enc/iso_8859_9.c

   1 /**********************************************************************
   2   iso8859_9.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_ISO_8859_9_TO_LOWER_CASE(c) EncISO_8859_9_ToLowerCaseTable[c]
  34 #define ENC_IS_ISO_8859_9_CTYPE(code,ctype) \
  35   ((EncISO_8859_9_CtypeTable[code] & CTYPE_TO_BIT(ctype)) != 0)
  36
  37 static const UChar EncISO_8859_9_ToLowerCaseTable[256] = {
  38   '\000', '\001', '\002', '\003', '\004', '\005', '\006', '\007',
  39   '\010', '\011', '\012', '\013', '\014', '\015', '\016', '\017',
  40   '\020', '\021', '\022', '\023', '\024', '\025', '\026', '\027',
  41   '\030', '\031', '\032', '\033', '\034', '\035', '\036', '\037',
  42   '\040', '\041', '\042', '\043', '\044', '\045', '\046', '\047',
  43   '\050', '\051', '\052', '\053', '\054', '\055', '\056', '\057',
  44   '\060', '\061', '\062', '\063', '\064', '\065', '\066', '\067',
  45   '\070', '\071', '\072', '\073', '\074', '\075', '\076', '\077',
  46   '\100', '\141', '\142', '\143', '\144', '\145', '\146', '\147',
  47   '\150', '\151', '\152', '\153', '\154', '\155', '\156', '\157',
  48   '\160', '\161', '\162', '\163', '\164', '\165', '\166', '\167',
  49   '\170', '\171', '\172', '\133', '\134', '\135', '\136', '\137',
  50   '\140', '\141', '\142', '\143', '\144', '\145', '\146', '\147',
  51   '\150', '\151', '\152', '\153', '\154', '\155', '\156', '\157',
  52   '\160', '\161', '\162', '\163', '\164', '\165', '\166', '\167',
  53   '\170', '\171', '\172', '\173', '\174', '\175', '\176', '\177',
  54   '\200', '\201', '\202', '\203', '\204', '\205', '\206', '\207',
  55   '\210', '\211', '\212', '\213', '\214', '\215', '\216', '\217',
  56   '\220', '\221', '\222', '\223', '\224', '\225', '\226', '\227',
  57   '\230', '\231', '\232', '\233', '\234', '\235', '\236', '\237',
  58   '\240', '\241', '\242', '\243', '\244', '\245', '\246', '\247',
  59   '\250', '\251', '\252', '\253', '\254', '\255', '\256', '\257',
  60   '\260', '\261', '\262', '\263', '\264', '\265', '\266', '\267',
  61   '\270', '\271', '\272', '\273', '\274', '\275', '\276', '\277',
  62   '\340', '\341', '\342', '\343', '\344', '\345', '\346', '\347',
  63   '\350', '\351', '\352', '\353', '\354', '\355', '\356', '\357',
  64   '\360', '\361', '\362', '\363', '\364', '\365', '\366', '\327',
  65   '\370', '\371', '\372', '\373', '\374', '\151', '\376', '\337',
  66   '\340', '\341', '\342', '\343', '\344', '\345', '\346', '\347',
  67   '\350', '\351', '\352', '\353', '\354', '\355', '\356', '\357',
  68   '\360', '\361', '\362', '\363', '\364', '\365', '\366', '\367',
  69   '\370', '\371', '\372', '\373', '\374', '\375', '\376', '\377'
  70 };
  71
  72 static const unsigned short EncISO_8859_9_CtypeTable[256] = {
  73   0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008,
  74   0x4008, 0x420c, 0x4209, 0x4208, 0x4208, 0x4208, 0x4008, 0x4008,
  75   0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008,
  76   0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008,
  77   0x4284, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0,
  78   0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0,
  79   0x78b0, 0x78b0, 0x78b0, 0x78b0, 0x78b0, 0x78b0, 0x78b0, 0x78b0,
  80   0x78b0, 0x78b0, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0,
  81   0x41a0, 0x7ca2, 0x7ca2, 0x7ca2, 0x7ca2, 0x7ca2, 0x7ca2, 0x74a2,
