2013-07-29 09:04:59 +00:00
|
|
|
/*
|
|
|
|
* Copyright (c) 2004-2013 Sergey Lyubka <valenok@gmail.com>
|
2013-08-01 08:14:09 +00:00
|
|
|
* Copyright (c) 2013 Cesanta Software Limited
|
2013-07-29 09:04:59 +00:00
|
|
|
* All rights reserved
|
|
|
|
*
|
|
|
|
* This library is dual-licensed: you can redistribute it and/or modify
|
|
|
|
* it under the terms of the GNU General Public License version 2 as
|
|
|
|
* published by the Free Software Foundation. For the terms of this
|
|
|
|
* license, see <http://www.gnu.org/licenses/>.
|
|
|
|
*
|
|
|
|
* You are free to use this library under the terms of the GNU General
|
|
|
|
* Public License, but WITHOUT ANY WARRANTY; without even the implied
|
|
|
|
* warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
|
|
|
|
* See the GNU General Public License for more details.
|
|
|
|
*
|
|
|
|
* Alternatively, you can license this library under a commercial
|
|
|
|
* license, as set out in <http://cesanta.com/products.html>.
|
|
|
|
*/
|
|
|
|
|
|
|
|
#include <stdio.h>
|
|
|
|
#include <ctype.h>
|
|
|
|
#include <string.h>
|
|
|
|
|
|
|
|
#include "slre.h"
|
|
|
|
|
|
|
|
#define MAX_BRANCHES 100
|
|
|
|
#define MAX_BRACKETS 100
|
2013-10-01 11:13:00 +00:00
|
|
|
#define FAIL_IF(condition, error_code) if (condition) return (error_code)
|
2013-07-29 09:04:59 +00:00
|
|
|
|
2014-08-22 16:54:40 +00:00
|
|
|
#ifndef ARRAY_SIZE
|
2014-08-25 07:20:56 +00:00
|
|
|
#define ARRAY_SIZE(ar) (sizeof(ar) / sizeof((ar)[0]))
|
2014-08-22 16:54:40 +00:00
|
|
|
#endif
|
|
|
|
|
2013-07-29 09:04:59 +00:00
|
|
|
#ifdef SLRE_DEBUG
|
|
|
|
#define DBG(x) printf x
|
|
|
|
#else
|
|
|
|
#define DBG(x)
|
|
|
|
#endif
|
|
|
|
|
2013-08-26 16:11:20 +00:00
|
|
|
struct bracket_pair {
|
|
|
|
const char *ptr; /* Points to the first char after '(' in regex */
|
|
|
|
int len; /* Length of the text between '(' and ')' */
|
|
|
|
int branches; /* Index in the branches array for this pair */
|
|
|
|
int num_branches; /* Number of '|' in this bracket pair */
|
|
|
|
};
|
|
|
|
|
|
|
|
struct branch {
|
2013-09-10 23:25:56 +00:00
|
|
|
int bracket_index; /* index for 'struct bracket_pair brackets' */
|
|
|
|
/* array defined below */
|
2013-08-26 16:11:20 +00:00
|
|
|
const char *schlong; /* points to the '|' character in the regex */
|
|
|
|
};
|
|
|
|
|
2013-07-29 09:04:59 +00:00
|
|
|
struct regex_info {
|
|
|
|
/*
|
|
|
|
* Describes all bracket pairs in the regular expression.
|
|
|
|
* First entry is always present, and grabs the whole regex.
|
|
|
|
*/
|
2013-08-26 16:11:20 +00:00
|
|
|
struct bracket_pair brackets[MAX_BRACKETS];
|
2013-08-01 08:14:09 +00:00
|
|
|
int num_brackets;
|
2013-07-29 09:04:59 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Describes alternations ('|' operators) in the regular expression.
|
|
|
|
* Each branch falls into a specific branch pair.
