PostgreSQL Source Code  git master
regex.h
Go to the documentation of this file.
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; /* bitmask of the following flags: */
60 #define REG_UBACKREF 000001 /* has back-reference (\n) */
61 #define REG_ULOOKAROUND 000002 /* has lookahead/lookbehind constraint */
62 #define REG_UBOUNDS 000004 /* has bounded quantifier ({m,n}) */
63 #define REG_UBRACES 000010 /* has { that doesn't begin a quantifier */
64 #define REG_UBSALNUM 000020 /* has backslash-alphanumeric in non-ARE */
65 #define REG_UPBOTCH 000040 /* has unmatched right paren in ERE (legal
66  * per spec, but that was a mistake) */
67 #define REG_UBBS 000100 /* has backslash within bracket expr */
68 #define REG_UNONPOSIX 000200 /* has any construct that extends POSIX */
69 #define REG_UUNSPEC 000400 /* has any case disallowed by POSIX, e.g.
70  * an empty branch */
71 #define REG_UUNPORT 001000 /* has numeric character code dependency */
72 #define REG_ULOCALE 002000 /* has locale dependency */
73 #define REG_UEMPTYMATCH 004000 /* can match a zero-length string */
74 #define REG_UIMPOSSIBLE 010000 /* provably cannot match anything */
75 #define REG_USHORTEST 020000 /* has non-greedy quantifier */
76  int re_csize; /* sizeof(character) */
77  char *re_endp; /* backward compatibility kludge */
78  Oid re_collation; /* Collation that defines LC_CTYPE behavior */
79  /* the rest is opaque pointers to hidden innards */
80  char *re_guts; /* `char *' is more portable than `void *' */
81  char *re_fns;
82 } regex_t;
83 
84 /* result reporting (may acquire more fields later) */
85 typedef struct
86 {
87  regoff_t rm_so; /* start of substring */
88  regoff_t rm_eo; /* end of substring */
89 } regmatch_t;
90 
91 /* supplementary control and reporting */
92 typedef struct
93 {
94  regmatch_t rm_extend; /* see REG_EXPECT */
95 } rm_detail_t;
96 
97 
98 
99 /*
100  * regex compilation flags
101  */
102 #define REG_BASIC 000000 /* BREs (convenience) */
103 #define REG_EXTENDED 000001 /* EREs */
104 #define REG_ADVF 000002 /* advanced features in EREs */
105 #define REG_ADVANCED 000003 /* AREs (which are also EREs) */
106 #define REG_QUOTE 000004 /* no special characters, none */
107 #define REG_NOSPEC REG_QUOTE /* historical synonym */
108 #define REG_ICASE 000010 /* ignore case */
109 #define REG_NOSUB 000020 /* caller doesn't need subexpr match data */
110 #define REG_EXPANDED 000040 /* expanded format, white space & comments */
111 #define REG_NLSTOP 000100 /* \n doesn't match . or [^ ] */
112 #define REG_NLANCH 000200 /* ^ matches after \n, $ before */
113 #define REG_NEWLINE 000300 /* newlines are line terminators */
114 #define REG_PEND 000400 /* ugh -- backward-compatibility hack */
115 #define REG_EXPECT 001000 /* report details on partial/limited matches */
116 #define REG_BOSONLY 002000 /* temporary kludge for BOS-only matches */
117 #define REG_DUMP 004000 /* none of your business :-) */
118 #define REG_FAKE 010000 /* none of your business :-) */
119 #define REG_PROGRESS 020000 /* none of your business :-) */
120 
121 
122 
123 /*
124  * regex execution flags
125  */
126 #define REG_NOTBOL 0001 /* BOS is not BOL */
127 #define REG_NOTEOL 0002 /* EOS is not EOL */
128 #define REG_STARTEND 0004 /* backward compatibility kludge */
129 #define REG_FTRACE 0010 /* none of your business */
130 #define REG_MTRACE 0020 /* none of your business */
131 #define REG_SMALL 0040 /* none of your business */
132 
133 
134 /*
135  * error reporting
136  * Be careful if modifying the list of error codes -- the table used by
137  * regerror() is generated automatically from this file!