  82   0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2,
  83   0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2,
  84   0x74a2, 0x74a2, 0x74a2, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x51a0,
  85   0x41a0, 0x78e2, 0x78e2, 0x78e2, 0x78e2, 0x78e2, 0x78e2, 0x70e2,
  86   0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2,
  87   0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2,
  88   0x70e2, 0x70e2, 0x70e2, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x4008,
  89   0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008,
  90   0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008,
  91   0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008,
  92   0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008,
  93   0x0284, 0x01a0, 0x00a0, 0x00a0, 0x00a0, 0x00a0, 0x00a0, 0x00a0,
  94   0x00a0, 0x00a0, 0x30e2, 0x01a0, 0x00a0, 0x01a0, 0x00a0, 0x00a0,
  95   0x00a0, 0x00a0, 0x10a0, 0x10a0, 0x00a0, 0x30e2, 0x00a0, 0x01a0,
  96   0x00a0, 0x10a0, 0x30e2, 0x01a0, 0x10a0, 0x10a0, 0x10a0, 0x01a0,
  97   0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2,
  98   0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2,
  99   0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x00a0,
 100   0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x30e2,
 101   0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2,
 102   0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2,
 103   0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x00a0,
 104   0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2
 105 };
 106
 107 static int
 108 mbc_case_fold(OnigCaseFoldType flag,
 109               const UChar** pp, const UChar* end ARG_UNUSED, UChar* lower,
 110               OnigEncoding enc ARG_UNUSED)
 111 {
 112   const UChar* p = *pp;
 113
 114   if (*p == SHARP_s && (flag & INTERNAL_ONIGENC_CASE_FOLD_MULTI_CHAR) != 0) {
 115     *lower++ = 's';
 116     *lower   = 's';
 117     (*pp)++;
 118     return 2;
 119   }
 120
 121   *lower = ENC_ISO_8859_9_TO_LOWER_CASE(*p);
 122   (*pp)++;
 123   return 1;
 124 }
 125
 126 #if 0
 127 static int
 128 is_mbc_ambiguous(OnigCaseFoldType flag, const UChar** pp, const UChar* end)
 129 {
 130   int v;
 131   const UChar* p = *pp;
 132
 133   if (*p == SHARP_s && (flag & INTERNAL_ONIGENC_CASE_FOLD_MULTI_CHAR) != 0) {
 134     (*pp)++;
 135     return TRUE;
 136   }
 137
 138   (*pp)++;
 139   v = (EncISO_8859_9_CtypeTable[*p] & (BIT_CTYPE_UPPER | BIT_CTYPE_LOWER));
 140   if ((v | BIT_CTYPE_LOWER) != 0) {
 141     /* 0xdf etc..  are lower case letter, but can't convert. */
 142     if (*p >= 0xaa && *p <= 0xba)
 143       return FALSE;
 144     else
 145       return TRUE;
 146   }
 147
 148   return (v != 0 ? TRUE : FALSE);
 149 }
 150 #endif
 151
 152 static int
 153 is_code_ctype(OnigCodePoint code, unsigned int ctype, OnigEncoding enc ARG_UNUSED)
 154 {
 155   if (code < 256)
 156     return ENC_IS_ISO_8859_9_CTYPE(code, ctype);
 157   else
 158     return FALSE;
 159 }
 160
 161 static const OnigPairCaseFoldCodes CaseFoldMap[] = {
 162   { 0xc0, 0xe0 },
 163   { 0xc1, 0xe1 },
 164   { 0xc2, 0xe2 },
 165   { 0xc3, 0xe3 },
 166   { 0xc4, 0xe4 },
 167   { 0xc5, 0xe5 },
 168   { 0xc6, 0xe6 },
 169   { 0xc7, 0xe7 },
 170   { 0xc8, 0xe8 },
 171   { 0xc9, 0xe9 },
 172   { 0xca, 0xea },
 173   { 0xcb, 0xeb },
 174   { 0xcc, 0xec },
 175   { 0xcd, 0xed },