|
|
|
|
*/
|
2013-08-26 16:11:20 +00:00
|
|
|
struct branch branches[MAX_BRANCHES];
|
2013-07-29 09:04:59 +00:00
|
|
|
int num_branches;
|
|
|
|
|
2013-08-07 10:01:28 +00:00
|
|
|
/* Array of captures provided by the user */
|
|
|
|
struct slre_cap *caps;
|
|
|
|
int num_caps;
|
|
|
|
|
2014-07-14 10:32:14 +00:00
|
|
|
/* E.g. SLRE_IGNORE_CASE. See enum below */
|
2013-07-29 09:04:59 +00:00
|
|
|
int flags;
|
|
|
|
};
|
|
|
|
|
2013-08-03 18:42:11 +00:00
|
|
|
static int is_metacharacter(const unsigned char *s) {
|
2014-08-25 09:45:01 +00:00
|
|
|
static const char *metacharacters = "^$().[]*+?|\\Ssdbfnrtv";
|
2013-10-01 11:13:00 +00:00
|
|
|
return strchr(metacharacters, *s) != NULL;
|
2013-08-03 18:42:11 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
static int op_len(const char *re) {
|
|
|
|
return re[0] == '\\' && re[1] == 'x' ? 4 : re[0] == '\\' ? 2 : 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int set_len(const char *re, int re_len) {
|
|
|
|
int len = 0;
|
|
|
|
|
|
|
|
while (len < re_len && re[len] != ']') {
|
|
|
|
len += op_len(re + len);
|
|
|
|
}
|
|
|
|
|
|
|
|
return len <= re_len ? len + 1 : -1;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int get_op_len(const char *re, int re_len) {
|
|
|
|
return re[0] == '[' ? set_len(re + 1, re_len - 1) + 1 : op_len(re);
|
2013-07-29 09:04:59 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
static int is_quantifier(const char *re) {
|
|
|
|
return re[0] == '*' || re[0] == '+' || re[0] == '?';
|
|
|
|
}
|
|
|
|
|
2013-08-02 13:21:48 +00:00
|
|
|
static int toi(int x) {
|
|
|
|
return isdigit(x) ? x - '0' : x - 'W';
|
|
|
|
}
|
2013-08-03 18:42:11 +00:00
|
|
|
|
2013-08-02 13:21:48 +00:00
|
|
|
static int hextoi(const unsigned char *s) {
|
|
|
|
return (toi(tolower(s[0])) << 4) | toi(tolower(s[1]));
|
|
|
|
}
|
|
|
|
|
2013-08-03 18:42:11 +00:00
|
|
|
static int match_op(const unsigned char *re, const unsigned char *s,
|
|
|
|
struct regex_info *info) {
|
|
|
|
int result = 0;
|
|
|
|
switch (*re) {
|
|
|
|
case '\\':
|
|
|
|
/* Metacharacters */
|
|
|
|
switch (re[1]) {
|
2014-08-25 09:45:01 +00:00
|
|
|
case 'S': FAIL_IF(isspace(*s), SLRE_NO_MATCH); result++; break;
|
|
|
|
case 's': FAIL_IF(!isspace(*s), SLRE_NO_MATCH); result++; break;
|
|
|
|
case 'd': FAIL_IF(!isdigit(*s), SLRE_NO_MATCH); result++; break;
|
|
|
|
case 'b': FAIL_IF(*s != '\b', SLRE_NO_MATCH); result++; break;
|
|
|
|
case 'f': FAIL_IF(*s != '\f', SLRE_NO_MATCH); result++; break;
|
|
|
|
case 'n': FAIL_IF(*s != '\n', SLRE_NO_MATCH); result++; break;
|
|
|
|
case 'r': FAIL_IF(*s != '\r', SLRE_NO_MATCH); result++; break;
|
|
|
|
case 't': FAIL_IF(*s != '\t', SLRE_NO_MATCH); result++; break;
|
|
|
|
case 'v': FAIL_IF(*s != '\v', SLRE_NO_MATCH); result++; break;
|
2013-08-03 18:42:11 +00:00
