1 #ifndef _REGEX_H_
2 #define _REGEX_H_				/* never again */
3 /*
4  * regular expressions
5  *
6  * Copyright (c) 1998, 1999 Henry Spencer.  All rights reserved.
7  *
8  * Development of this software was funded, in part, by Cray Research Inc.,
9  * UUNET Communications Services Inc., Sun Microsystems Inc., and Scriptics
10  * Corporation, none of whom are responsible for the results.  The author
11  * thanks all of them.
12  *
13  * Redistribution and use in source and binary forms -- with or without
14  * modification -- are permitted for any purpose, provided that
15  * redistributions in source form retain this entire copyright notice and
16  * indicate the origin and nature of any modifications.
17  *
18  * I'd appreciate being given credit for this package in the documentation
19  * of software which uses it, but that is not a requirement.
20  *
21  * THIS SOFTWARE IS PROVIDED ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES,
22  * INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY
23  * AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL
24  * HENRY SPENCER BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
25  * EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
26  * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS;
27  * OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
28  * WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR
29  * OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF
30  * ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
31  *
32  * src/include/regex/regex.h
33  */
34 
35 /*
36  * Add your own defines, if needed, here.
37  */
38 #include "mb/pg_wchar.h"
39 
40 /*
41  * interface types etc.
42  */
43 
44 /*
45  * regoff_t has to be large enough to hold either off_t or ssize_t,
46  * and must be signed; it's only a guess that long is suitable.
47  */
48 typedef long regoff_t;
49 
50 /*
51  * other interface types
52  */
53 
54 /* the biggie, a compiled RE (or rather, a front end to same) */
55 typedef struct
56 {
57 	int			re_magic;		/* magic number */
58 	size_t		re_nsub;		/* number of subexpressions */
59 	long		re_info;		/* information about RE */
60 #define  REG_UBACKREF		 000001
61 #define  REG_ULOOKAROUND	 000002
62 #define  REG_UBOUNDS	 000004
63 #define  REG_UBRACES	 000010
64 #define  REG_UBSALNUM		 000020
65 #define  REG_UPBOTCH	 000040
66 #define  REG_UBBS		 000100
67 #define  REG_UNONPOSIX		 000200
68 #define  REG_UUNSPEC	 000400
69 #define  REG_UUNPORT	 001000
70 #define  REG_ULOCALE	 002000
71 #define  REG_UEMPTYMATCH	 004000
72 #define  REG_UIMPOSSIBLE	 010000
73 #define  REG_USHORTEST		 020000
74 	int			re_csize;		/* sizeof(character) */
75 	char	   *re_endp;		/* backward compatibility kludge */
76 	Oid			re_collation;	/* Collation that defines LC_CTYPE behavior */
77 	/* the rest is opaque pointers to hidden innards */
78 	char	   *re_guts;		/* `char *' is more portable than `void *' */
79 	char	   *re_fns;
80 } regex_t;
81 
82 /* result reporting (may acquire more fields later) */
83 typedef struct
84 {
85 	regoff_t	rm_so;			/* start of substring */
86 	regoff_t	rm_eo;			/* end of substring */
87 } regmatch_t;
88 
89 /* supplementary control and reporting */
90 typedef struct
91 {
92 	regmatch_t	rm_extend;		/* see REG_EXPECT */
93 } rm_detail_t;
94 
95 
96 
97 /*
98  * regex compilation flags
99  */
100 #define REG_BASIC	000000		/* BREs (convenience) */
101 #define REG_EXTENDED	000001	/* EREs */
102 #define REG_ADVF	000002		/* advanced features in EREs */
103 #define REG_ADVANCED	000003	/* AREs (which are also EREs) */
