1 #ifndef REGENC_H 2 #define REGENC_H 3 /********************************************************************** 4 regenc.h - Oniguruma (regular expression library) 5 **********************************************************************/ 6 /*- 7 * Copyright (c) 2002-2020 K.Kosako 8 * All rights reserved. 9 * 10 * Redistribution and use in source and binary forms, with or without 11 * modification, are permitted provided that the following conditions 12 * are met: 13 * 1. Redistributions of source code must retain the above copyright 14 * notice, this list of conditions and the following disclaimer. 15 * 2. Redistributions in binary form must reproduce the above copyright 16 * notice, this list of conditions and the following disclaimer in the 17 * documentation and/or other materials provided with the distribution. 18 * 19 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND 20 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 21 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 22 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE 23 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 24 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 25 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 26 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 27 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 28 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 29 * SUCH DAMAGE. 30 */ 31 32 #ifndef ONIGURUMA_EXPORT 33 #define ONIGURUMA_EXPORT 34 #endif 35 36 #include "config.h" 37 38 #ifndef ONIG_NO_STANDARD_C_HEADERS 39 #include <stddef.h> 40 #endif 41 42 #ifdef ONIG_ESCAPE_UCHAR_COLLISION 43 #undef ONIG_ESCAPE_UCHAR_COLLISION 44 #endif 45 46 #include "oniguruma.h" 47 48 typedef struct { 49 OnigCodePoint from; 50 OnigCodePoint to; 51 } OnigPairCaseFoldCodes; 52 53 54 #ifndef NULL 55 #define NULL ((void* )0) 56 #endif 57 58 #ifndef TRUE 59 #define TRUE 1 60 #endif 61 62 #ifndef FALSE 63 #define FALSE 0 64 #endif 65 66 #ifndef ARG_UNUSED 67 #if defined(__GNUC__) 68 # define ARG_UNUSED __attribute__ ((unused)) 69 #else 70 # define ARG_UNUSED 71 #endif 72 #endif 73 74 #define ONIG_IS_NULL(p) (((void*)(p)) == (void*)0) 75 #define ONIG_IS_NOT_NULL(p) (((void*)(p)) != (void*)0) 76 #define ONIG_CHECK_NULL_RETURN(p) if (ONIG_IS_NULL(p)) return NULL 77 #define ONIG_CHECK_NULL_RETURN_VAL(p,val) if (ONIG_IS_NULL(p)) return (val) 78 79 #define MAX_CODE_POINT (~((OnigCodePoint )0)) 80 #define ASCII_LIMIT 127 81 #define NEWLINE_CODE 0x0a 82 83 #define enclen(enc,p) ONIGENC_MBC_ENC_LEN(enc,p) 84 85 /* character types bit flag */ 86 #define BIT_CTYPE_NEWLINE (1<< ONIGENC_CTYPE_NEWLINE) 87 #define BIT_CTYPE_ALPHA (1<< ONIGENC_CTYPE_ALPHA) 88 #define BIT_CTYPE_BLANK (1<< ONIGENC_CTYPE_BLANK) 89 #define BIT_CTYPE_CNTRL (1<< ONIGENC_CTYPE_CNTRL) 90 #define BIT_CTYPE_DIGIT (1<< ONIGENC_CTYPE_DIGIT) 91 #define BIT_CTYPE_GRAPH (1<< ONIGENC_CTYPE_GRAPH) 92 #define BIT_CTYPE_LOWER (1<< ONIGENC_CTYPE_LOWER) 93 #define BIT_CTYPE_PRINT (1<< ONIGENC_CTYPE_PRINT) 94 #define BIT_CTYPE_PUNCT (1<< ONIGENC_CTYPE_PUNCT) 95 #define BIT_CTYPE_SPACE (1<< ONIGENC_CTYPE_SPACE) 96 #define BIT_CTYPE_UPPER (1<< ONIGENC_CTYPE_UPPER) 97 #define BIT_CTYPE_XDIGIT (1<< ONIGENC_CTYPE_XDIGIT) 98 #define