|
|
|
|
|
|
|
case 'x':
|
|
|
|
/* Match byte, \xHH where HH is hexadecimal byte representaion */
|
2013-10-01 11:13:00 +00:00
|
|
|
FAIL_IF(hextoi(re + 2) != *s, SLRE_NO_MATCH);
|
2013-08-03 18:42:11 +00:00
|
|
|
result++;
|
|
|
|
break;
|
|
|
|
|
|
|
|
default:
|
2013-09-30 18:30:50 +00:00
|
|
|
/* Valid metacharacter check is done in bar() */
|
2013-10-01 11:13:00 +00:00
|
|
|
FAIL_IF(re[1] != s[0], SLRE_NO_MATCH);
|
2013-09-30 18:30:50 +00:00
|
|
|
result++;
|
2013-08-03 18:42:11 +00:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
break;
|
|
|
|
|
2013-10-01 11:13:00 +00:00
|
|
|
case '|': FAIL_IF(1, SLRE_INTERNAL_ERROR); break;
|
|
|
|
case '$': FAIL_IF(1, SLRE_NO_MATCH); break;
|
2013-08-03 18:42:11 +00:00
|
|
|
case '.': result++; break;
|
|
|
|
|
|
|
|
default:
|
2014-07-14 10:32:14 +00:00
|
|
|
if (info->flags & SLRE_IGNORE_CASE) {
|
2013-10-01 11:13:00 +00:00
|
|
|
FAIL_IF(tolower(*re) != tolower(*s), SLRE_NO_MATCH);
|
2013-08-03 18:42:11 +00:00
|
|
|
} else {
|
2013-10-01 11:13:00 +00:00
|
|
|
FAIL_IF(*re != *s, SLRE_NO_MATCH);
|
2013-08-03 18:42:11 +00:00
|
|
|
}
|
|
|
|
result++;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
return result;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int match_set(const char *re, int re_len, const char *s,
|
|
|
|
struct regex_info *info) {
|
2013-09-30 18:30:50 +00:00
|
|
|
int len = 0, result = -1, invert = re[0] == '^';
|
2013-08-03 18:42:11 +00:00
|
|
|
|
|
|
|
if (invert) re++, re_len--;
|
|
|
|
|
2013-09-30 18:30:50 +00:00
|
|
|
while (len <= re_len && re[len] != ']' && result <= 0) {
|
2013-08-03 18:42:11 +00:00
|
|
|
/* Support character range */
|
|
|
|
if (re[len] != '-' && re[len + 1] == '-' && re[len + 2] != ']' &&
|
|
|
|
re[len + 2] != '\0') {
|
2015-07-17 06:34:26 +00:00
|
|
|
result = info->flags & SLRE_IGNORE_CASE ?
|
2015-06-10 10:19:43 +00:00
|
|
|
tolower(*s) >= tolower(re[len]) && tolower(*s) <= tolower(re[len + 2]) :
|
|
|
|
*s >= re[len] && *s <= re[len + 2];
|
2013-08-03 18:42:11 +00:00
|
|
|
len += 3;
|
|
|
|
} else {
|
2016-10-27 15:09:34 +00:00
|
|
|
result = match_op((const unsigned char *) re + len, (const unsigned char *) s, info);
|
2013-08-07 10:01:28 +00:00
|
|
|
len += op_len(re + len);
|
2013-08-03 18:42:11 +00:00
|
|
|
}
|
|
|
|
}
|
2013-09-30 18:30:50 +00:00
|
|
|
return (!invert && result > 0) || (invert && result <= 0) ? 1 : -1;
|
2013-08-03 18:42:11 +00:00
|
|
|
}
|
|
|
|
|
2013-08-07 10:01:28 +00:00
|
|
|
static int doh(const char *s, int s_len, struct regex_info *info, int bi);
|
2013-07-29 22:24:41 +00:00
|
|
|
|
2013-08-01 08:14:09 +00:00
|
|
|
static int bar(const char *re, int re_len, const char *s, int s_len,
|
2013-08-07 10:01:28 +00:00
|
|
|
struct regex_info *info, int bi) {
|
2013-08-01 08:14:09 +00:00
|
