )0xdf;
 129       return 2;
 130     }
 131     else
 132       return 1;
 133   }
 134   else if (0x61 <= *p && *p <= 0x7a) {
 135     items[0].byte_len = 1;
 136     items[0].code_len = 1;
 137     items[0].code[0] = (OnigCodePoint )(*p - 0x20);
 138     if (*p == 0x73 && end > p + 1
 139         && (*(p+1) == 0x73 || *(p+1) == 0x53)) { /* ss */
 140       items[1].byte_len = 2;
 141       items[1].code_len = 1;
 142       items[1].code[0] = (OnigCodePoint )0xdf;
 143       return 2;
 144     }
 145     else
 146       return 1;
 147   }
 148   else if (0xc0 <= *p && *p <= 0xcf) {
 149     items[0].byte_len = 1;
 150     items[0].code_len = 1;
 151     items[0].code[0] = (OnigCodePoint )(*p + 0x20);
 152     return 1;
 153   }
 154   else if (0xd0 <= *p && *p <= 0xdf) {
 155     if (*p == 0xdf) {
 156       items[0].byte_len = 1;
 157       items[0].code_len = 2;
 158       items[0].code[0] = (OnigCodePoint )'s';
 159       items[0].code[1] = (OnigCodePoint )'s';
 160
 161       items[1].byte_len = 1;
 162       items[1].code_len = 2;
 163       items[1].code[0] = (OnigCodePoint )'S';
 164       items[1].code[1] = (OnigCodePoint )'S';
 165
 166       items[2].byte_len = 1;
 167       items[2].code_len = 2;
 168       items[2].code[0] = (OnigCodePoint )'s';
 169       items[2].code[1] = (OnigCodePoint )'S';
 170
 171       items[3].byte_len = 1;
 172       items[3].code_len = 2;
 173       items[3].code[0] = (OnigCodePoint )'S';
 174       items[3].code[1] = (OnigCodePoint )'s';
 175
 176       return 4;
 177     }
 178     else if (*p != 0xd7) {
 179       items[0].byte_len = 1;
 180       items[0].code_len = 1;
 181       items[0].code[0] = (OnigCodePoint )(*p + 0x20);
 182       return 1;
 183     }
 184   }
 185   else if (0xe0 <= *p && *p <= 0xef) {
 186     items[0].byte_len = 1;
 187     items[0].code_len = 1;
 188     items[0].code[0] = (OnigCodePoint )(*p - 0x20);
 189     return 1;
 190   }
 191   else if (0xf0 <= *p && *p <= 0xfe) {
 192     if (*p != 0xf7) {
 193       items[0].byte_len = 1;
 194       items[0].code_len = 1;
 195       items[0].code[0] = (OnigCodePoint )(*p - 0x20);
 196       return 1;
 197     }
 198   }
 199
 200   return 0;
 201 }
 202
 203 static int
 204 mbc_case_fold(OnigCaseFoldType flag, const UChar** pp, const UChar* end ARG_UNUSED,
 205               UChar* lower, OnigEncoding enc ARG_UNUSED)
 206 {
 207   const UChar* p = *pp;
 208
 209   if (*p == 0xdf && (flag & INTERNAL_ONIGENC_CASE_FOLD_MULTI_CHAR) != 0) {
 210     *lower++ = 's';
 211     *lower   = 's';
 212     (*pp)++;
 213     return 2;
 214   }
 215
 216   *lower = ONIGENC_ISO_8859_1_TO_LOWER_CASE(*p);
 217   (*pp)++;
 218   return 1;
 219 }
 220
 221 #if 0
 222 static int
 223 is_mbc_ambiguous(OnigCaseFoldType flag,
 224                  const UChar** pp, const UChar* end)
 225 {
 226   int v;
 227   const UChar* p = *pp;
 228
 229   if (*p == 0xdf && (flag & INTERNAL_ONIGENC_CASE_FOLD_MULTI_CHAR) != 0) {
 230     (*pp)++;
 231     return TRUE;
 232   }
 233
 234   (*pp)++;
 235   v = (EncISO_8859_1_CtypeTable[*p] & (BIT_CTYPE_UPPER | BIT_CTYPE_LOWER));
 236   if ((v | BIT_CTYPE_LOWER) != 0) {
 237     /* 0xdf, 0xaa, 0xb5, 0xba are lower case letter, but can't convert. */
 238     if (*p >= 0xaa && *p <= 0xba)
 239       return FALSE;
 240     else
 241       return TRUE;
 242   }
 243
 244   return (v != 0 ? TRUE : FALSE);
 245 }
 246 #endif
 247
 248 static int
 249 is_code_ctype(OnigCodePoint code, unsigned int ctype, OnigEncoding enc ARG_UNUSED)
 250 {
 251   if (code < 256)
 252     return ENC_IS_ISO_8859_1_CTYPE(code, ctype);
 253   else
 254     return FALSE;
 255 }
 256
 257 OnigEncodingDefine(iso_8859_1, ISO_8859_1) = {
 258   onigenc_single_byte_mbc_enc_len,
 259   "ISO-8859-1",  /* name */
 260   1,             /* max enc length */
 261   1,             /* min enc length */
 262   onigenc_is_mbc_newline_0x0a,
 263   onigenc_single_byte_mbc_to_code,
 264   onigenc_single_byte_code_to_mbclen,
 265   onigenc_single_byte_code_to_mbc,
 266   mbc_case_fold,
 267   apply_all_case_fold,
 268   get_case_fold_codes_by_str,
 269   onigenc_minimum_property_name_to_ctype,
 270   is_code_ctype,
 271   onigenc_not_support_get_ctype_code_range,
 272   onigenc_single_byte_left_adjust_char_head,
 273   onigenc_always_true_is_allowed_reverse_match
 274 };
 275 ENC_ALIAS("ISO8859-1", "ISO-8859-1")
 276
 277 /*
 278  * Name: windows-1252
 279  * MIBenum: 2252
 280  * Link: http://www.iana.org/assignments/character-sets
 281  * Link: http://www.microsoft.com/globaldev/reference/sbcs/1252.mspx
 282  * Link: http://en.wikipedia.org/wiki/Windows-1252
 283  */
 284 ENC_REPLICATE("Windows-1252", "ISO-8859-1")
 285 ENC_ALIAS("CP1252", "Windows-1252")