BIT_CTYPE_WORD (1<< ONIGENC_CTYPE_WORD) 99 #define BIT_CTYPE_ALNUM (1<< ONIGENC_CTYPE_ALNUM) 100 #define BIT_CTYPE_ASCII (1<< ONIGENC_CTYPE_ASCII) 101 102 #define CTYPE_TO_BIT(ctype) (1<<(ctype)) 103 #define CTYPE_IS_WORD_GRAPH_PRINT(ctype) \ 104 ((ctype) == ONIGENC_CTYPE_WORD || (ctype) == ONIGENC_CTYPE_GRAPH ||\ 105 (ctype) == ONIGENC_CTYPE_PRINT) 106 107 108 typedef struct { 109 UChar *name; 110 int ctype; 111 short int len; 112 } PosixBracketEntryType; 113 114 struct PropertyNameCtype { 115 char *name; 116 int ctype; 117 }; 118 119 /* #define USE_CRNL_AS_LINE_TERMINATOR */ 120 #define USE_UNICODE_PROPERTIES 121 #define USE_UNICODE_EXTENDED_GRAPHEME_CLUSTER 122 #define USE_UNICODE_WORD_BREAK 123 /* #define USE_UNICODE_CASE_FOLD_TURKISH_AZERI */ 124 /* #define USE_UNICODE_ALL_LINE_TERMINATORS */ /* see Unicode.org UTS #18 */ 125 126 127 #define ONIG_ENCODING_INIT_DEFAULT ONIG_ENCODING_ASCII 128 129 130 #define ENC_SKIP_OFFSET_1_OR_0 7 131 132 #define ENC_FLAG_ASCII_COMPATIBLE (1<<0) 133 #define ENC_FLAG_UNICODE (1<<1) 134 #define ENC_FLAG_SKIP_OFFSET_MASK (7<<2) 135 #define ENC_FLAG_SKIP_OFFSET_0 0 136 #define ENC_FLAG_SKIP_OFFSET_1 (1<<2) 137 #define ENC_FLAG_SKIP_OFFSET_2 (2<<2) 138 #define ENC_FLAG_SKIP_OFFSET_3 (3<<2) 139 #define ENC_FLAG_SKIP_OFFSET_4 (4<<2) 140 #define ENC_FLAG_SKIP_OFFSET_1_OR_0 (ENC_SKIP_OFFSET_1_OR_0<<2) 141 142 #define ENC_GET_SKIP_OFFSET(enc) \ 143 (((enc)->flag & ENC_FLAG_SKIP_OFFSET_MASK)>>2) 144 145 #define CASE_FOLD_IS_ASCII_ONLY(flag) \ 146 (((flag) & ONIGENC_CASE_FOLD_ASCII_ONLY) != 0) 147 #define CASE_FOLD_IS_NOT_ASCII_ONLY(flag) \ 148 (((flag) & ONIGENC_CASE_FOLD_ASCII_ONLY) == 0) 149 150 /* for encoding system implementation (internal) */ 151 extern int onigenc_end(void); 152 extern int onigenc_ascii_apply_all_case_fold P_((OnigCaseFoldType flag, OnigApplyAllCaseFoldFunc f, void* arg)); 153 extern int onigenc_ascii_get_case_fold_codes_by_str P_((OnigCaseFoldType flag, const OnigUChar* p, const OnigUChar* end, OnigCaseFoldCodeItem items[])); 154 extern int onigenc_apply_all_case_fold_with_map P_((int map_size, const OnigPairCaseFoldCodes map[], int ess_tsett_flag, OnigCaseFoldType flag, OnigApplyAllCaseFoldFunc f, void* arg)); 155 extern int onigenc_get_case_fold_codes_by_str_with_map P_((int map_size, const OnigPairCaseFoldCodes map[], int ess_tsett_flag, OnigCaseFoldType flag, const OnigUChar* p, const OnigUChar* end, OnigCaseFoldCodeItem items[])); 156 extern int onigenc_not_support_get_ctype_code_range P_((OnigCtype ctype, OnigCodePoint* sb_out, const OnigCodePoint* ranges[])); 157 extern int onigenc_is_mbc_newline_0x0a P_((const UChar* p, const UChar* end)); 158 159 160 /* methods for single byte encoding */ 161 extern int onigenc_ascii_mbc_case_fold P_((OnigCaseFoldType flag, const UChar** p, const UChar* end, UChar* lower)); 162 extern int onigenc_single_byte_mbc_enc_len P_((const UChar* p)); 163 extern OnigCodePoint onigenc_single_byte_mbc_to_code P_((const UChar* p, const UChar* end)); 164 extern int onigenc_single_byte_code_to_mbclen P_((OnigCodePoint code)); 165 extern int onigenc_single_byte_code_to_mbc P_((OnigCodePoint code, UChar *buf)); 166 extern UChar* onigenc_single_byte_left_adjust_char_head P_((const UChar* start, const