|
|
/* i is offset in re, j is offset in s, bi is brackets index */
|
|
|
|
int i, j, n, step;
|
2013-07-29 09:04:59 +00:00
|
|
|
|
2013-09-28 01:06:18 +00:00
|
|
|
for (i = j = 0; i < re_len && j <= s_len; i += step) {
|
2013-07-29 09:04:59 +00:00
|
|
|
|
2013-08-03 09:02:13 +00:00
|
|
|
/* Handle quantifiers. Get the length of the chunk. */
|
2013-08-03 18:42:11 +00:00
|
|
|
step = re[i] == '(' ? info->brackets[bi + 1].len + 2 :
|
|
|
|
get_op_len(re + i, re_len - i);
|
2013-08-03 09:02:13 +00:00
|
|
|
|
2013-09-30 18:30:50 +00:00
|
|
|
DBG(("%s [%.*s] [%.*s] re_len=%d step=%d i=%d j=%d\n", __func__,
|
2013-08-03 09:02:13 +00:00
|
|
|
re_len - i, re + i, s_len - j, s + j, re_len, step, i, j));
|
2013-07-29 09:04:59 +00:00
|
|
|
|
2013-10-01 11:13:00 +00:00
|
|
|
FAIL_IF(is_quantifier(&re[i]), SLRE_UNEXPECTED_QUANTIFIER);
|
|
|
|
FAIL_IF(step <= 0, SLRE_INVALID_CHARACTER_SET);
|
2013-07-29 09:04:59 +00:00
|
|
|
|
|
|
|
if (i + step < re_len && is_quantifier(re + i + step)) {
|
2013-10-01 07:19:01 +00:00
|
|
|
DBG(("QUANTIFIER: [%.*s]%c [%.*s]\n", step, re + i,
|
|
|
|
re[i + step], s_len - j, s + j));
|
2013-07-29 09:04:59 +00:00
|
|
|
if (re[i + step] == '?') {
|
2013-09-30 18:30:50 +00:00
|
|
|
int result = bar(re + i, step, s + j, s_len - j, info, bi);
|
|
|
|
j += result > 0 ? result : 0;
|
2013-07-29 09:04:59 +00:00
|
|
|
i++;
|
2013-07-31 07:56:42 +00:00
|
|
|
} else if (re[i + step] == '+' || re[i + step] == '*') {
|
2013-10-01 07:19:01 +00:00
|
|
|
int j2 = j, nj = j, n1, n2 = -1, ni, non_greedy = 0;
|
2013-07-29 09:04:59 +00:00
|
|
|
|
|
|
|
/* Points to the regexp code after the quantifier */
|
2013-08-03 18:42:11 +00:00
|
|
|
ni = i + step + 1;
|
|
|
|
if (ni < re_len && re[ni] == '?') {
|
2013-07-31 07:56:42 +00:00
|
|
|
non_greedy = 1;
|
2013-08-03 18:42:11 +00:00
|
|
|
ni++;
|
2013-07-31 07:56:42 +00:00
|
|
|
}
|
2013-07-29 09:04:59 +00:00
|
|
|
|
2013-10-17 14:30:46 +00:00
|
|
|
do {
|
2014-02-08 10:09:02 +00:00
|
|
|
if ((n1 = bar(re + i, step, s + j2, s_len - j2, info, bi)) > 0) {
|
|
|
|
j2 += n1;
|
|
|
|
}
|
2013-10-17 14:30:46 +00:00
|
|
|
if (re[i + step] == '+' && n1 < 0) break;
|
|
|
|
|
2013-07-29 09:04:59 +00:00
|
|
|
if (ni >= re_len) {
|
|
|
|
/* After quantifier, there is nothing */
|
2013-10-17 14:30:46 +00:00
|
|
|
nj = j2;
|
|
|
|
} else if ((n2 = bar(re + ni, re_len - ni, s + j2,
|
|
|
|
s_len - j2, info, bi)) >= 0) {
|
|
|
|
/* Regex after quantifier matched */
|
|
|
|
nj = j2 + n2;
|
2013-07-29 09:04:59 +00:00
|
|
|
}
|
2013-09-29 20:25:14 +00:00
|
|
|
if (nj > j && non_greedy) break;
|
2013-10-17 14:30:46 +00:00
|
|
|
} while (n1 > 0);
|
|
|
|
|
2015-05-19 13:30:48 +00:00
|
|
|
/*
|
|
|
|
* Even if we found one or more pattern, this branch will be executed,
|
|
|
|
* changing the next captures.
