| 1 |
/*************************************************
|
| 2 |
* Perl-Compatible Regular Expressions *
|
| 3 |
*************************************************/
|
| 4 |
|
| 5 |
/*
|
| 6 |
This is a library of functions to support regular expressions whose syntax
|
| 7 |
and semantics are as close as possible to those of the Perl 5 language. See
|
| 8 |
the file Tech.Notes for some information on the internals.
|
| 9 |
|
| 10 |
This module is a wrapper that provides a POSIX API to the underlying PCRE
|
| 11 |
functions.
|
| 12 |
|
| 13 |
Written by: Philip Hazel <ph10@cam.ac.uk>
|
| 14 |
|
| 15 |
Copyright (c) 1997-1999 University of Cambridge
|
| 16 |
|
| 17 |
-----------------------------------------------------------------------------
|
| 18 |
Permission is granted to anyone to use this software for any purpose on any
|
| 19 |
computer system, and to redistribute it freely, subject to the following
|
| 20 |
restrictions:
|
| 21 |
|
| 22 |
1. This software is distributed in the hope that it will be useful,
|
| 23 |
but WITHOUT ANY WARRANTY; without even the implied warranty of
|
| 24 |
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
|
| 25 |
|
| 26 |
2. The origin of this software must not be misrepresented, either by
|
| 27 |
explicit claim or by omission.
|
| 28 |
|
| 29 |
3. Altered versions must be plainly marked as such, and must not be
|
| 30 |
misrepresented as being the original software.
|
| 31 |
|
| 32 |
4. If PCRE is embedded in any software that is released under the GNU
|
| 33 |
General Purpose Licence (GPL), then the terms of that licence shall
|
| 34 |
supersede any condition above with which it is incompatible.
|
| 35 |
-----------------------------------------------------------------------------
|
| 36 |
*/
|
| 37 |
|
| 38 |
#include "internal.h"
|
| 39 |
#include "pcreposix.h"
|
| 40 |
#include "stdlib.h"
|
| 41 |
|
| 42 |
|
| 43 |
|
| 44 |
/* Corresponding tables of PCRE error messages and POSIX error codes. */
|
| 45 |
|
| 46 |
static const char *estring[] = {
|
| 47 |
ERR1, ERR2, ERR3, ERR4, ERR5, ERR6, ERR7, ERR8, ERR9, ERR10,
|
| 48 |
ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19, ERR20,
|
| 49 |
ERR21, ERR22, ERR23, ERR24, ERR25 };
|
| 50 |
|
| 51 |
static int eint[] = {
|
| 52 |
REG_EESCAPE, /* "\\ at end of pattern" */
|
| 53 |
REG_EESCAPE, /* "\\c at end of pattern" */
|
| 54 |
REG_EESCAPE, /* "unrecognized character follows \\" */
|
| 55 |
REG_BADBR, /* "numbers out of order in {} quantifier" */
|
| 56 |
REG_BADBR, /* "number too big in {} quantifier" */
|
| 57 |
REG_EBRACK, /* "missing terminating ] for character class" */
|
| 58 |
REG_ECTYPE, /* "invalid escape sequence in character class" */
|
| 59 |
REG_ERANGE, /* "range out of order in character class" */
|
| 60 |
REG_BADRPT, /* "nothing to repeat" */
|
| 61 |
REG_BADRPT, /* "operand of unlimited repeat could match the empty string" */
|
| 62 |
REG_ASSERT, /* "internal error: unexpected repeat" */
|
| 63 |
REG_BADPAT, /* "unrecognized character after (?" */
|
| 64 |
REG_ESIZE, /* "too many capturing parenthesized sub-patterns" */
|
| 65 |
REG_EPAREN, /* "missing )" */
