/[pcre]/code/branches/pcre16/pcreposix.c
ViewVC logotype

Contents of /code/branches/pcre16/pcreposix.c

Parent Directory Parent Directory | Revision Log Revision Log


Revision 3 - (show annotations) (download)
Sat Feb 24 21:38:01 2007 UTC (7 years, 7 months ago) by nigel
Original Path: code/trunk/pcreposix.c
File MIME type: text/plain
File size: 7546 byte(s)
Load pcre-1.00 into code/trunk.

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 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 */
33
34 #include "internal.h"
35 #include "pcreposix.h"
36 #include "stdlib.h"
37
38
39
40 /* Corresponding tables of PCRE error messages and POSIX error codes. */
41
42 static char *estring[] = {
43 ERR1, ERR2, ERR3, ERR4, ERR5, ERR6, ERR7, ERR8, ERR9, ERR10,
44 ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19, ERR20,
45 ERR21, ERR22, ERR23 };
46
47 static int eint[] = {
48 REG_EESCAPE, /* "\\ at end of pattern" */
49 REG_EESCAPE, /* "\\c at end of pattern" */
50 REG_EESCAPE, /* "unrecognized character follows \\" */
51 REG_BADBR, /* "numbers out of order in {} quantifier" */
52 REG_BADBR, /* "number too big in {} quantifier" */
53 REG_EBRACK, /* "missing terminating ] for character class" */
54 REG_ECTYPE, /* "invalid escape sequence in character class" */
55 REG_ERANGE, /* "range out of order in character class" */
56 REG_BADRPT, /* "nothing to repeat" */
57 REG_BADRPT, /* "operand of unlimited repeat could match the empty string" */
58 REG_ASSERT, /* "internal error: unexpected repeat" */
59 REG_BADPAT, /* "unrecognized character after (?" */
60 REG_ESIZE, /* "too many capturing parenthesized sub-patterns" */
61 REG_EPAREN, /* "missing )" */
62 REG_ESUBREG, /* "back reference to non-existent subpattern" */
63 REG_INVARG, /* "erroffset passed as NULL" */
64 REG_INVARG, /* "unknown option bit(s) set" */
65 REG_EPAREN, /* "missing ) after comment" */
66 REG_ESIZE, /* "too many sets of parentheses" */
67 REG_ESIZE, /* "regular expression too large" */
68 REG_ESPACE, /* "failed to get memory" */
69 REG_EPAREN, /* "unmatched brackets" */
70 REG_ASSERT /* "internal error: code overflow" */
71 };
72
73 /* Table of texts corresponding to POSIX error codes */
74
75 static char *pstring[] = {
76 "", /* Dummy for value 0 */
77 "internal error", /* REG_ASSERT */
78 "invalid repeat counts in {}", /* BADBR */
79 "pattern error", /* BADPAT */
80 "? * + invalid", /* BADRPT */
81 "unbalanced {}", /* EBRACE */
82 "unbalanced []", /* EBRACK */
83 "collation error - not relevant", /* ECOLLATE */
84 "bad class", /* ECTYPE */
85 "bad escape sequence", /* EESCAPE */
86 "empty expression", /* EMPTY */
87 "unbalanced ()", /* EPAREN */
88 "bad range inside []", /* ERANGE */
89 "expression too big", /* ESIZE */
90 "failed to get memory", /* ESPACE */
91 "bad back reference", /* ESUBREG */
92 "bad argument", /* INVARG */
93 "match failed" /* NOMATCH */
94 };
95
96
97
98
99 /*************************************************
100 * Translate PCRE text code to int *
101 *************************************************/
102
103 /* PCRE compile-time errors are given as strings defined as macros. We can just
104 look them up in a table to turn them into POSIX-style error codes. */
105
106 static int
107 pcre_posix_error_code(const char *s)
108 {
109 int i;
110 for (i = 0; i < sizeof(estring)/sizeof(char *); i++)
111 if (strcmp(s, estring[i]) == 0) return eint[i];
112 return REG_ASSERT;
113 }
114
115
116