|
|
|
|
*/
|
2015-05-19 16:08:58 +00:00
|
|
|
if (n1 < 0 && n2 < 0 && re[i + step] == '*' &&
|
2014-02-08 10:09:02 +00:00
|
|
|
(n2 = bar(re + ni, re_len - ni, s + j, s_len - j, info, bi)) > 0) {
|
|
|
|
nj = j + n2;
|
|
|
|
}
|
|
|
|
|
|
|
|
DBG(("STAR/PLUS END: %d %d %d %d %d\n", j, nj, re_len - ni, n1, n2));
|
2013-10-01 11:13:00 +00:00
|
|
|
FAIL_IF(re[i + step] == '+' && nj == j, SLRE_NO_MATCH);
|
2013-09-30 09:52:57 +00:00
|
|
|
|
2013-10-01 07:19:01 +00:00
|
|
|
/* If while loop body above was not executed for the * quantifier, */
|
|
|
|
/* make sure the rest of the regex matches */
|
2013-10-01 11:13:00 +00:00
|
|
|
FAIL_IF(nj == j && ni < re_len && n2 < 0, SLRE_NO_MATCH);
|
2013-10-01 07:19:01 +00:00
|
|
|
|
2013-09-30 09:52:57 +00:00
|
|
|
/* Returning here cause we've matched the rest of RE already */
|
2013-07-29 09:04:59 +00:00
|
|
|
return nj;
|
|
|
|
}
|
2013-09-29 20:25:14 +00:00
|
|
|
continue;
|
2013-07-29 09:04:59 +00:00
|
|
|
}
|
|
|
|
|
2013-08-03 18:42:11 +00:00
|
|
|
if (re[i] == '[') {
|
|
|
|
n = match_set(re + i + 1, re_len - (i + 2), s + j, info);
|
|
|
|
DBG(("SET %.*s [%.*s] -> %d\n", step, re + i, s_len - j, s + j, n));
|
2013-10-01 11:13:00 +00:00
|
|
|
FAIL_IF(n <= 0, SLRE_NO_MATCH);
|
2013-08-03 18:42:11 +00:00
|
|
|
j += n;
|
|
|
|
} else if (re[i] == '(') {
|
2014-02-24 11:21:55 +00:00
|
|
|
n = SLRE_NO_MATCH;
|
2013-08-03 18:42:11 +00:00
|
|
|
bi++;
|
2013-10-01 11:13:00 +00:00
|
|
|
FAIL_IF(bi >= info->num_brackets, SLRE_INTERNAL_ERROR);
|
2014-02-08 10:09:02 +00:00
|
|
|
DBG(("CAPTURING [%.*s] [%.*s] [%s]\n",
|
|
|
|
step, re + i, s_len - j, s + j, re + i + step));
|
|
|
|
|
|
|
|
if (re_len - (i + step) <= 0) {
|
|
|
|
/* Nothing follows brackets */
|
|
|
|
n = doh(s + j, s_len - j, info, bi);
|
|
|
|
} else {
|
|
|
|
int j2;
|
2014-02-24 11:21:55 +00:00
|
|
|
for (j2 = 0; j2 <= s_len - j; j2++) {
|
2014-02-08 10:09:02 +00:00
|
|
|
if ((n = doh(s + j, s_len - (j + j2), info, bi)) >= 0 &&
|
|
|
|
bar(re + i + step, re_len - (i + step),
|
|
|
|
s + j + n, s_len - (j + n), info, bi) >= 0) break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2013-08-03 18:42:11 +00:00
|
|
|
DBG(("CAPTURED [%.*s] [%.*s]:%d\n", step, re + i, s_len - j, s + j, n));
|
2013-10-01 11:13:00 +00:00
|
|
|
FAIL_IF(n < 0, n);
|
2015-05-19 13:30:48 +00:00
|
|
|
if (info->caps != NULL && n > 0) {
|
2013-08-07 10:01:28 +00:00
|
|
|
info->caps[bi - 1].ptr = s + j;
|
|
|
|
info->caps[bi - 1].len = n;
|
2013-08-03 18:42:11 +00:00
|
|
|
}
|
|
|
|
j += n;
|
|
|
|
} else if (re[i] == '^') {
|
2013-10-01 11:13:00 +00:00
|
|
|
FAIL_IF(j != 0, SLRE_NO_MATCH);
|
2013-09-28 01:06:18 +00:00
|
|
|
} else if (re[i] == '$') {
|
2013-10-01 11:13:00 +00:00
|
|
|
FAIL_IF(j != s_len, SLRE_NO_MATCH);
|
2013-08-03 18:42:11 +00:00
|
|
|
} else {
|
2013-10-01 11:13:00 +00:00
|
|
|
