xref: /illumos-gate/usr/src/head/regex.h (revision edb348833aaacfa1176e502ad38875fd0b2717ab)
1 /*
2  * CDDL HEADER START
3  *
4  * The contents of this file are subject to the terms of the
5  * Common Development and Distribution License, Version 1.0 only
6  * (the "License").  You may not use this file except in compliance
7  * with the License.
8  *
9  * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
10  * or http://www.opensolaris.org/os/licensing.
11  * See the License for the specific language governing permissions
12  * and limitations under the License.
13  *
14  * When distributing Covered Code, include this CDDL HEADER in each
15  * file and include the License file at usr/src/OPENSOLARIS.LICENSE.
16  * If applicable, add the following below this CDDL HEADER, with the
17  * fields enclosed by brackets "[]" replaced with your own identifying
18  * information: Portions Copyright [yyyy] [name of copyright owner]
19  *
20  * CDDL HEADER END
21  */
22 /*
23  * Copyright 2005 Sun Microsystems, Inc.  All rights reserved.
24  * Use is subject to license terms.
25  */
26 
27 /*
28  * Copyright 1989, 1994 by Mortice Kern Systems Inc.
29  * All rights reserved.
30  */
31 /*
32  * Copyright 2010 Nexenta Systems, Inc.  All rights reserved.
33  * Use is subject to license terms.
34  */
35 
36 #ifndef	_REGEX_H
37 #define	_REGEX_H
38 
39 #include <sys/feature_tests.h>
40 #include <sys/types.h>
41 
42 #ifdef	__cplusplus
43 extern "C" {
44 #endif
45 
46 
47 /*
48  * wchar_t is a built-in type in standard C++ and as such is not
49  * defined here when using standard C++. However, the GNU compiler
50  * fixincludes utility nonetheless creates its own version of this
51  * header for use by gcc and g++. In that version it adds a redundant
52  * guard for __cplusplus. To avoid the creation of a gcc/g++ specific
53  * header we need to include the following magic comment:
54  *
55  * we must use the C++ compiler's type
56  *
57  * The above comment should not be removed or changed until GNU
58  * gcc/fixinc/inclhack.def is updated to bypass this header.
59  */
60 #if !defined(__cplusplus) || (__cplusplus < 199711L && !defined(__GNUG__))
61 #ifndef _WCHAR_T
62 #define	_WCHAR_T
63 #if defined(_LP64)
64 typedef int	wchar_t;
65 #else
66 typedef long    wchar_t;
67 #endif
68 #endif	/* !_WCHAR_T */
69 #endif	/* !defined(__cplusplus) ... */
70 
71 typedef ssize_t regoff_t;
72 
73 /* regcomp flags */
74 #define	REG_BASIC	0x00
75 #define	REG_EXTENDED	0x01		/* Use Extended Regular Expressions */
76 #define	REG_NEWLINE	0x08		/* Treat \n as regular character */
77 #define	REG_ICASE	0x04		/* Ignore case in match */
78 #define	REG_NOSUB	0x02		/* Don't set subexpression */
79 #define	REG_EGREP	0x1000		/* running as egrep(1) */
80 
81 /* non-standard flags - note that most of these are not supported */
82 #define	REG_DELIM	0x10		/* string[0] is delimiter */
83 #define	REG_DEBUG	0x20		/* Debug recomp and regexec */
84 #define	REG_ANCHOR	0x40		/* Implicit ^ and $ */
85 #define	REG_WORDS	0x80		/* \< and \> match word boundries */
86 
87 /* FreeBSD additions */
88 #define	REG_DUMP	0x2000
89 #define	REG_PEND	0x4000
90 #define	REG_NOSPEC	0x8000
91 #define	REG_STARTEND	0x10000
92 
93 /* internal flags */
94 #define	REG_MUST	0x100		/* check for regmust substring */