104 #define REG_QUOTE	000004		/* no special characters, none */
105 #define REG_NOSPEC	REG_QUOTE	/* historical synonym */
106 #define REG_ICASE	000010		/* ignore case */
107 #define REG_NOSUB	000020		/* don't care about subexpressions */
108 #define REG_EXPANDED	000040	/* expanded format, white space & comments */
109 #define REG_NLSTOP	000100		/* \n doesn't match . or [^ ] */
110 #define REG_NLANCH	000200		/* ^ matches after \n, $ before */
111 #define REG_NEWLINE 000300		/* newlines are line terminators */
112 #define REG_PEND	000400		/* ugh -- backward-compatibility hack */
113 #define REG_EXPECT	001000		/* report details on partial/limited matches */
114 #define REG_BOSONLY 002000		/* temporary kludge for BOS-only matches */
115 #define REG_DUMP	004000		/* none of your business :-) */
116 #define REG_FAKE	010000		/* none of your business :-) */
117 #define REG_PROGRESS	020000	/* none of your business :-) */
118 
119 
120 
121 /*
122  * regex execution flags
123  */
124 #define REG_NOTBOL	0001		/* BOS is not BOL */
125 #define REG_NOTEOL	0002		/* EOS is not EOL */
126 #define REG_STARTEND	0004	/* backward compatibility kludge */
127 #define REG_FTRACE	0010		/* none of your business */
128 #define REG_MTRACE	0020		/* none of your business */
129 #define REG_SMALL	0040		/* none of your business */
130 
131 
132 /*
133  * error reporting
134  * Be careful if modifying the list of error codes -- the table used by
135  * regerror() is generated automatically from this file!
136  */
137 #define REG_OKAY	 0			/* no errors detected */
138 #define REG_NOMATCH  1			/* failed to match */
139 #define REG_BADPAT	 2			/* invalid regexp */
140 #define REG_ECOLLATE	 3		/* invalid collating element */
141 #define REG_ECTYPE	 4			/* invalid character class */
142 #define REG_EESCAPE  5			/* invalid escape \ sequence */
143 #define REG_ESUBREG  6			/* invalid backreference number */
144 #define REG_EBRACK	 7			/* brackets [] not balanced */
145 #define REG_EPAREN	 8			/* parentheses () not balanced */
146 #define REG_EBRACE	 9			/* braces {} not balanced */
147 #define REG_BADBR	10			/* invalid repetition count(s) */
148 #define REG_ERANGE	11			/* invalid character range */
149 #define REG_ESPACE	12			/* out of memory */
150 #define REG_BADRPT	13			/* quantifier operand invalid */
151 #define REG_ASSERT	15			/* "can't happen" -- you found a bug */
152 #define REG_INVARG	16			/* invalid argument to regex function */
153 #define REG_MIXED	17			/* character widths of regex and string differ */
154 #define REG_BADOPT	18			/* invalid embedded option */
155 #define REG_ETOOBIG 19			/* regular expression is too complex */
156 #define REG_ECOLORS 20			/* too many colors */
157 #define REG_CANCEL	21			/* operation cancelled */
158 /* two specials for debugging and testing */
159 #define REG_ATOI	101			/* convert error-code name to number */
160 #define REG_ITOA	102			/* convert error-code number to name */
161 /* non-error result codes for pg_regprefix */
162 #define REG_PREFIX	(-1)		/* identified a common prefix */
163 #define REG_EXACT	(-2)		/* identified an exact match */
164 
165 
166 
167 /*
168  * the prototypes for exported functions
169  */
170 extern int	pg_regcomp(regex_t *, const pg_wchar *, size_t, int, Oid);
171 extern int	pg_regexec(regex_t *, const pg_wchar *, size_t, size_t, rm_detail_t *, size_t, regmatch_t[], int);
172 extern int	pg_regprefix(regex_t *, pg_wchar **, size_t *);
173 extern void pg_regfree(regex_t *);
174 extern size_t pg_regerror(int, const regex_t *, char *, size_t);
175 extern void pg_set_regex_collation(Oid collation);
176 
177 #endif   /* _REGEX_H_ */
178