FAIL_IF(j >= s_len, SLRE_NO_MATCH);
|
2016-10-27 15:09:34 +00:00
|
|
|
n = match_op((const unsigned char *) (re + i), (const unsigned char *) (s + j), info);
|
2013-10-01 11:13:00 +00:00
|
|
|
FAIL_IF(n <= 0, n);
|
2013-08-03 18:42:11 +00:00
|
|
|
j += n;
|
2013-07-29 09:04:59 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
return j;
|
|
|
|
}
|
|
|
|
|
2013-08-01 08:14:09 +00:00
|
|
|
/* Process branch points */
|
2013-08-07 10:01:28 +00:00
|
|
|
static int doh(const char *s, int s_len, struct regex_info *info, int bi) {
|
2013-08-01 08:14:09 +00:00
|
|
|
const struct bracket_pair *b = &info->brackets[bi];
|
|
|
|
int i = 0, len, result;
|
|
|
|
const char *p;
|
|
|
|
|
|
|
|
do {
|
|
|
|
p = i == 0 ? b->ptr : info->branches[b->branches + i - 1].schlong + 1;
|
|
|
|
len = b->num_branches == 0 ? b->len :
|
2014-08-25 08:10:46 +00:00
|
|
|
i == b->num_branches ? (int) (b->ptr + b->len - p) :
|
|
|
|
(int) (info->branches[b->branches + i].schlong - p);
|
2013-10-01 07:19:01 +00:00
|
|
|
DBG(("%s %d %d [%.*s] [%.*s]\n", __func__, bi, i, len, p, s_len, s));
|
2013-08-07 10:01:28 +00:00
|
|
|
result = bar(p, len, s, s_len, info, bi);
|
2013-10-01 07:19:01 +00:00
|
|
|
DBG(("%s <- %d\n", __func__, result));
|
2013-09-29 20:25:14 +00:00
|
|
|
} while (result <= 0 && i++ < b->num_branches); /* At least 1 iteration */
|
2013-08-01 08:14:09 +00:00
|
|
|
|
|
|
|
return result;
|
|
|
|
}
|
|
|
|
|
2013-09-30 09:52:57 +00:00
|
|
|
static int baz(const char *s, int s_len, struct regex_info *info) {
|
2013-09-30 18:30:50 +00:00
|
|
|
int i, result = -1, is_anchored = info->brackets[0].ptr[0] == '^';
|
2013-09-30 09:52:57 +00:00
|
|
|
|
2013-09-30 18:30:50 +00:00
|
|
|
for (i = 0; i <= s_len; i++) {
|
2013-09-30 09:52:57 +00:00
|
|
|
result = doh(s + i, s_len - i, info, 0);
|
2013-09-30 18:30:50 +00:00
|
|
|
if (result >= 0) {
|
2013-09-30 09:52:57 +00:00
|
|
|
result += i;
|
|
|
|
break;
|
|
|
|
}
|
2013-09-30 18:30:50 +00:00
|
|
|
if (is_anchored) break;
|
2013-09-30 09:52:57 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
return result;
|
|
|
|
}
|
|
|
|
|
2013-08-01 08:14:09 +00:00
|
|
|
static void setup_branch_points(struct regex_info *info) {
|
|
|
|
int i, j;
|
|
|
|
struct branch tmp;
|
|
|
|
|
|
|
|
/* First, sort branches. Must be stable, no qsort. Use bubble algo. */
|
|
|
|
for (i = 0; i < info->num_branches; i++) {
|
|
|
|
for (j = i + 1; j < info->num_branches; j++) {
|
|
|
|
if (info->branches[i].bracket_index > info->branches[j].bracket_index) {
|
|
|
|
tmp = info->branches[i];
|
|
|
|
info->branches[i] = info->branches[j];
|
|
|
|
info->branches[j] = tmp;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* For each bracket, set their branch points. This way, for every bracket
|
|
|
|
* (i.e. every chunk of regex) we know all branch points before matching.
