1*287247a8SAlexander Pyhalov /*	$NetBSD: fparseln.c,v 1.9 1999/09/20 04:48:06 lukem Exp $	*/
2*287247a8SAlexander Pyhalov 
3*287247a8SAlexander Pyhalov /*
4*287247a8SAlexander Pyhalov  * Copyright (c) 1997 Christos Zoulas.  All rights reserved.
5*287247a8SAlexander Pyhalov  *
6*287247a8SAlexander Pyhalov  * Redistribution and use in source and binary forms, with or without
7*287247a8SAlexander Pyhalov  * modification, are permitted provided that the following conditions
8*287247a8SAlexander Pyhalov  * are met:
9*287247a8SAlexander Pyhalov  * 1. Redistributions of source code must retain the above copyright
10*287247a8SAlexander Pyhalov  *    notice, this list of conditions and the following disclaimer.
11*287247a8SAlexander Pyhalov  * 2. Redistributions in binary form must reproduce the above copyright
12*287247a8SAlexander Pyhalov  *    notice, this list of conditions and the following disclaimer in the
13*287247a8SAlexander Pyhalov  *    documentation and/or other materials provided with the distribution.
14*287247a8SAlexander Pyhalov  * 3. All advertising materials mentioning features or use of this software
15*287247a8SAlexander Pyhalov  *    must display the following acknowledgement:
16*287247a8SAlexander Pyhalov  *	This product includes software developed by Christos Zoulas.
17*287247a8SAlexander Pyhalov  * 4. The name of the author may not be used to endorse or promote products
18*287247a8SAlexander Pyhalov  *    derived from this software without specific prior written permission.
19*287247a8SAlexander Pyhalov  *
20*287247a8SAlexander Pyhalov  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
21*287247a8SAlexander Pyhalov  * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
22*287247a8SAlexander Pyhalov  * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
23*287247a8SAlexander Pyhalov  * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
24*287247a8SAlexander Pyhalov  * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
25*287247a8SAlexander Pyhalov  * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
26*287247a8SAlexander Pyhalov  * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
27*287247a8SAlexander Pyhalov  * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
28*287247a8SAlexander Pyhalov  * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
29*287247a8SAlexander Pyhalov  * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
30*287247a8SAlexander Pyhalov  */
31*287247a8SAlexander Pyhalov 
32*287247a8SAlexander Pyhalov #include <stdio.h>
33*287247a8SAlexander Pyhalov #include <string.h>
34*287247a8SAlexander Pyhalov #include <stdlib.h>
35*287247a8SAlexander Pyhalov #include <compat.h>
36*287247a8SAlexander Pyhalov 
37*287247a8SAlexander Pyhalov static int isescaped(const char *, const char *, int);
38*287247a8SAlexander Pyhalov 
39*287247a8SAlexander Pyhalov /*
40*287247a8SAlexander Pyhalov  * isescaped():
41*287247a8SAlexander Pyhalov  *	Return true if the character in *p that belongs to a string
42*287247a8SAlexander Pyhalov  *	that starts in *sp, is escaped by the escape character esc.
43*287247a8SAlexander Pyhalov  */
44*287247a8SAlexander Pyhalov static int
isescaped(const char * sp,const char * p,int esc)45*287247a8SAlexander Pyhalov isescaped(const char *sp, const char *p, int esc)
46*287247a8SAlexander Pyhalov {
47*287247a8SAlexander Pyhalov 	const char	*cp;
48*287247a8SAlexander Pyhalov 	size_t		ne;
49*287247a8SAlexander Pyhalov 
50*287247a8SAlexander Pyhalov 	/* No escape character */
51*287247a8SAlexander Pyhalov 	if (esc == '\0')
52*287247a8SAlexander Pyhalov 		return (1);
53*287247a8SAlexander Pyhalov 
54*287247a8SAlexander Pyhalov 	/* Count the number of escape characters that precede ours */
55*287247a8SAlexander Pyhalov 	for (ne = 0, cp = p; --cp >= sp && *cp == esc; ne++)
56*287247a8SAlexander Pyhalov 		continue;
57*287247a8SAlexander Pyhalov 
58*287247a8SAlexander Pyhalov 	/* Return true if odd number of escape characters */
59*287247a8SAlexander Pyhalov 	return ((ne & 1) != 0);
60*287247a8SAlexander Pyhalov }
61*287247a8SAlexander Pyhalov 
62*287247a8SAlexander Pyhalov 
63*287247a8SAlexander Pyhalov /*
64*287247a8SAlexander Pyhalov  * fparseln():
65*287247a8SAlexander Pyhalov  *	Read a line from a file parsing continuations ending in \
66*287247a8SAlexander Pyhalov  *	and eliminating trailing newlines, or comments starting with
67*287247a8SAlexander Pyhalov  *	the comment char.