UChar* s)); 167 extern int onigenc_always_true_is_allowed_reverse_match P_((const UChar* s, const UChar* end)); 168 extern int onigenc_always_false_is_allowed_reverse_match P_((const UChar* s, const UChar* end)); 169 extern int onigenc_always_true_is_valid_mbc_string P_((const UChar* s, const UChar* end)); 170 extern int onigenc_length_check_is_valid_mbc_string P_((OnigEncoding enc, const UChar* s, const UChar* end)); 171 172 /* methods for multi byte encoding */ 173 extern OnigCodePoint onigenc_mbn_mbc_to_code P_((OnigEncoding enc, const UChar* p, const UChar* end)); 174 extern int onigenc_mbn_mbc_case_fold P_((OnigEncoding enc, OnigCaseFoldType flag, const UChar** p, const UChar* end, UChar* lower)); 175 extern int onigenc_mb2_code_to_mbc P_((OnigEncoding enc, OnigCodePoint code, UChar *buf)); 176 extern int onigenc_minimum_property_name_to_ctype P_((OnigEncoding enc, UChar* p, UChar* end)); 177 extern int onigenc_unicode_property_name_to_ctype P_((OnigEncoding enc, UChar* p, UChar* end)); 178 extern int onigenc_is_mbc_word_ascii P_((OnigEncoding enc, UChar* s, const UChar* end)); 179 extern int onigenc_mb2_is_code_ctype P_((OnigEncoding enc, OnigCodePoint code, unsigned int ctype)); 180 extern int onigenc_mb4_code_to_mbc P_((OnigEncoding enc, OnigCodePoint code, UChar *buf)); 181 extern int onigenc_mb4_is_code_ctype P_((OnigEncoding enc, OnigCodePoint code, unsigned int ctype)); 182 extern struct PropertyNameCtype* onigenc_euc_jp_lookup_property_name P_((register const char *str, register size_t len)); 183 extern struct PropertyNameCtype* onigenc_sjis_lookup_property_name P_((register const char *str, register size_t len)); 184 185 /* in unicode.c */ 186 extern int onigenc_unicode_is_code_ctype P_((OnigCodePoint code, unsigned int ctype)); 187 extern int onigenc_utf16_32_get_ctype_code_range P_((OnigCtype ctype, OnigCodePoint *sb_out, const OnigCodePoint* ranges[])); 188 extern int onigenc_unicode_ctype_code_range P_((OnigCtype ctype, const OnigCodePoint* ranges[])); 189 extern int onigenc_unicode_get_case_fold_codes_by_str P_((OnigEncoding enc, OnigCaseFoldType flag, const OnigUChar* p, const OnigUChar* end, OnigCaseFoldCodeItem items[])); 190 extern int onigenc_unicode_mbc_case_fold P_((OnigEncoding enc, OnigCaseFoldType flag, const UChar** pp, const UChar* end, UChar* fold)); 191 extern int onigenc_unicode_apply_all_case_fold P_((OnigCaseFoldType flag, OnigApplyAllCaseFoldFunc f, void* arg)); 192 193 extern int onigenc_egcb_is_break_position P_((OnigEncoding enc, UChar* p, UChar* prev, const UChar* start, const UChar* end)); 194 195 #ifdef USE_UNICODE_WORD_BREAK 196 extern int onigenc_wb_is_break_position P_((OnigEncoding enc, UChar* p, UChar* prev, const UChar* start, const UChar* end)); 197 #endif 198 199 #define UTF16_IS_SURROGATE_FIRST(c) (((c) & 0xfc) == 0xd8) 200 #define UTF16_IS_SURROGATE_SECOND(c) (((c) & 0xfc) == 0xdc) 201 202 /* from unicode generated codes */ 203 #define FOLDS1_FOLD(i) (OnigUnicodeFolds1 + (i)) 204 #define FOLDS2_FOLD(i) (OnigUnicodeFolds2 + (i)) 205 #define FOLDS3_FOLD(i) (OnigUnicodeFolds3 + (i)) 206 #define FOLDS1_UNFOLDS_NUM(i) (OnigUnicodeFolds1[(i)+1]) 207 #define FOLDS2_UNFOLDS_NUM(i) (OnigUnicodeFolds2[(i)+2]) 208 #define