xref: /openbsd/bin/ed/re.c (revision 5af055cd)
1 /*	$OpenBSD: re.c,v 1.16 2015/10/09 21:24:05 tobias Exp $	*/
2 /*	$NetBSD: re.c,v 1.14 1995/03/21 09:04:48 cgd Exp $	*/
3 
4 /* re.c: This file contains the regular expression interface routines for
5    the ed line editor. */
6 /*-
7  * Copyright (c) 1993 Andrew Moore, Talke Studio.
8  * All rights reserved.
9  *
10  * Redistribution and use in source and binary forms, with or without
11  * modification, are permitted provided that the following conditions
12  * are met:
13  * 1. Redistributions of source code must retain the above copyright
14  *    notice, this list of conditions and the following disclaimer.
15  * 2. Redistributions in binary form must reproduce the above copyright
16  *    notice, this list of conditions and the following disclaimer in the
17  *    documentation and/or other materials provided with the distribution.
18  *
19  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
20  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
21  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
22  * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
23  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
24  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
25  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
26  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
27  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
28  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
29  * SUCH DAMAGE.
30  */
31 
32 #include "ed.h"
33 
34 static char *extract_pattern(int);
35 static char *parse_char_class(char *);
36 
37 extern int patlock;
38 
39 
40 /* get_compiled_pattern: return pointer to compiled pattern from command
41    buffer */
42 regex_t *
43 get_compiled_pattern(void)
44 {
45 	static regex_t *exp = NULL;
46 	char errbuf[128] = "";
47 
48 	char *exps;
49 	char delimiter;
50 	int n;
51 
52 	if ((delimiter = *ibufp) == ' ') {
53 		seterrmsg("invalid pattern delimiter");
54 		return NULL;
55 	} else if (delimiter == '\n' || *++ibufp == '\n' || *ibufp == delimiter) {
56 		if (!exp)
57 			seterrmsg("no previous pattern");
58 		return exp;
59 	} else if ((exps = extract_pattern(delimiter)) == NULL)
60 		return NULL;
61 	/* buffer alloc'd && not reserved */
62 	if (exp && !patlock)
63 		regfree(exp);
64 	else if ((exp = malloc(sizeof(regex_t))) == NULL) {
65 		perror(NULL);
66 		seterrmsg("out of memory");
67 		return NULL;
68 	}
69 	patlock = 0;
70 	if ((n = regcomp(exp, exps, 0)) != 0) {
71 		regerror(n, exp, errbuf, sizeof errbuf);
72 		seterrmsg(errbuf);
73 		free(exp);
74 		return exp = NULL;
75 	}
76 	return exp;
77 }
78 
79 
80 /* extract_pattern: copy a pattern string from the command buffer; return
81    pointer to the copy */
82 static char *
83 extract_pattern(int delimiter)
84 {
85 	static char *lhbuf = NULL;	/* buffer */
86 	static int lhbufsz = 0;		/* buffer size */
87 
88 	char *nd;
89 	int len;
90 
91 	for (nd = ibufp; *nd != delimiter && *nd != '\n'; nd++)
92 		switch (*nd) {
93 		default:
94 			break;
95 		case '[':
96 			if ((nd = parse_char_class(++nd)) == NULL) {
97 				seterrmsg("unbalanced brackets ([])");
98 				return NULL;
99 			}
100 			break;
101 		case '\\':
102 			if (*++nd == '\n') {
103 				seterrmsg("trailing backslash (\\)");
104 				return NULL;
105 			}
106 			break;
107 		}
108 	len = nd - ibufp;
109 	REALLOC(lhbuf, lhbufsz, len + 1, NULL);
110 	memcpy(lhbuf, ibufp, len);
111 	lhbuf[len] = '\0';
112 	ibufp = nd;
113 	return (isbinary) ? NUL_TO_NEWLINE(lhbuf, len) : lhbuf;
114 }
115 
116 
117 /* parse_char_class: expand a POSIX character class */
118 static char *
119 parse_char_class(char *s)
120 {
121 	int c, d;
122 
123 	if (*s == '^')
124 		s++;
125 	if (*s == ']')
126 		s++;
127 	for (; *s != ']' && *s != '\n'; s++)
128 		if (*s == '[' && ((d = *(s+1)) == '.' || d == ':' || d == '='))
129 			for (s++, c = *++s; *s != ']' || c != d; s++)
130 				if ((c = *s) == '\n')
131 					return NULL;
132 	return  (*s == ']') ? s : NULL;
133 }
134