68*287247a8SAlexander Pyhalov  */
69*287247a8SAlexander Pyhalov char *
fparseln(FILE * fp,size_t * size,size_t * lineno,const char str[3],int flags)70*287247a8SAlexander Pyhalov fparseln(FILE *fp, size_t *size, size_t *lineno, const char str[3], int flags)
71*287247a8SAlexander Pyhalov {
72*287247a8SAlexander Pyhalov 	static const char dstr[3] = { '\\', '\\', '#' };
73*287247a8SAlexander Pyhalov 
74*287247a8SAlexander Pyhalov 	size_t	s, len;
75*287247a8SAlexander Pyhalov 	char   *buf;
76*287247a8SAlexander Pyhalov 	char   *ptr, *cp;
77*287247a8SAlexander Pyhalov 	int	cnt;
78*287247a8SAlexander Pyhalov 	char	esc, con, nl, com;
79*287247a8SAlexander Pyhalov 
80*287247a8SAlexander Pyhalov 	len = 0;
81*287247a8SAlexander Pyhalov 	buf = NULL;
82*287247a8SAlexander Pyhalov 	cnt = 1;
83*287247a8SAlexander Pyhalov 
84*287247a8SAlexander Pyhalov 	if (str == NULL)
85*287247a8SAlexander Pyhalov 		str = dstr;
86*287247a8SAlexander Pyhalov 
87*287247a8SAlexander Pyhalov 	esc = str[0];
88*287247a8SAlexander Pyhalov 	con = str[1];
89*287247a8SAlexander Pyhalov 	com = str[2];
90*287247a8SAlexander Pyhalov 	/*
91*287247a8SAlexander Pyhalov 	 * XXX: it would be cool to be able to specify the newline character,
92*287247a8SAlexander Pyhalov 	 * but unfortunately, fgetln does not let us
93*287247a8SAlexander Pyhalov 	 */
94*287247a8SAlexander Pyhalov 	nl  = '\n';
95*287247a8SAlexander Pyhalov 
96*287247a8SAlexander Pyhalov 	while (cnt) {
97*287247a8SAlexander Pyhalov 		cnt = 0;
98*287247a8SAlexander Pyhalov 
99*287247a8SAlexander Pyhalov 		if (lineno)
100*287247a8SAlexander Pyhalov 			(*lineno)++;
101*287247a8SAlexander Pyhalov 
102*287247a8SAlexander Pyhalov 		if ((ptr = fgetln(fp, &s)) == NULL)
103*287247a8SAlexander Pyhalov 			break;
104*287247a8SAlexander Pyhalov 
105*287247a8SAlexander Pyhalov 		if (s && com) {		/* Check and eliminate comments */
106*287247a8SAlexander Pyhalov 			for (cp = ptr; cp < ptr + s; cp++)
107*287247a8SAlexander Pyhalov 				if (*cp == com && !isescaped(ptr, cp, esc)) {
108*287247a8SAlexander Pyhalov 					s = cp - ptr;
109*287247a8SAlexander Pyhalov 					cnt = s == 0 && buf == NULL;
110*287247a8SAlexander Pyhalov 					break;
111*287247a8SAlexander Pyhalov 				}
112*287247a8SAlexander Pyhalov 		}
113*287247a8SAlexander Pyhalov 
114*287247a8SAlexander Pyhalov 		if (s && nl) { 		/* Check and eliminate newlines */
115*287247a8SAlexander Pyhalov 			cp = &ptr[s - 1];
116*287247a8SAlexander Pyhalov 
117*287247a8SAlexander Pyhalov 			if (*cp == nl)
118*287247a8SAlexander Pyhalov 				s--;	/* forget newline */
119*287247a8SAlexander Pyhalov 		}
120*287247a8SAlexander Pyhalov 
121*287247a8SAlexander Pyhalov 		if (s && con) {		/* Check and eliminate continuations */
122*287247a8SAlexander Pyhalov 			cp = &ptr[s - 1];
123*287247a8SAlexander Pyhalov 
124*287247a8SAlexander Pyhalov 			if (*cp == con && !isescaped(ptr, cp, esc)) {
125*287247a8SAlexander Pyhalov 				s--;	/* forget escape */
126*287247a8SAlexander Pyhalov 				cnt = 1;
127*287247a8SAlexander Pyhalov 			}
128*287247a8SAlexander Pyhalov 		}
129*287247a8SAlexander Pyhalov 
130*287247a8SAlexander Pyhalov 		if (s == 0 && buf != NULL)
131*287247a8SAlexander Pyhalov 			continue;
132*287247a8SAlexander Pyhalov 
133*287247a8SAlexander Pyhalov 		if ((cp = realloc(buf, len + s + 1)) == NULL) {
134*287247a8SAlexander Pyhalov 			free(buf);
135*287247a8SAlexander Pyhalov 			return (NULL);
136*287247a8SAlexander Pyhalov 		}
137*287247a8SAlexander Pyhalov 		buf = cp;
138*287247a8SAlexander Pyhalov 