 176   { 0xce, 0xee },
 177   { 0xcf, 0xef },
 178
 179   { 0xd0, 0xf0 },
 180   { 0xd1, 0xf1 },
 181   { 0xd2, 0xf2 },
 182   { 0xd3, 0xf3 },
 183   { 0xd4, 0xf4 },
 184   { 0xd5, 0xf5 },
 185   { 0xd6, 0xf6 },
 186   { 0xd8, 0xf8 },
 187   { 0xd9, 0xf9 },
 188   { 0xda, 0xfa },
 189   { 0xdb, 0xfb },
 190   { 0xdc, 0xfc },
 191   /*{ 0xdd, 0xfd }, exclude dotless i/I with dot; not a case pair */
 192   { 0xde, 0xfe }
 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, 1,
 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, 1,
 213              flag, p, end, items);
 214 }
 215
 216 #define DOTLESS_i        (0xFD)
 217 #define I_WITH_DOT_ABOVE (0xDD)
 218 static int
 219 case_map(OnigCaseFoldType* flagP, const OnigUChar** pp,
 220          const OnigUChar* end, OnigUChar* to, OnigUChar* to_end,
 221          const struct OnigEncodingTypeST* enc)
 222 {
 223   OnigCodePoint code;
 224   OnigUChar *to_start = to;
 225   OnigCaseFoldType flags = *flagP;
 226
 227   while (*pp < end && to < to_end) {
 228     code = *(*pp)++;
 229     if (code == SHARP_s) {
 230       if (flags & ONIGENC_CASE_UPCASE) {
 231         flags |= ONIGENC_CASE_MODIFIED;
 232         *to++ = 'S';
 233         code = (flags & ONIGENC_CASE_TITLECASE) ? 's' : 'S';
 234       }
 235       else if (flags & ONIGENC_CASE_FOLD) {
 236         flags |= ONIGENC_CASE_MODIFIED;
 237         *to++ = 's';
 238         code = 's';
 239       }
 240     }
 241     else if (code == 0xAA || code == 0xB5 || code == 0xBA || code == 0xFF)
 242       ;
 243     else if ((EncISO_8859_9_CtypeTable[code] & BIT_CTYPE_UPPER)
 244              && (flags & (ONIGENC_CASE_DOWNCASE | ONIGENC_CASE_FOLD))) {
 245       flags |= ONIGENC_CASE_MODIFIED;
 246       if (code == 'I')
 247         code = flags & ONIGENC_CASE_FOLD_TURKISH_AZERI ? DOTLESS_i : 'i';
 248       else
 249         code = ENC_ISO_8859_9_TO_LOWER_CASE(code);
 250     }
 251     else if ((EncISO_8859_9_CtypeTable[code]&BIT_CTYPE_LOWER)
 252              && (flags & ONIGENC_CASE_UPCASE)) {
 253       flags |= ONIGENC_CASE_MODIFIED;
 254       if (code == 'i')
 255         code = flags & ONIGENC_CASE_FOLD_TURKISH_AZERI ? I_WITH_DOT_ABOVE : 'I';
 256       else if (code == DOTLESS_i)
 257         code = 'I';
 258       else
 259         code -= 0x20;
 260     }
 261     *to++ = code;
 262     if (flags & ONIGENC_CASE_TITLECASE)  /* switch from titlecase to lowercase for capitalize */
 263       flags ^= (ONIGENC_CASE_UPCASE | ONIGENC_CASE_DOWNCASE | ONIGENC_CASE_TITLECASE);
 264   }
 265   *flagP = flags;
 266   return (int )(to - to_start);
 267 }
 268
 269 OnigEncodingDefine(iso_8859_9, ISO_8859_9) = {
 270   onigenc_single_byte_mbc_enc_len,
 271   "ISO-8859-9",  /* name */
 272   1,             /* max enc length */
 273   1,             /* min enc length */
 274   onigenc_is_mbc_newline_0x0a,
 275   onigenc_single_byte_mbc_to_code,
 276   onigenc_single_byte_code_to_mbclen,
 277   onigenc_single_byte_code_to_mbc,
 278   mbc_case_fold,
 279   apply_all_case_fold,
 280   get_case_fold_codes_by_str,
 281   onigenc_minimum_property_name_to_ctype,
 282   is_code_ctype,
 283   onigenc_not_support_get_ctype_code_range,
 284   onigenc_single_byte_left_adjust_char_head,
 285   onigenc_always_true_is_allowed_reverse_match,
 286   case_map,
 287   0,
 288   ONIGENC_FLAG_NONE,
 289 };
 290 ENC_ALIAS("ISO8859-9", "ISO-8859-9")