2 * The authors of this software are Rob Pike and Ken Thompson.
3 * Copyright (c) 2002 by Lucent Technologies.
5 * Permission to use, copy, modify, and distribute this software for any
6 * purpose without fee is hereby granted, provided that this entire notice
7 * is included in all copies of any software which is or includes a copy
8 * or modification of this software and in all copies of the supporting
9 * documentation for such software.
10 * THIS SOFTWARE IS BEING PROVIDED "AS IS", WITHOUT ANY EXPRESS OR IMPLIED
11 * WARRANTY. IN PARTICULAR, NEITHER THE AUTHORS NOR LUCENT TECHNOLOGIES MAKE ANY
12 * REPRESENTATION OR WARRANTY OF ANY KIND CONCERNING THE MERCHANTABILITY
13 * OF THIS SOFTWARE OR ITS FITNESS FOR ANY PARTICULAR PURPOSE.
15 * heavily modified by Ketmar // Vampire Avalon
28 * A regular expression specifies a set of strings of characters. A
29 * member of this set of strings is said to be matched by the regular
30 * expression. In the following specification for regular expressions the
31 * word `character' means any character (rune) but newline.
33 * The syntax for a regular expression e0 is
34 * e3: literal | charclass | '.' | '^' | '$' | '(?:' e0 ')' | '(' e0 ')'
39 * REP: '*' | '+' | '?'
48 * A literal is any non-metacharacter, or a metacharacter (one of .*+?[]()|\^$).
50 * A charclass is a nonempty string s bracketed [s] (or [^s]); it matches any
51 * character in (or not in) s. A negated character class never matches newline.
52 * A substring a-b, with a and b in ascending order, stands for the inclusive
53 * range of characters between a and b. In s, the metacharacters '-', ']', an
54 * initial '^' must be preceded by a '\'; other metacharacters have no special
55 * meaning and may appear unescaped.
57 * A '^' matches the beginning of a line; '$' matches the end of the line.
58 * A '.' matches any character.
59 * A '\z' matches character with code 0 ('\0').
60 * A '\d' matches digits ([0-9]).
61 * A '\s' matches space ([\f\t\r\n\v ]).
62 * A '\w' matches 'word character' ([0-9A-Za-z_]).
63 * A '\b' matches 'word boundary'.
64 * One can use 'negative metacharacters' too ('\Z', '\D', etc.).
66 * The REP operators match zero or more (*), one or more (+), zero or one (?),
67 * instances respectively of the preceding regular expression e2.
69 * An alternative regular expression, e0|e1, matches either a match to e0 or a match to e1.
71 * A match to any part of a regular expression extends as far as possible without preventing
72 * a match to the remainder of the regular expression.
74 * Remember that POSIX classes ([:alpha:], etc) covers only ASCII characters!
81 * define this to allow RE9_FLAG_CASEINSENS work with non-ascii characters
83 * RE9_DISABLE_NONGREEDY
84 * define this to disable non-greedy closures (the engine will be somewhat faster)
86 * RE9_DISABLE_POSIX_CLASSES
87 * disable parsing of posix classes like [:space:], [:digit:], etc.
89 * RE9_DISABLE_LEARNING
90 * disable regexp 'learning' in compiler
93 #ifdef REGEXP9_DEBUG_MEMSIZE
94 extern int re9_memused
;
98 /* maximum is 127, vm opcode limits this */
99 enum { RE9_SUBEXP_MAX
= 16 };
102 #if defined(__GNUC__) && __GNUC__ >= 3
103 # define REGEXP9_ATTR_PURE __attribute__((pure))
105 # define REGEXP9_ATTR_PURE
109 /* subexpression matches */
116 /* program definition */
117 typedef struct re9_prog_s re9_prog_t
;
122 RE9_FLAG_NONUTF8
= 0x01, /* for both compile and execute */
123 RE9_FLAG_ANYDOT
= 0x02, /* '.' matches newline too; for both compile and execute */
124 /* only for compiler */
125 RE9_FLAG_LITERAL
= 0x10,
126 RE9_FLAG_NONGREEDY
= 0x20, /* invert default repetition mode */
127 RE9_FLAG_CASEINSENS
= 0x40, /* only for ASCII */
128 /* only for interpreter */
129 RE9_FLAG_MT0_RANGE
= 0x100, /* use match[0].sp and match[0].ep as string start and end */
131 RE9_FLAG_DUMPPRG
= 0x1000
136 * re9_compile() compiles a regular expression and returns a pointer to the generated description.