95 
96 /* regexec flags */
97 #define	REG_NOTBOL	0x200		/* string is not BOL */
98 #define	REG_NOTEOL	0x400		/* string has no EOL */
99 #define	REG_NOOPT	0x800		/* don't do regmust optimization */
100 
101 /* regcomp and regexec return codes */
102 #define	REG_OK		0		/* success (non-standard) */
103 #define	REG_NOMATCH	1		/* regexec failed to match */
104 #define	REG_ECOLLATE	2		/* invalid collation element ref. */
105 #define	REG_EESCAPE	3		/* trailing \ in pattern */
106 #define	REG_ENEWLINE	4		/* \n found before end of pattern */
107 #define	REG_ENSUB	5		/* more than 9 \( \) pairs (OBS) */
108 #define	REG_ESUBREG	6		/* number in \[0-9] invalid */
109 #define	REG_EBRACK	7		/* [ ] imbalance */
110 #define	REG_EPAREN	8		/* ( ) imbalance */
111 #define	REG_EBRACE	9		/* \{ \} imbalance */
112 #define	REG_ERANGE	10		/* invalid endpoint in range */
113 #define	REG_ESPACE	11		/* no memory for compiled pattern */
114 #define	REG_BADRPT	12		/* invalid repetition */
115 #define	REG_ECTYPE	13		/* invalid char-class type */
116 #define	REG_BADPAT	14		/* syntax error */
117 #define	REG_BADBR	15		/* \{ \} contents bad */
118 #define	REG_EFATAL	16		/* internal error, not POSIX.2 */
119 #define	REG_ECHAR	17		/* invalid mulitbyte character */
120 #define	REG_STACK	18		/* backtrack stack overflow */
121 #define	REG_ENOSYS	19		/* function not supported (XPG4) */
122 #define	REG__LAST	20		/* first unused code */
123 #define	REG_EBOL	21		/* ^ anchor and not BOL */
124 #define	REG_EEOL	22		/* $ anchor and not EOL */
125 #define	_REG_BACKREF_MAX 9		/* Max # of subexp. backreference */
126 
127 typedef struct {		/* regcomp() data saved for regexec() */
128 	size_t  re_nsub;	/* # of subexpressions in RE pattern */
129 
130 	/*
131 	 * Internal use only.  Note that any changes to this structure
132 	 * have to preserve sizing, as it is baked into applications.
133 	 */
134 	struct re_guts *re_g;
135 	int re_magic;
136 	const char *re_endp;
137 
138 	/* here for compat */
139 	size_t	re_len;		/* # wchar_t chars in compiled pattern */
140 	struct _regex_ext_t *re_sc;	/* for binary compatibility */
141 } regex_t;
142 
143 /* subexpression positions */
144 typedef struct {
145 #ifdef __STDC__
146 	const char	*rm_sp, *rm_ep;	/* Start pointer, end pointer */
147 #else
148 	char		*rm_sp, *rm_ep;	/* Start pointer, end pointer */
149 #endif
150 	regoff_t	rm_so, rm_eo;	/* Start offset, end offset */
151 	int		rm_ss, rm_es;	/* Used internally */
152 } regmatch_t;
153 
154 
155 /*
156  * Additional API and structs to support regular expression manipulations
157  * on wide characters.
158  */
159 
160 #if defined(__STDC__)
161 
162 extern int regcomp(regex_t *_RESTRICT_KYWD, const char *_RESTRICT_KYWD, int);
163 extern int regexec(const regex_t *_RESTRICT_KYWD, const char *_RESTRICT_KYWD,
164 	size_t, regmatch_t *_RESTRICT_KYWD, int);
165 extern size_t regerror(int, const regex_t *_RESTRICT_KYWD,
166 	char *_RESTRICT_KYWD, size_t);
167 extern void regfree(regex_t *);
168 
169 #else  /* defined(__STDC__) */
170 
171 extern int regcomp();
172 extern int regexec();
173 extern size_t regerror();
174 extern void regfree();
175 
176 #endif  /* defined(__STDC__) */
177 
178 #ifdef	__cplusplus
179 }
180 #endif
181 
182 #endif	/* _REGEX_H */
183