1 #ifndef REGENC_H 2 #define REGENC_H 3 /********************************************************************** 4 regenc.h - Oniguruma (regular expression library) 5 **********************************************************************/ 6 /*- 7 * Copyright (c) 2002-2019 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 #include <stddef.h> 38 39 #ifdef ONIG_ESCAPE_UCHAR_COLLISION 40 #undef ONIG_ESCAPE_UCHAR_COLLISION 41 #endif 42 43 #include "oniguruma.h" 44 45 typedef struct { 46 OnigCodePoint from; 47 OnigCodePoint to; 48 } OnigPairCaseFoldCodes; 49 50 51 #ifndef NULL 52 #define NULL ((void* )0) 53 #endif 54 55 #ifndef TRUE 56 #define TRUE 1 57 #endif 58 59 #ifndef FALSE 60 #define FALSE 0 61 #endif 62 63 #ifndef ARG_UNUSED 64 #if defined(__GNUC__) 65 # define ARG_UNUSED __attribute__ ((unused)) 66 #else 67 # define ARG_UNUSED 68 #endif 69 #endif 70 71 #define ONIG_IS_NULL(p) (((void*)(p)) == (void*)0) 72 #define ONIG_IS_NOT_NULL(p) (((void*)(p)) != (void*)0) 73 #define ONIG_CHECK_NULL_RETURN(p) if (ONIG_IS_NULL(p)) return NULL 74 #define ONIG_CHECK_NULL_RETURN_VAL(p,val) if (ONIG_IS_NULL(p)) return (val) 75 76 #define MAX_CODE_POINT (~((OnigCodePoint )0)) 77 78 #define enclen(enc,p) ONIGENC_MBC_ENC_LEN(enc,p) 79 80 /* character types bit flag */ 81 #define BIT_CTYPE_NEWLINE (1<< ONIGENC_CTYPE_NEWLINE) 82 #define BIT_CTYPE_ALPHA (1<< ONIGENC_CTYPE_ALPHA) 83 #define BIT_CTYPE_BLANK (1<< ONIGENC_CTYPE_BLANK) 84 #define BIT_CTYPE_CNTRL (1<< ONIGENC_CTYPE_CNTRL) 85 #define BIT_CTYPE_DIGIT (1<< ONIGENC_CTYPE_DIGIT) 86 #define BIT_CTYPE_GRAPH (1<< ONIGENC_CTYPE_GRAPH) 87 #define BIT_CTYPE_LOWER (1<< ONIGENC_CTYPE_LOWER) 88 #define BIT_CTYPE_PRINT (1<< ONIGENC_CTYPE_PRINT) 89 #define BIT_CTYPE_PUNCT (1<< ONIGENC_CTYPE_PUNCT) 90 #define BIT_CTYPE_SPACE (1<< ONIGENC_CTYPE_SPACE) 91 #define BIT_CTYPE_UPPER (1<< ONIGENC_CTYPE_UPPER) 92 #define BIT_CTYPE_XDIGIT (1<< ONIGENC_CTYPE_XDIGIT) 93 #define BIT_CTYPE_WORD (1<< ONIGENC_CTYPE_WORD) 94 #define BIT_CTYPE_ALNUM (1<< ONIGENC_CTYPE_ALNUM) 95 #define BIT_CTYPE_ASCII (1<< ONIGENC_CTYPE_ASCII) 96 97 #define CTYPE_TO_BIT(ctype) (1<<(ctype)) 98 #define CTYPE_IS_WORD_GRAPH_PRINT(ctype) \ 99 ((ctype) == ONIGENC_CTYPE_WORD || (ctype) == ONIGENC_CTYPE_GRAPH ||\ 100 (ctype) == ONIGENC_CTYPE_PRINT) 101 102 103 typedef struct { 104 UChar *name; 105 int ctype; 106 short int len; 107 } PosixBracketEntryType; 108 109 struct PropertyNameCtype { 110 char *name; 111 int ctype; 112 }; 113 114 /* #define USE_CRNL_AS_LINE_TERMINATOR */ 115 #define USE_UNICODE_PROPERTIES 116 #define USE_UNICODE_EXTENDED_GRAPHEME_CLUSTER 117 #define USE_UNICODE_WORD_BREAK 118 /* #define USE_UNICODE_CASE_FOLD_TURKISH_AZERI */ 119 /* #define USE_UNICODE_ALL_LINE_TERMINATORS */ /* see Unicode.org UTS #18 */ 120 121 122 #define ONIG_ENCODING_INIT_DEFAULT