|
|
|
|
*/
|
|
|
|
for (i = j = 0; i < info->num_brackets; i++) {
|
|
|
|
info->brackets[i].num_branches = 0;
|
|
|
|
info->brackets[i].branches = j;
|
|
|
|
while (j < info->num_branches && info->branches[j].bracket_index == i) {
|
|
|
|
info->brackets[i].num_branches++;
|
|
|
|
j++;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static int foo(const char *re, int re_len, const char *s, int s_len,
|
2013-08-07 10:01:28 +00:00
|
|
|
struct regex_info *info) {
|
2013-09-30 09:52:57 +00:00
|
|
|
int i, step, depth = 0;
|
2013-07-29 09:04:59 +00:00
|
|
|
|
2013-08-01 08:14:09 +00:00
|
|
|
/* First bracket captures everything */
|
|
|
|
info->brackets[0].ptr = re;
|
|
|
|
info->brackets[0].len = re_len;
|
|
|
|
info->num_brackets = 1;
|
2013-07-29 09:04:59 +00:00
|
|
|
|
2013-08-01 08:14:09 +00:00
|
|
|
/* Make a single pass over regex string, memorize brackets and branches */
|
2013-07-29 09:04:59 +00:00
|
|
|
for (i = 0; i < re_len; i += step) {
|
2013-08-03 18:42:11 +00:00
|
|
|
step = get_op_len(re + i, re_len - i);
|
2013-07-29 09:04:59 +00:00
|
|
|
|
|
|
|
if (re[i] == '|') {
|
2014-08-25 07:20:56 +00:00
|
|
|
FAIL_IF(info->num_branches >= (int) ARRAY_SIZE(info->branches),
|
2013-10-01 11:13:00 +00:00
|
|
|
SLRE_TOO_MANY_BRANCHES);
|
2013-08-01 08:14:09 +00:00
|
|
|
info->branches[info->num_branches].bracket_index =
|
|
|
|
info->brackets[info->num_brackets - 1].len == -1 ?