138  */
139 #define REG_OKAY 0 /* no errors detected */
140 #define REG_NOMATCH 1 /* failed to match */
141 #define REG_BADPAT 2 /* invalid regexp */
142 #define REG_ECOLLATE 3 /* invalid collating element */
143 #define REG_ECTYPE 4 /* invalid character class */
144 #define REG_EESCAPE 5 /* invalid escape \ sequence */
145 #define REG_ESUBREG 6 /* invalid backreference number */
146 #define REG_EBRACK 7 /* brackets [] not balanced */
147 #define REG_EPAREN 8 /* parentheses () not balanced */
148 #define REG_EBRACE 9 /* braces {} not balanced */
149 #define REG_BADBR 10 /* invalid repetition count(s) */
150 #define REG_ERANGE 11 /* invalid character range */
151 #define REG_ESPACE 12 /* out of memory */
152 #define REG_BADRPT 13 /* quantifier operand invalid */
153 #define REG_ASSERT 15 /* "can't happen" -- you found a bug */
154 #define REG_INVARG 16 /* invalid argument to regex function */
155 #define REG_MIXED 17 /* character widths of regex and string differ */
156 #define REG_BADOPT 18 /* invalid embedded option */
157 #define REG_ETOOBIG 19 /* regular expression is too complex */
158 #define REG_ECOLORS 20 /* too many colors */
159 /* two specials for debugging and testing */
160 #define REG_ATOI 101 /* convert error-code name to number */
161 #define REG_ITOA 102 /* convert error-code number to name */
162 /* non-error result codes for pg_regprefix */
163 #define REG_PREFIX (-1) /* identified a common prefix */
164 #define REG_EXACT (-2) /* identified an exact match */
165 
166 
167 
168 /*
169  * the prototypes for exported functions
170  */
171 
172 /* regcomp.c */
173 extern int pg_regcomp(regex_t *re, const pg_wchar *string, size_t len,
174  int flags, Oid collation);
175 extern int pg_regexec(regex_t *re, const pg_wchar *string, size_t len,
176  size_t search_start, rm_detail_t *details,
177  size_t nmatch, regmatch_t pmatch[], int flags);
178 extern int pg_regprefix(regex_t *re, pg_wchar **string, size_t *slength);
179 extern void pg_regfree(regex_t *re);
180 extern size_t pg_regerror(int errcode, const regex_t *preg, char *errbuf,
181  size_t errbuf_size);
182 
183 /* regexp.c */
184 extern regex_t *RE_compile_and_cache(text *text_re, int cflags, Oid collation);
185 extern bool RE_compile_and_execute(text *text_re, char *dat, int dat_len,
186  int cflags, Oid collation,
187  int nmatch, regmatch_t *pmatch);
188 
189 #endif /* _REGEX_H_ */
int errcode(int sqlerrcode)
Definition: elog.c:859
unsigned int pg_wchar
Definition: mbprint.c:31
const void size_t len
unsigned int Oid
Definition: postgres_ext.h:31
bool RE_compile_and_execute(text *text_re, char *dat, int dat_len, int cflags, Oid collation, int nmatch, regmatch_t *pmatch)
Definition: regexp.c:358
size_t pg_regerror(int errcode, const regex_t *preg, char *errbuf, size_t errbuf_size)
Definition: regerror.c:60
int pg_regprefix(regex_t *re, pg_wchar **string, size_t *slength)
Definition: regprefix.c:46
int pg_regcomp(regex_t *re, const pg_wchar *string, size_t len, int flags, Oid collation)
Definition: regcomp.c:370
regex_t * RE_compile_and_cache(text *text_re, int cflags, Oid collation)
Definition: regexp.c:141
int pg_regexec(regex_t *re, const pg_wchar *string, size_t len, size_t search_start, rm_detail_t *details, size_t nmatch, regmatch_t pmatch[], int flags)
Definition: regexec.c:185
void pg_regfree(regex_t *re)
Definition: regfree.c:49
long regoff_t
Definition: regex.h:48
Definition: regex.h:56
long re_info
Definition: regex.h:59
size_t re_nsub
Definition: regex.h:58
int re_magic
Definition: regex.h:57
Definition: c.h:674