ONIG_ENCODING_ASCII 123 124 125 #define ENC_SKIP_OFFSET_1_OR_0 7 126 127 #define ENC_FLAG_ASCII_COMPATIBLE (1<<0) 128 #define ENC_FLAG_UNICODE (1<<1) 129 #define ENC_FLAG_SKIP_OFFSET_MASK (7<<2) 130 #define ENC_FLAG_SKIP_OFFSET_0 0 131 #define ENC_FLAG_SKIP_OFFSET_1 (1<<2) 132 #define ENC_FLAG_SKIP_OFFSET_2 (2<<2) 133 #define ENC_FLAG_SKIP_OFFSET_3 (3<<2) 134 #define ENC_FLAG_SKIP_OFFSET_4 (4<<2) 135 #define ENC_FLAG_SKIP_OFFSET_1_OR_0 (ENC_SKIP_OFFSET_1_OR_0<<2) 136 137 #define ENC_GET_SKIP_OFFSET(enc) \ 138 (((enc)->flag & ENC_FLAG_SKIP_OFFSET_MASK)>>2) 139 140 141 /* for encoding system implementation (internal) */ 142 extern int onigenc_end(void); 143 extern int onigenc_ascii_apply_all_case_fold P_((OnigCaseFoldType flag, OnigApplyAllCaseFoldFunc f, void* arg)); 144 extern int onigenc_ascii_get_case_fold_codes_by_str P_((OnigCaseFoldType flag, const OnigUChar* p, const OnigUChar* end, OnigCaseFoldCodeItem items[])); 145 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)); 146 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[])); 147 extern int onigenc_not_support_get_ctype_code_range P_((OnigCtype ctype, OnigCodePoint* sb_out, const OnigCodePoint* ranges[])); 148 extern int onigenc_is_mbc_newline_0x0a P_((const UChar* p, const UChar* end)); 149 150 151 /* methods for single byte encoding */ 152 extern int onigenc_ascii_mbc_case_fold P_((OnigCaseFoldType flag, const UChar** p, const UChar* end, UChar* lower)); 153 extern int onigenc_single_byte_mbc_enc_len P_((const UChar* p)); 154 extern OnigCodePoint onigenc_single_byte_mbc_to_code P_((const UChar* p, const UChar* end)); 155 extern int onigenc_single_byte_code_to_mbclen P_((OnigCodePoint code)); 156 extern int onigenc_single_byte_code_to_mbc P_((OnigCodePoint code, UChar *buf)); 157 extern UChar* onigenc_single_byte_left_adjust_char_head P_((const UChar* start, const UChar* s)); 158 extern int onigenc_always_true_is_allowed_reverse_match P_((const UChar* s, const UChar* end)); 159 extern int onigenc_always_false_is_allowed_reverse_match P_((const UChar* s, const UChar* end)); 160 extern int onigenc_always_true_is_valid_mbc_string P_((const UChar* s, const UChar* end)); 161 extern int onigenc_length_check_is_valid_mbc_string P_((OnigEncoding enc, const UChar* s, const UChar* end)); 162 163 /* methods for multi byte encoding */ 164 extern OnigCodePoint onigenc_mbn_mbc_to_code P_((OnigEncoding enc, const UChar* p, const UChar* end)); 165 extern int onigenc_mbn_mbc_case_fold P_((OnigEncoding enc, OnigCaseFoldType flag, const UChar** p, const UChar* end, UChar* lower)); 166 extern int onigenc_mb2_code_to_mbc P_((OnigEncoding enc, OnigCodePoint code, UChar *buf)); 167 extern int onigenc_minimum_property_name_to_ctype P_((OnigEncoding enc, UChar* p, UChar* end)); 168 extern int