137 * re9_compile() returns 0 for an illegal expression or other failure.
138 * Compiler is thread-safe.
139 * errmsg can be NULL; if *errmsg is not NULL, it SHOULD NOT be free()d or modified!
141 extern re9_prog_t
*re9_compile (const char *s
, int flags
, const char **errmsg
);
143 extern re9_prog_t
*re9_compile_ex (const char *s
, const char *eol
, int flags
, const char **errmsg
);
145 /* return number of captures in this regexp; 0th capture is always full match */
146 extern int re9_nsub (const re9_prog_t
*p
);
149 * free compiled regular expression
151 extern void re9_free (re9_prog_t
*p
);
153 /* re9_execute() matches a null-terminated string against the compiled regular expression in prog.
154 * If it matches, regexec returns 1 and fills in the array match with character pointers to the
155 * substrings of string that correspond to the parenthesized subexpressions of exp: match[i].sp
156 * points to the beginning and match[i].ep points just beyond the end of the ith substring.
157 * (Subexpression i begins at the ith left parenthesis, counting from 1.) Pointers in match[0]
158 * pick out the substring that corresponds to the whole regular expression.
159 * If match[0].sp is nonzero on entry, regexec starts matching at that point within string.
160 * If match[0].ep is nonzero on entry, the last character matched is the one preceding that point.
161 * Unused elements of match are filled with zeros. Matches involving and are extended as far as
162 * possible. The number of array elements in match is given by msize.
163 * mp can be NULL and ms should be 0 in this case.
164 * re9_execute() returns 0 if string is not matched.
165 * Executor is thread-safe, one program can be used in multiple threads simultaneously.
167 /* progp: program to run
168 * bol: string to run machine on
169 * mp: subexpression elements (can be NULL)
170 * ms: number of elements at mp (should be 0 if mp is NULL)
172 extern int re9_execute (const re9_prog_t
*progp
, int flags
, const char *bol
, re9_sub_t
*mp
, int ms
);
174 /* executor can eat up to maxmem memory */
175 extern int re9_execute_ex (const re9_prog_t
*progp
, int flags
, const char *bol
, re9_sub_t
*mp
, int ms
, size_t maxmem
);
178 /* 'prepared for execution' program */
179 typedef struct re9_prog_prepared_s re9_prog_prepared_t
;
182 /* 'prepare' program -- allocate some memory, etc.
183 * this can be used to avoid excessive malloc()s.
184 * return NULL on error.
186 extern re9_prog_prepared_t
*re9_prepare (const re9_prog_t
*progp
);
188 /* executor can eat up to maxmem memory */
189 extern re9_prog_prepared_t
*re9_prepare_ex (const re9_prog_t
*progp
, size_t maxmem
);
191 /* execute 'prepared' program */
192 extern int re9_prepared_execute (re9_prog_prepared_t
*pp
, int flags
, const char *bol
, re9_sub_t
*mp
, int ms
);
194 /* free 'prepared' program */
195 extern void re9_prepared_free (re9_prog_prepared_t
*pp
);
198 /* re9_sub() places in dp a substitution instance of sp in the context of the last regexec
199 * performed using match. Each instance of '\n', where n is a digit, is replaced by the string
200 * delimited by match[n].sp and match[n].ep. Each instance of is replaced by the string
201 * delimited by match[0].sp and match[0].ep. The substitution will always be null terminated
202 * and trimmed to fit into dlen bytes.
203 * Function will return number of bytes needed to successfully insert everything (including trailing 0).
204 * Use '\{num}' to insert 10th and bigger match.
205 * Use '\z' to insert character with code 0.
208 * dp: destination string
209 * dlen: destination string size
210 * mp: subexpression elements
211 * ms: number of elements at mp
213 extern int re9_sub (char *dp
, size_t dlen
, const char *sp
, const re9_sub_t
*mp
, int ms
) REGEXP9_ATTR_PURE
;