117 /*************************************************
118 * Translate error code to string *
119 *************************************************/
120
121 size_t
122 regerror(int errcode, const regex_t *preg, char *errbuf, size_t errbuf_size)
123 {
124 char *message, *addmessage;
125 int length, adlength;
126
127 message = (errcode >= sizeof(pstring)/sizeof(char *))?
128 "unknown error code" : pstring[errcode];
129
130 length = (int)strlen(message) + 1;
131
132 if (preg != NULL && (int)preg->re_erroffset != -1)
133 {
134 addmessage = " at offset ";
135 adlength = (int)strlen(addmessage) + 6;
136 }
137 else adlength = 0;
138
139 if (errbuf_size > 0)
140 {
141 if (adlength > 0 && errbuf_size >= length + adlength)
142 sprintf(errbuf, "%s%s%-6d", message, addmessage, preg->re_erroffset);
143 else
144 {
145 strncpy(errbuf, message, errbuf_size - 1);
146 errbuf[errbuf_size-1] = 0;
147 }
148 }
149
150 return length + adlength;
151 }
152
153
154
155
156 /*************************************************
157 * Free store held by a regex *
158 *************************************************/
159
160 void
161 regfree(regex_t *preg)
162 {
163 (pcre_free)(preg->re_pcre);
164 }
165
166
167
168
169 /*************************************************
170 * Compile a regular expression *
171 *************************************************/
172
173 /*
174 Arguments:
175 preg points to a structure for recording the compiled expression
176 pattern the pattern to compile
177 cflags compilation flags
178
179 Returns: 0 on success
180 various non-zero codes on failure
181 */
182
183 int
184 regcomp(regex_t *preg, const char *pattern, int cflags)
185 {
186 char *errorptr;
187 int erroffset;
188 int options = 0;
189
190 if ((cflags & REG_ICASE) != 0) options |= PCRE_CASELESS;
191 if ((cflags & REG_NEWLINE) != 0) options |= PCRE_MULTILINE;
192
193 preg->re_pcre = pcre_compile(pattern, options, &errorptr, &erroffset);
194 preg->re_erroffset = erroffset;
195
196 if (preg->re_pcre == NULL) return pcre_posix_error_code(errorptr);
197
198 preg->re_nsub = pcre_info(preg->re_pcre, NULL, NULL);
199 return 0;
200 }
201
202
203
204
205 /*************************************************
206 * Match a regular expression *
207 *************************************************/
208
209 int
210 regexec(regex_t *preg, const char *string, size_t nmatch,
211 regmatch_t pmatch[], int eflags)
212 {
213 int rc;
214 int options = 0;
215
216 if ((eflags & REG_NOTBOL) != 0) options |= PCRE_NOTBOL;
217 if ((eflags & REG_NOTEOL) != 0) options |= PCRE_NOTEOL;
218
219 preg->re_erroffset = -1; /* Only has meaning after compile */
220
221 rc = pcre_exec(preg->re_pcre, NULL, string, (int)strlen(string), options,
222 (int *)pmatch, nmatch * 2);
223
224 if (rc == 0) return 0; /* All pmatch were filled in */
225
226 if (rc > 0)
227 {
228 int i;
229 for (i = rc; i < nmatch; i++) pmatch[i].rm_so = pmatch[i].rm_eo = -1;
230 return 0;
231 }
232
233 else switch(rc)
234 {
235 case PCRE_ERROR_NOMATCH: return REG_NOMATCH;
236 case PCRE_ERROR_BADREF: return REG_ESUBREG;
237 case PCRE_ERROR_NULL: return REG_INVARG;
238 case PCRE_ERROR_BADOPTION: return REG_INVARG;
239 case PCRE_ERROR_BADMAGIC: return REG_INVARG;
240 case PCRE_ERROR_UNKNOWN_NODE: return REG_ASSERT;
241 case PCRE_ERROR_NOMEMORY: return REG_ESPACE;
242 default: return REG_ASSERT;
243 }
244 }
245
246 /* End of pcreposix.c */

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12