onigenc_unicode_property_name_to_ctype P_((OnigEncoding enc, UChar* p, UChar* end)); 169 extern int onigenc_is_mbc_word_ascii P_((OnigEncoding enc, UChar* s, const UChar* end)); 170 extern int onigenc_mb2_is_code_ctype P_((OnigEncoding enc, OnigCodePoint code, unsigned int ctype)); 171 extern int onigenc_mb4_code_to_mbc P_((OnigEncoding enc, OnigCodePoint code, UChar *buf)); 172 extern int onigenc_mb4_is_code_ctype P_((OnigEncoding enc, OnigCodePoint code, unsigned int ctype)); 173 extern struct PropertyNameCtype* onigenc_euc_jp_lookup_property_name P_((register const char *str, register size_t len)); 174 extern struct PropertyNameCtype* onigenc_sjis_lookup_property_name P_((register const char *str, register size_t len)); 175 176 /* in unicode.c */ 177 extern int onigenc_unicode_is_code_ctype P_((OnigCodePoint code, unsigned int ctype)); 178 extern int onigenc_utf16_32_get_ctype_code_range P_((OnigCtype ctype, OnigCodePoint *sb_out, const OnigCodePoint* ranges[])); 179 extern int onigenc_unicode_ctype_code_range P_((OnigCtype ctype, const OnigCodePoint* ranges[])); 180 extern int onigenc_unicode_get_case_fold_codes_by_str P_((OnigEncoding enc, OnigCaseFoldType flag, const OnigUChar* p, const OnigUChar* end, OnigCaseFoldCodeItem items[])); 181 extern int onigenc_unicode_mbc_case_fold P_((OnigEncoding enc, OnigCaseFoldType flag, const UChar** pp, const UChar* end, UChar* fold)); 182 extern int onigenc_unicode_apply_all_case_fold P_((OnigCaseFoldType flag, OnigApplyAllCaseFoldFunc f, void* arg)); 183 184 extern int onigenc_egcb_is_break_position P_((OnigEncoding enc, UChar* p, UChar* prev, const UChar* start, const UChar* end)); 185 186 #ifdef USE_UNICODE_WORD_BREAK 187 extern int onigenc_wb_is_break_position P_((OnigEncoding enc, UChar* p, UChar* prev, const UChar* start, const UChar* end)); 188 #endif 189 190 #define UTF16_IS_SURROGATE_FIRST(c) (((c) & 0xfc) == 0xd8) 191 #define UTF16_IS_SURROGATE_SECOND(c) (((c) & 0xfc) == 0xdc) 192 193 /* from unicode generated codes */ 194 #define FOLDS1_FOLD(i) (OnigUnicodeFolds1 + (i)) 195 #define FOLDS2_FOLD(i) (OnigUnicodeFolds2 + (i)) 196 #define FOLDS3_FOLD(i) (OnigUnicodeFolds3 + (i)) 197 #define FOLDS1_UNFOLDS_NUM(i) (OnigUnicodeFolds1[(i)+1]) 198 #define FOLDS2_UNFOLDS_NUM(i) (OnigUnicodeFolds2[(i)+2]) 199 #define FOLDS3_UNFOLDS_NUM(i) (OnigUnicodeFolds3[(i)+3]) 200 #define FOLDS1_UNFOLDS(i) (OnigUnicodeFolds1 + (i) + 2) 201 #define FOLDS2_UNFOLDS(i) (OnigUnicodeFolds2 + (i) + 3) 202 #define FOLDS3_UNFOLDS(i) (OnigUnicodeFolds3 + (i) + 4) 203 #define FOLDS1_NEXT_INDEX(i) ((i) + 2 + OnigUnicodeFolds1[(i)+1]) 204 #define FOLDS2_NEXT_INDEX(i) ((i) + 3 + OnigUnicodeFolds2[(i)+2]) 205 #define FOLDS3_NEXT_INDEX(i) ((i) + 4 + OnigUnicodeFolds3[(i)+3]) 206 207 #define FOLDS_FOLD_ADDR_BUK(buk, addr) do {\ 208 