139*287247a8SAlexander Pyhalov 		(void) memcpy(buf + len, ptr, s);
140*287247a8SAlexander Pyhalov 		len += s;
141*287247a8SAlexander Pyhalov 		buf[len] = '\0';
142*287247a8SAlexander Pyhalov 	}
143*287247a8SAlexander Pyhalov 
144*287247a8SAlexander Pyhalov 	if ((flags & FPARSELN_UNESCALL) != 0 && esc && buf != NULL &&
145*287247a8SAlexander Pyhalov 	    strchr(buf, esc) != NULL) {
146*287247a8SAlexander Pyhalov 		ptr = cp = buf;
147*287247a8SAlexander Pyhalov 		while (cp[0] != '\0') {
148*287247a8SAlexander Pyhalov 			int skipesc;
149*287247a8SAlexander Pyhalov 
150*287247a8SAlexander Pyhalov 			while (cp[0] != '\0' && cp[0] != esc)
151*287247a8SAlexander Pyhalov 				*ptr++ = *cp++;
152*287247a8SAlexander Pyhalov 			if (cp[0] == '\0' || cp[1] == '\0')
153*287247a8SAlexander Pyhalov 				break;
154*287247a8SAlexander Pyhalov 
155*287247a8SAlexander Pyhalov 			skipesc = 0;
156*287247a8SAlexander Pyhalov 			if (cp[1] == com)
157*287247a8SAlexander Pyhalov 				skipesc += (flags & FPARSELN_UNESCCOMM);
158*287247a8SAlexander Pyhalov 			if (cp[1] == con)
159*287247a8SAlexander Pyhalov 				skipesc += (flags & FPARSELN_UNESCCONT);
160*287247a8SAlexander Pyhalov 			if (cp[1] == esc)
161*287247a8SAlexander Pyhalov 				skipesc += (flags & FPARSELN_UNESCESC);
162*287247a8SAlexander Pyhalov 			if (cp[1] != com && cp[1] != con && cp[1] != esc)
163*287247a8SAlexander Pyhalov 				skipesc = (flags & FPARSELN_UNESCREST);
164*287247a8SAlexander Pyhalov 
165*287247a8SAlexander Pyhalov 			if (skipesc)
166*287247a8SAlexander Pyhalov 				cp++;
167*287247a8SAlexander Pyhalov 			else
168*287247a8SAlexander Pyhalov 				*ptr++ = *cp++;
169*287247a8SAlexander Pyhalov 			*ptr++ = *cp++;
170*287247a8SAlexander Pyhalov 		}
171*287247a8SAlexander Pyhalov 		*ptr = '\0';
172*287247a8SAlexander Pyhalov 		len = strlen(buf);
173*287247a8SAlexander Pyhalov 	}
174*287247a8SAlexander Pyhalov 
175*287247a8SAlexander Pyhalov 	if (size) {
176*287247a8SAlexander Pyhalov 		*size = len;
177*287247a8SAlexander Pyhalov 	}
178*287247a8SAlexander Pyhalov 	return (buf);
179*287247a8SAlexander Pyhalov }
180*287247a8SAlexander Pyhalov 
181*287247a8SAlexander Pyhalov #ifdef TEST
182*287247a8SAlexander Pyhalov 
183*287247a8SAlexander Pyhalov int
main(int argc,char * argv[])184*287247a8SAlexander Pyhalov main(int argc, char *argv[])
185*287247a8SAlexander Pyhalov {
186*287247a8SAlexander Pyhalov 	char   *ptr;
187*287247a8SAlexander Pyhalov 	size_t	size, line;
188*287247a8SAlexander Pyhalov 
189*287247a8SAlexander Pyhalov 	line = 0;
190*287247a8SAlexander Pyhalov 	while ((ptr = fparseln(stdin, &size, &line, NULL,
191*287247a8SAlexander Pyhalov 	    FPARSELN_UNESCALL)) != NULL) {
192*287247a8SAlexander Pyhalov 		printf("line %d (%d) |%s|\n", line, size, ptr);
193*287247a8SAlexander Pyhalov 	}
194*287247a8SAlexander Pyhalov 	return (0);
195*287247a8SAlexander Pyhalov }
196*287247a8SAlexander Pyhalov 
197*287247a8SAlexander Pyhalov /*
198*287247a8SAlexander Pyhalov  * # This is a test
199*287247a8SAlexander Pyhalov  * line 1
200*287247a8SAlexander Pyhalov  * line 2 \
201*287247a8SAlexander Pyhalov  * line 3 # Comment
202*287247a8SAlexander Pyhalov  * line 4 \# Not comment \\\\
203*287247a8SAlexander Pyhalov  *
204*287247a8SAlexander Pyhalov  * # And a comment \
205*287247a8SAlexander Pyhalov  * line 5 \\\
206*287247a8SAlexander Pyhalov  * line 6
207*287247a8SAlexander Pyhalov  *
208*287247a8SAlexander Pyhalov  */
209*287247a8SAlexander Pyhalov 
210*287247a8SAlexander Pyhalov #endif /* TEST */
211