]>
Commit | Line | Data |
---|---|---|
14b0e578 CS |
1 | #ifndef REGENC_H\r |
2 | #define REGENC_H\r | |
3 | /**********************************************************************\r | |
4 | regenc.h - Oniguruma (regular expression library)\r | |
5 | **********************************************************************/\r | |
6 | /*-\r | |
7 | * Copyright (c) 2002-2008 K.Kosako <sndgk393 AT ybb DOT ne DOT jp>\r | |
8 | * All rights reserved.\r | |
9 | *\r | |
10 | * Redistribution and use in source and binary forms, with or without\r | |
11 | * modification, are permitted provided that the following conditions\r | |
12 | * are met:\r | |
13 | * 1. Redistributions of source code must retain the above copyright\r | |
14 | * notice, this list of conditions and the following disclaimer.\r | |
15 | * 2. Redistributions in binary form must reproduce the above copyright\r | |
16 | * notice, this list of conditions and the following disclaimer in the\r | |
17 | * documentation and/or other materials provided with the distribution.\r | |
18 | *\r | |
19 | * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND\r | |
20 | * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE\r | |
21 | * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE\r | |
22 | * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE\r | |
23 | * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL\r | |
24 | * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS\r | |
25 | * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)\r | |
26 | * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT\r | |
27 | * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY\r | |
28 | * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF\r | |
29 | * SUCH DAMAGE.\r | |
30 | */\r | |
31 | \r | |
32 | #ifndef PACKAGE\r | |
33 | /* PACKAGE is defined in config.h */\r | |
34 | //#include "config.h"\r | |
35 | #endif\r | |
36 | \r | |
37 | #ifdef ONIG_ESCAPE_UCHAR_COLLISION\r | |
38 | #undef ONIG_ESCAPE_UCHAR_COLLISION\r | |
39 | #endif\r | |
40 | \r | |
41 | #include "oniguruma.h"\r | |
42 | \r | |
43 | typedef struct {\r | |
44 | OnigCodePoint from;\r | |
45 | OnigCodePoint to;\r | |
46 | } OnigPairCaseFoldCodes;\r | |
47 | \r | |
48 | \r | |
49 | #ifndef NULL\r | |
50 | #define NULL ((void* )0)\r | |
51 | #endif\r | |
52 | \r | |
53 | #ifndef TRUE\r | |
54 | #define TRUE 1\r | |
55 | #endif\r | |
56 | \r | |
57 | #ifndef FALSE\r | |
58 | #define FALSE 0\r | |
59 | #endif\r | |
60 | \r | |
61 | #ifndef ARG_UNUSED\r | |
62 | #if defined(__GNUC__)\r | |
63 | # define ARG_UNUSED __attribute__ ((unused))\r | |
64 | #else\r | |
65 | # define ARG_UNUSED\r | |
66 | #endif\r | |
67 | #endif\r | |
68 | \r | |
69 | #define ONIG_IS_NULL(p) (((void*)(p)) == (void*)0)\r | |
70 | #define ONIG_IS_NOT_NULL(p) (((void*)(p)) != (void*)0)\r | |
71 | #define ONIG_CHECK_NULL_RETURN(p) if (ONIG_IS_NULL(p)) return NULL\r | |
72 | #define ONIG_CHECK_NULL_RETURN_VAL(p,val) if (ONIG_IS_NULL(p)) return (val)\r | |
73 | \r | |
74 | #define enclen(enc,p) ONIGENC_MBC_ENC_LEN(enc,p)\r | |
75 | \r | |
76 | /* character types bit flag */\r | |
77 | #define BIT_CTYPE_NEWLINE (1<< ONIGENC_CTYPE_NEWLINE)\r | |
78 | #define BIT_CTYPE_ALPHA (1<< ONIGENC_CTYPE_ALPHA)\r | |
79 | #define BIT_CTYPE_BLANK (1<< ONIGENC_CTYPE_BLANK)\r | |
80 | #define BIT_CTYPE_CNTRL (1<< ONIGENC_CTYPE_CNTRL)\r | |
81 | #define BIT_CTYPE_DIGIT (1<< ONIGENC_CTYPE_DIGIT)\r | |
82 | #define BIT_CTYPE_GRAPH (1<< ONIGENC_CTYPE_GRAPH)\r | |