FOLDS3_UNFOLDS_NUM(i) (OnigUnicodeFolds3[(i)+3]) 209 #define FOLDS1_UNFOLDS(i) (FOLDS1_FOLD(i) + 2) 210 #define FOLDS2_UNFOLDS(i) (FOLDS2_FOLD(i) + 3) 211 #define FOLDS3_UNFOLDS(i) (FOLDS3_FOLD(i) + 4) 212 #define FOLDS1_NEXT_INDEX(i) ((i) + 2 + FOLDS1_UNFOLDS_NUM(i)) 213 #define FOLDS2_NEXT_INDEX(i) ((i) + 3 + FOLDS2_UNFOLDS_NUM(i)) 214 #define FOLDS3_NEXT_INDEX(i) ((i) + 4 + FOLDS3_UNFOLDS_NUM(i)) 215 216 #define FOLDS_FOLD_ADDR_BUK(buk, addr) do {\ 217 if ((buk)->fold_len == 1)\ 218 addr = OnigUnicodeFolds1 + (buk)->index;\ 219 else if ((buk)->fold_len == 2)\ 220 addr = OnigUnicodeFolds2 + (buk)->index;\ 221 else if ((buk)->fold_len == 3)\ 222 addr = OnigUnicodeFolds3 + (buk)->index;\ 223 else\ 224 return ONIGERR_INVALID_CODE_POINT_VALUE;\ 225 } while (0) 226 227 extern OnigCodePoint OnigUnicodeFolds1[]; 228 extern OnigCodePoint OnigUnicodeFolds2[]; 229 extern OnigCodePoint OnigUnicodeFolds3[]; 230 231 struct ByUnfoldKey { 232 OnigCodePoint code; 233 short int index; 234 short int fold_len; 235 }; 236 237 extern const struct ByUnfoldKey* onigenc_unicode_unfold_key(OnigCodePoint code); 238 extern int onigenc_unicode_fold1_key(OnigCodePoint code[]); 239 extern int onigenc_unicode_fold2_key(OnigCodePoint code[]); 240 extern int onigenc_unicode_fold3_key(OnigCodePoint code[]); 241 242 extern int onig_codes_cmp(OnigCodePoint a[], OnigCodePoint b[], int n); 243 extern int onig_codes_byte_at(OnigCodePoint code[], int at); 244 245 246 247 #define ONIGENC_ISO_8859_1_TO_LOWER_CASE(c) \ 248 OnigEncISO_8859_1_ToLowerCaseTable[c] 249 #define ONIGENC_ISO_8859_1_TO_UPPER_CASE(c) \ 250 OnigEncISO_8859_1_ToUpperCaseTable[c] 251 252 extern const UChar OnigEncISO_8859_1_ToLowerCaseTable[]; 253 extern const UChar OnigEncISO_8859_1_ToUpperCaseTable[]; 254 255 extern int 256 onigenc_with_ascii_strncmp P_((OnigEncoding enc, const UChar* p, const UChar* end, const UChar* sascii /* ascii */, int n)); 257 extern UChar* 258 onigenc_step P_((OnigEncoding enc, const UChar* p, const UChar* end, int n)); 259 260 /* defined in regexec.c, but used in enc/xxx.c */ 261 extern int onig_is_in_code_range P_((const UChar* p, OnigCodePoint code)); 262 263 extern OnigEncoding OnigEncDefaultCharEncoding; 264 extern const UChar OnigEncAsciiToLowerCaseTable[]; 265 extern const UChar OnigEncAsciiToUpperCaseTable[]; 266 extern const unsigned short OnigEncAsciiCtypeTable[]; 267 268 269 #define ONIGENC_IS_ASCII_CODE(code) ((code) < 0x80) 270 #define ONIGENC_ASCII_CODE_TO_LOWER_CASE(c) OnigEncAsciiToLowerCaseTable[c] 271 #define ONIGENC_ASCII_CODE_TO_UPPER_CASE(c) OnigEncAsciiToUpperCaseTable[c] 272 #define ONIGENC_IS_ASCII_CODE_CTYPE(code,ctype) \ 273 ((OnigEncAsciiCtypeTable[code] & CTYPE_TO_BIT(ctype)) != 0) 274 #define ONIGENC_IS_ASCII_CODE_WORD(code) \ 275 ((OnigEncAsciiCtypeTable[code] & CTYPE_TO_BIT(ONIGENC_CTYPE_WORD)) != 0) 276 #define ONIGENC_IS_ASCII_CODE_CASE_AMBIG(code) \ 277 (ONIGENC_IS_ASCII_CODE_CTYPE(code, ONIGENC_CTYPE_UPPER) ||\ 278 ONIGENC_IS_ASCII_CODE_CTYPE(code, ONIGENC_CTYPE_LOWER)) 279 280 #define ONIGENC_IS_UNICODE_ENCODING(enc) \ 281 (((enc)->flag & ENC_FLAG_UNICODE) != 0) 282 283 #define ONIGENC_IS_ASCII_COMPATIBLE_ENCODING(enc) \ 284 (((enc)->flag & ENC_FLAG_ASCII_COMPATIBLE) != 0) 285 286 #endif /* REGENC_H */ 287