if ((buk)->fold_len == 1)\ 209 addr = OnigUnicodeFolds1 + (buk)->index;\ 210 else if ((buk)->fold_len == 2)\ 211 addr = OnigUnicodeFolds2 + (buk)->index;\ 212 else if ((buk)->fold_len == 3)\ 213 addr = OnigUnicodeFolds3 + (buk)->index;\ 214 else\ 215 return ONIGERR_INVALID_CODE_POINT_VALUE;\ 216 } while (0) 217 218 extern OnigCodePoint OnigUnicodeFolds1[]; 219 extern OnigCodePoint OnigUnicodeFolds2[]; 220 extern OnigCodePoint OnigUnicodeFolds3[]; 221 222 struct ByUnfoldKey { 223 OnigCodePoint code; 224 short int index; 225 short int fold_len; 226 }; 227 228 extern const struct ByUnfoldKey* onigenc_unicode_unfold_key(OnigCodePoint code); 229 extern int onigenc_unicode_fold1_key(OnigCodePoint code[]); 230 extern int onigenc_unicode_fold2_key(OnigCodePoint code[]); 231 extern int onigenc_unicode_fold3_key(OnigCodePoint code[]); 232 233 extern int onig_codes_cmp(OnigCodePoint a[], OnigCodePoint b[], int n); 234 extern int onig_codes_byte_at(OnigCodePoint code[], int at); 235 236 237 238 #define ONIGENC_ISO_8859_1_TO_LOWER_CASE(c) \ 239 OnigEncISO_8859_1_ToLowerCaseTable[c] 240 #define ONIGENC_ISO_8859_1_TO_UPPER_CASE(c) \ 241 OnigEncISO_8859_1_ToUpperCaseTable[c] 242 243 extern const UChar OnigEncISO_8859_1_ToLowerCaseTable[]; 244 extern const UChar OnigEncISO_8859_1_ToUpperCaseTable[]; 245 246 extern int 247 onigenc_with_ascii_strncmp P_((OnigEncoding enc, const UChar* p, const UChar* end, const UChar* sascii /* ascii */, int n)); 248 extern UChar* 249 onigenc_step P_((OnigEncoding enc, const UChar* p, const UChar* end, int n)); 250 251 /* defined in regexec.c, but used in enc/xxx.c */ 252 extern int onig_is_in_code_range P_((const UChar* p, OnigCodePoint code)); 253 254 extern OnigEncoding OnigEncDefaultCharEncoding; 255 extern const UChar OnigEncAsciiToLowerCaseTable[]; 256 extern const UChar OnigEncAsciiToUpperCaseTable[]; 257 extern const unsigned short OnigEncAsciiCtypeTable[]; 258 259 260 #define ONIGENC_IS_ASCII_CODE(code) ((code) < 0x80) 261 #define ONIGENC_ASCII_CODE_TO_LOWER_CASE(c) OnigEncAsciiToLowerCaseTable[c] 262 #define ONIGENC_ASCII_CODE_TO_UPPER_CASE(c) OnigEncAsciiToUpperCaseTable[c] 263 #define ONIGENC_IS_ASCII_CODE_CTYPE(code,ctype) \ 264 ((OnigEncAsciiCtypeTable[code] & CTYPE_TO_BIT(ctype)) != 0) 265 #define ONIGENC_IS_ASCII_CODE_WORD(code) \ 266 ((OnigEncAsciiCtypeTable[code] & CTYPE_TO_BIT(ONIGENC_CTYPE_WORD)) != 0) 267 #define ONIGENC_IS_ASCII_CODE_CASE_AMBIG(code) \ 268 (ONIGENC_IS_ASCII_CODE_CTYPE(code, ONIGENC_CTYPE_UPPER) ||\ 269 ONIGENC_IS_ASCII_CODE_CTYPE(code, ONIGENC_CTYPE_LOWER)) 270 271 #define ONIGENC_IS_UNICODE_ENCODING(enc) \ 272 (((enc)->flag & ENC_FLAG_UNICODE) != 0) 273 274 #define ONIGENC_IS_ASCII_COMPATIBLE_ENCODING(enc) \ 275 (((enc)->flag & ENC_FLAG_ASCII_COMPATIBLE) != 0) 276 277 #endif /* REGENC_H */ 278