|
| 66 |
REG_ESUBREG, /* "back reference to non-existent subpattern" */
|
| 67 |
REG_INVARG, /* "erroffset passed as NULL" */
|
| 68 |
REG_INVARG, /* "unknown option bit(s) set" */
|
| 69 |
REG_EPAREN, /* "missing ) after comment" */
|
| 70 |
REG_ESIZE, /* "too many sets of parentheses" */
|
| 71 |
REG_ESIZE, /* "regular expression too large" */
|
| 72 |
REG_ESPACE, /* "failed to get memory" */
|
| 73 |
REG_EPAREN, /* "unmatched brackets" */
|
| 74 |
REG_ASSERT, /* "internal error: code overflow" */
|
| 75 |
REG_BADPAT, /* "unrecognized character after (?<" */
|
| 76 |
REG_BADPAT, /* "lookbehind assertion is not fixed length" */
|
| 77 |
REG_BADPAT, /* "malformed number after (?(" */
|
| 78 |
REG_BADPAT, /* "conditional group containe more than two branches" */
|
| 79 |
REG_BADPAT /* "assertion expected after (?(" */
|
| 80 |
};
|
| 81 |
|
| 82 |
/* Table of texts corresponding to POSIX error codes */
|
| 83 |
|
| 84 |
static const char *pstring[] = {
|
| 85 |
"", /* Dummy for value 0 */
|
| 86 |
"internal error", /* REG_ASSERT */
|
| 87 |
"invalid repeat counts in {}", /* BADBR */
|
| 88 |
"pattern error", /* BADPAT */
|
| 89 |
"? * + invalid", /* BADRPT */
|
| 90 |
"unbalanced {}", /* EBRACE */
|
| 91 |
"unbalanced []", /* EBRACK */
|
| 92 |
"collation error - not relevant", /* ECOLLATE */
|
| 93 |
"bad class", /* ECTYPE */
|
| 94 |
"bad escape sequence", /* EESCAPE */
|
| 95 |
"empty expression", /* EMPTY */
|
| 96 |
"unbalanced ()", /* EPAREN */
|
| 97 |
"bad range inside []", /* ERANGE */
|
| 98 |
"expression too big", /* ESIZE */
|
| 99 |
"failed to get memory", /* ESPACE */
|
| 100 |
"bad back reference", /* ESUBREG */
|
| 101 |
"bad argument", /* INVARG */
|
| 102 |
"match failed" /* NOMATCH */
|
| 103 |
};
|
| 104 |
|
| 105 |
|
| 106 |
|
| 107 |
|
| 108 |
/*************************************************
|
| 109 |
* Translate PCRE text code to int *
|
| 110 |
*************************************************/
|
| 111 |
|
| 112 |
/* PCRE compile-time errors are given as strings defined as macros. We can just
|
| 113 |
look them up in a table to turn them into POSIX-style error codes. */
|
| 114 |
|
| 115 |
static int
|
| 116 |
pcre_posix_error_code(const char *s)
|
| 117 |
{
|
| 118 |
size_t i;
|
| 119 |
for (i = 0; i < sizeof(estring)/sizeof(char *); i++)
|
| 120 |
if (strcmp(s, estring[i]) == 0) return eint[i];
|
| 121 |
return REG_ASSERT;
|
| 122 |
}
|
| 123 |
|
| 124 |
|
| 125 |
|
| 126 |
/*************************************************
|
| 127 |
* Translate error code to string *
|
| 128 |
*************************************************/
|
| 129 |
|
| 130 |
size_t
|
| 131 |
regerror(int errcode, const regex_t *preg, char *errbuf, size_t errbuf_size)
|
| 132 |
{
|
| 133 |
const char *message, *addmessage;
|
| 134 |
size_t length, addlength;
|
| 135 |
|
| 136 |
message = (errcode >= (int)(sizeof(pstring)/sizeof(char *)))?
|
| 137 |
"unknown error code" : pstring[errcode];
|
| 138 |
length = strlen(message) + 1;
|
| 139 |
|
| 140 |
addmessage = " at offset ";
|
| 141 |
addlength = (preg != NULL && (int)preg->re_erroffset != -1)?