|
|
|
|
info->num_brackets - 1 : depth;
|
2013-07-29 09:04:59 +00:00
|
|
|
info->branches[info->num_branches].schlong = &re[i];
|
|
|
|
info->num_branches++;
|
2013-09-30 18:30:50 +00:00
|
|
|
} else if (re[i] == '\\') {
|
2013-10-01 11:13:00 +00:00
|
|
|
FAIL_IF(i >= re_len - 1, SLRE_INVALID_METACHARACTER);
|
2013-09-30 18:30:50 +00:00
|
|
|
if (re[i + 1] == 'x') {
|
|
|
|
/* Hex digit specification must follow */
|
|
|
|
FAIL_IF(re[i + 1] == 'x' && i >= re_len - 3,
|
2013-10-01 11:13:00 +00:00
|
|
|
SLRE_INVALID_METACHARACTER);
|
2013-09-30 18:30:50 +00:00
|
|
|
FAIL_IF(re[i + 1] == 'x' && !(isxdigit(re[i + 2]) &&
|
2013-10-01 11:13:00 +00:00
|
|
|
isxdigit(re[i + 3])), SLRE_INVALID_METACHARACTER);
|
2013-09-30 18:30:50 +00:00
|
|
|
} else {
|
2016-10-27 15:09:34 +00:00
|
|
|
FAIL_IF(!is_metacharacter((const unsigned char *) re + i + 1),
|
2013-10-01 11:13:00 +00:00
|
|
|
SLRE_INVALID_METACHARACTER);
|
2013-09-30 18:30:50 +00:00
|
|
|
}
|
2013-07-29 09:04:59 +00:00
|
|
|
} else if (re[i] == '(') {
|
2014-08-25 07:20:56 +00:00
|
|
|
FAIL_IF(info->num_brackets >= (int) ARRAY_SIZE(info->brackets),
|
2013-10-01 11:13:00 +00:00
|
|
|
SLRE_TOO_MANY_BRACKETS);
|
2013-07-29 09:04:59 +00:00
|
|
|
depth++; /* Order is important here. Depth increments first. */
|
2013-08-01 08:14:09 +00:00
|
|
|
info->brackets[info->num_brackets].ptr = re + i + 1;
|
|
|
|
info->brackets[info->num_brackets].len = -1;
|
|
|
|
info->num_brackets++;
|
2013-08-07 10:01:28 +00:00
|
|
|
FAIL_IF(info->num_caps > 0 && info->num_brackets - 1 > info->num_caps,
|
2013-10-01 11:13:00 +00:00
|
|
|
SLRE_CAPS_ARRAY_TOO_SMALL);
|
2013-07-29 09:04:59 +00:00
|
|
|
} else if (re[i] == ')') {
|
2013-08-01 08:14:09 +00:00
|
|
|
int ind = info->brackets[info->num_brackets - 1].len == -1 ?
|
|
|
|
info->num_brackets - 1 : depth;
|
2014-08-25 08:10:46 +00:00
|
|
|
info->brackets[ind].len = (int) (&re[i] - info->brackets[ind].ptr);
|
2013-08-01 08:14:09 +00:00
|
|
|
DBG(("SETTING BRACKET %d [%.*s]\n",
|
|
|
|
ind, info->brackets[ind].len, info->brackets[ind].ptr));
|
2013-07-29 09:04:59 +00:00
|
|
|
depth--;
|
2013-10-01 11:13:00 +00:00
|
|
|
FAIL_IF(depth < 0, SLRE_UNBALANCED_BRACKETS);
|
|
|
|
FAIL_IF(i > 0 && re[i - 1] == '(', SLRE_NO_MATCH);
|
2013-07-29 09:04:59 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2013-10-01 11:13:00 +00:00
|
|
|
FAIL_IF(depth != 0, SLRE_UNBALANCED_BRACKETS);
|
2013-08-01 08:14:09 +00:00
|
|
|
setup_branch_points(info);
|
|
|
|
|
2013-09-30 09:52:57 +00:00
|
|
|
return baz(s, s_len, info);
|
2013-07-29 09:04:59 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
int slre_match(const char *regexp, const char *s, int s_len,
|
2014-07-14 10:32:14 +00:00
|
|
|
struct slre_cap *caps, int num_caps, int flags) {
|
2013-07-29 09:04:59 +00:00
|
|
|
struct regex_info info;
|
|
|
|
|
2013-08-01 08:14:09 +00:00
|
|
|
/* Initialize info structure */
|
2014-07-14 10:32:14 +00:00
|
|
|
info.flags = flags;
|
|
|
|
info.num_brackets = info.num_branches = 0;
|
2013-08-07 10:01:28 +00:00
|
|
|
info.num_caps = num_caps;
|
|
|
|
info.caps = caps;
|
2013-07-29 09:04:59 +00:00
|
|
|
|
2013-07-29 22:24:41 +00:00
|
|
|
DBG(("========================> [%s] [%.*s]\n", regexp, s_len, s));
|
2014-08-25 08:10:46 +00:00
|
|
|
return foo(regexp, (int) strlen(regexp), s, s_len, &info);
|
2013-07-29 09:04:59 +00:00
|
|
|
}
|