83 | #define BIT_CTYPE_LOWER (1<< ONIGENC_CTYPE_LOWER)\r | |
84 | #define BIT_CTYPE_PRINT (1<< ONIGENC_CTYPE_PRINT)\r | |
85 | #define BIT_CTYPE_PUNCT (1<< ONIGENC_CTYPE_PUNCT)\r | |
86 | #define BIT_CTYPE_SPACE (1<< ONIGENC_CTYPE_SPACE)\r | |
87 | #define BIT_CTYPE_UPPER (1<< ONIGENC_CTYPE_UPPER)\r | |
88 | #define BIT_CTYPE_XDIGIT (1<< ONIGENC_CTYPE_XDIGIT)\r | |
89 | #define BIT_CTYPE_WORD (1<< ONIGENC_CTYPE_WORD)\r | |
90 | #define BIT_CTYPE_ALNUM (1<< ONIGENC_CTYPE_ALNUM)\r | |
91 | #define BIT_CTYPE_ASCII (1<< ONIGENC_CTYPE_ASCII)\r | |
92 | \r | |
93 | #define CTYPE_TO_BIT(ctype) (1<<(ctype))\r | |
94 | #define CTYPE_IS_WORD_GRAPH_PRINT(ctype) \\r | |
95 | ((ctype) == ONIGENC_CTYPE_WORD || (ctype) == ONIGENC_CTYPE_GRAPH ||\\r | |
96 | (ctype) == ONIGENC_CTYPE_PRINT)\r | |
97 | \r | |
98 | \r | |
99 | typedef struct {\r | |
100 | UChar *name;\r | |
101 | int ctype;\r | |
102 | short int len;\r | |
103 | } PosixBracketEntryType;\r | |
104 | \r | |
105 | \r | |
106 | /* #define USE_CRNL_AS_LINE_TERMINATOR */\r | |
107 | #define USE_UNICODE_PROPERTIES\r | |
108 | /* #define USE_UNICODE_CASE_FOLD_TURKISH_AZERI */\r | |
109 | /* #define USE_UNICODE_ALL_LINE_TERMINATORS */ /* see Unicode.org UTF#18 */\r | |
110 | \r | |
111 | \r | |
112 | #define ONIG_ENCODING_INIT_DEFAULT ONIG_ENCODING_ASCII\r | |
113 | \r | |
114 | /* for encoding system implementation (internal) */\r | |
115 | ONIG_EXTERN int onigenc_ascii_apply_all_case_fold P_((OnigCaseFoldType flag, OnigApplyAllCaseFoldFunc f, void* arg));\r | |
116 | ONIG_EXTERN int onigenc_ascii_get_case_fold_codes_by_str P_((OnigCaseFoldType flag, const OnigUChar* p, const OnigUChar* end, OnigCaseFoldCodeItem items[]));\r | |
117 | ONIG_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));\r | |
118 | ONIG_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[]));\r | |
119 | ONIG_EXTERN int onigenc_not_support_get_ctype_code_range P_((OnigCtype ctype, OnigCodePoint* sb_out, const OnigCodePoint* ranges[]));\r | |
120 | ONIG_EXTERN int onigenc_is_mbc_newline_0x0a P_((const UChar* p, const UChar* end));\r | |
121 | \r | |
122 | \r | |
123 | /* methods for single byte encoding */\r | |
124 | ONIG_EXTERN int onigenc_ascii_mbc_case_fold P_((OnigCaseFoldType flag, const UChar** p, const UChar* end, UChar* lower));\r | |
125 | ONIG_EXTERN int onigenc_single_byte_mbc_enc_len P_((const UChar* p));\r | |
126 | ONIG_EXTERN OnigCodePoint onigenc_single_byte_mbc_to_code P_((const UChar* p, const UChar* end));\r | |
127 | ONIG_EXTERN int onigenc_single_byte_code_to_mbclen P_((OnigCodePoint code));\r | |
128 | ONIG_EXTERN int onigenc_single_byte_code_to_mbc P_((OnigCodePoint code, UChar *buf));\r | |
129 | ONIG_EXTERN UChar* onigenc_single_byte_left_adjust_char_head P_((const UChar* start, const UChar* s));\r | |
130 | ONIG_EXTERN int onigenc_always_true_is_allowed_reverse_match P_((const UChar* s, const UChar* end));\r | |
131 | ONIG_EXTERN int onigenc_always_false_is_allowed_reverse_match P_((const UChar* s, const UChar* end));\r | |
132 | \r | |
133 | /* methods for multi byte encoding */\r | |
134 | ONIG_EXTERN OnigCodePoint onigenc_mbn_mbc_to_code P_((OnigEncoding enc, const UChar* p, const UChar* end));\r | |
135 | ONIG_EXTERN int onigenc_mbn_mbc_case_fold P_((OnigEncoding enc, OnigCaseFoldType flag, const UChar** p, const UChar* end, UChar* lower));\r | |
136 | ONIG_EXTERN int onigenc_mb2_code_to_mbclen P_((OnigCodePoint code));\r | |
137 | ONIG_EXTERN int onigenc_mb2_code_to_mbc P_((OnigEncoding enc, OnigCodePoint code, UChar *buf));\r | |