|
| 142 |
strlen(addmessage) + 6 : 0;
|
| 143 |
|
| 144 |
if (errbuf_size > 0)
|
| 145 |
{
|
| 146 |
if (addlength > 0 && errbuf_size >= length + addlength)
|
| 147 |
sprintf(errbuf, "%s%s%-6d", message, addmessage, (int)preg->re_erroffset);
|
| 148 |
else
|
| 149 |
{
|
| 150 |
strncpy(errbuf, message, errbuf_size - 1);
|
| 151 |
errbuf[errbuf_size-1] = 0;
|
| 152 |
}
|
| 153 |
}
|
| 154 |
|
| 155 |
return length + addlength;
|
| 156 |
}
|
| 157 |
|
| 158 |
|
| 159 |
|
| 160 |
|
| 161 |
/*************************************************
|
| 162 |
* Free store held by a regex *
|
| 163 |
*************************************************/
|
| 164 |
|
| 165 |
void
|
| 166 |
regfree(regex_t *preg)
|
| 167 |
{
|
| 168 |
(pcre_free)(preg->re_pcre);
|
| 169 |
}
|
| 170 |
|
| 171 |
|
| 172 |
|
| 173 |
|
| 174 |
/*************************************************
|
| 175 |
* Compile a regular expression *
|
| 176 |
*************************************************/
|
| 177 |
|
| 178 |
/*
|
| 179 |
Arguments:
|
| 180 |
preg points to a structure for recording the compiled expression
|
| 181 |
pattern the pattern to compile
|
| 182 |
cflags compilation flags
|
| 183 |
|
| 184 |
Returns: 0 on success
|
| 185 |
various non-zero codes on failure
|
| 186 |
*/
|
| 187 |
|
| 188 |
int
|
| 189 |
regcomp(regex_t *preg, const char *pattern, int cflags)
|
| 190 |
{
|
| 191 |
const char *errorptr;
|
| 192 |
int erroffset;
|
| 193 |
int options = 0;
|
| 194 |
|
| 195 |
if ((cflags & REG_ICASE) != 0) options |= PCRE_CASELESS;
|
| 196 |
if ((cflags & REG_NEWLINE) != 0) options |= PCRE_MULTILINE;
|
| 197 |
|
| 198 |
preg->re_pcre = pcre_compile(pattern, options, &errorptr, &erroffset, NULL);
|
| 199 |
preg->re_erroffset = erroffset;
|
| 200 |
|
| 201 |
if (preg->re_pcre == NULL) return pcre_posix_error_code(errorptr);
|
| 202 |
|
| 203 |
preg->re_nsub = pcre_info(preg->re_pcre, NULL, NULL);
|
| 204 |
return 0;
|
| 205 |
}
|
| 206 |
|
| 207 |
|
| 208 |
|
| 209 |
|
| 210 |
/*************************************************
|
| 211 |
* Match a regular expression *
|
| 212 |
*************************************************/
|
| 213 |
|
| 214 |
int
|
| 215 |
regexec(regex_t *preg, const char *string, size_t nmatch,
|
| 216 |
regmatch_t pmatch[], int eflags)
|
| 217 |
{
|
| 218 |
int rc;
|
| 219 |
int options = 0;
|
| 220 |
|
| 221 |
if ((eflags & REG_NOTBOL) != 0) options |= PCRE_NOTBOL;
|
| 222 |
if ((eflags & REG_NOTEOL) != 0) options |= PCRE_NOTEOL;
|
| 223 |
|
| 224 |
preg->re_erroffset = (size_t)(-1); /* Only has meaning after compile */
|
| 225 |
|
| 226 |
rc = pcre_exec(preg->re_pcre, NULL, string, (int)strlen(string), 0, options,
|
| 227 |
(int *)pmatch, nmatch * 2);
|
| 228 |
|
| 229 |
if (rc == 0) return 0; /* All pmatch were filled in */
|
| 230 |
|
| 231 |
if (rc > 0)
|
| 232 |
{
|
| 233 |
size_t i;
|
| 234 |
for (i = rc; i < nmatch; i++) pmatch[i].rm_so = pmatch[i].rm_eo = -1;
|
| 235 |
return 0;
|
| 236 |
}
|
| 237 |
|
| 238 |
else switch(rc)
|
| 239 |
{
|
| 240 |
case PCRE_ERROR_NOMATCH: return REG_NOMATCH;
|
| 241 |
case PCRE_ERROR_NULL: return REG_INVARG;
|
| 242 |
case PCRE_ERROR_BADOPTION: return REG_INVARG;
|
| 243 |
case PCRE_ERROR_BADMAGIC: return REG_INVARG;
|
| 244 |
case PCRE_ERROR_UNKNOWN_NODE: return REG_ASSERT;
|
| 245 |
case PCRE_ERROR_NOMEMORY: return REG_ESPACE;
|
| 246 |
default: return REG_ASSERT;
|
| 247 |
}
|
| 248 |
}
|
| 249 |
|
| 250 |
/* End of pcreposix.c */
|
Parent Directory
| 
Revision Log