138 | ONIG_EXTERN int onigenc_minimum_property_name_to_ctype P_((OnigEncoding enc, UChar* p, UChar* end));\r | |
139 | ONIG_EXTERN int onigenc_unicode_property_name_to_ctype P_((OnigEncoding enc, UChar* p, UChar* end));\r | |
140 | ONIG_EXTERN int onigenc_mb2_is_code_ctype P_((OnigEncoding enc, OnigCodePoint code, unsigned int ctype));\r | |
141 | ONIG_EXTERN int onigenc_mb4_code_to_mbclen P_((OnigCodePoint code));\r | |
142 | ONIG_EXTERN int onigenc_mb4_code_to_mbc P_((OnigEncoding enc, OnigCodePoint code, UChar *buf));\r | |
143 | ONIG_EXTERN int onigenc_mb4_is_code_ctype P_((OnigEncoding enc, OnigCodePoint code, unsigned int ctype));\r | |
144 | \r | |
145 | \r | |
146 | /* in enc/unicode.c */\r | |
147 | ONIG_EXTERN int onigenc_unicode_is_code_ctype P_((OnigCodePoint code, unsigned int ctype));\r | |
148 | ONIG_EXTERN int onigenc_utf16_32_get_ctype_code_range P_((OnigCtype ctype, OnigCodePoint *sb_out, const OnigCodePoint* ranges[]));\r | |
149 | ONIG_EXTERN int onigenc_unicode_ctype_code_range P_((int ctype, const OnigCodePoint* ranges[]));\r | |
150 | ONIG_EXTERN int onigenc_unicode_get_case_fold_codes_by_str P_((OnigEncoding enc, OnigCaseFoldType flag, const OnigUChar* p, const OnigUChar* end, OnigCaseFoldCodeItem items[]));\r | |
151 | ONIG_EXTERN int onigenc_unicode_mbc_case_fold P_((OnigEncoding enc, OnigCaseFoldType flag, const UChar** pp, const UChar* end, UChar* fold));\r | |
152 | ONIG_EXTERN int onigenc_unicode_apply_all_case_fold P_((OnigCaseFoldType flag, OnigApplyAllCaseFoldFunc f, void* arg));\r | |
153 | \r | |
154 | \r | |
155 | #define UTF16_IS_SURROGATE_FIRST(c) (((c) & 0xfc) == 0xd8)\r | |
156 | #define UTF16_IS_SURROGATE_SECOND(c) (((c) & 0xfc) == 0xdc)\r | |
157 | \r | |
158 | #define ONIGENC_ISO_8859_1_TO_LOWER_CASE(c) \\r | |
159 | OnigEncISO_8859_1_ToLowerCaseTable[c]\r | |
160 | #define ONIGENC_ISO_8859_1_TO_UPPER_CASE(c) \\r | |
161 | OnigEncISO_8859_1_ToUpperCaseTable[c]\r | |
162 | \r | |
163 | ONIG_EXTERN const UChar OnigEncISO_8859_1_ToLowerCaseTable[];\r | |
164 | ONIG_EXTERN const UChar OnigEncISO_8859_1_ToUpperCaseTable[];\r | |
165 | \r | |
166 | ONIG_EXTERN int\r | |
167 | onigenc_with_ascii_strncmp P_((OnigEncoding enc, const UChar* p, const UChar* end, const UChar* sascii /* ascii */, int n));\r | |
168 | ONIG_EXTERN UChar*\r | |
169 | onigenc_step P_((OnigEncoding enc, const UChar* p, const UChar* end, int n));\r | |
170 | \r | |
171 | /* defined in regexec.c, but used in enc/xxx.c */\r | |
172 | extern int onig_is_in_code_range P_((const UChar* p, OnigCodePoint code));\r | |
173 | \r | |
174 | ONIG_EXTERN OnigEncoding OnigEncDefaultCharEncoding;\r | |
175 | ONIG_EXTERN const UChar OnigEncAsciiToLowerCaseTable[];\r | |
176 | ONIG_EXTERN const UChar OnigEncAsciiToUpperCaseTable[];\r | |
177 | ONIG_EXTERN const unsigned short OnigEncAsciiCtypeTable[];\r | |
178 | \r | |
179 | #define ONIGENC_IS_ASCII_CODE(code) ((code) < 0x80)\r | |
180 | #define ONIGENC_ASCII_CODE_TO_LOWER_CASE(c) OnigEncAsciiToLowerCaseTable[c]\r | |
181 | #define ONIGENC_ASCII_CODE_TO_UPPER_CASE(c) OnigEncAsciiToUpperCaseTable[c]\r | |
182 | #define ONIGENC_IS_ASCII_CODE_CTYPE(code,ctype) \\r | |
183 | ((OnigEncAsciiCtypeTable[code] & CTYPE_TO_BIT(ctype)) != 0)\r | |
184 | #define ONIGENC_IS_ASCII_CODE_CASE_AMBIG(code) \\r | |
185 | (ONIGENC_IS_ASCII_CODE_CTYPE(code, ONIGENC_CTYPE_UPPER) ||\\r | |
186 | ONIGENC_IS_ASCII_CODE_CTYPE(code, ONIGENC_CTYPE_LOWER))\r | |
187 | \r | |
188 | \r | |
189 | #endif /* REGENC_H */\r |