xref: /dflybsd-src/lib/libc/stdio/fparseln.c (revision d67f4a112e5c4403ad148406202cde8874e5ebbf)
1*d67f4a11SSascha Wildner /*	$NetBSD: fparseln.c,v 1.10 2009/10/21 01:07:45 snj Exp $	*/
2*d67f4a11SSascha Wildner 
3*d67f4a11SSascha Wildner /*
4*d67f4a11SSascha Wildner  * Copyright (c) 1997 Christos Zoulas.  All rights reserved.
5*d67f4a11SSascha Wildner  *
6*d67f4a11SSascha Wildner  * Redistribution and use in source and binary forms, with or without
7*d67f4a11SSascha Wildner  * modification, are permitted provided that the following conditions
8*d67f4a11SSascha Wildner  * are met:
9*d67f4a11SSascha Wildner  * 1. Redistributions of source code must retain the above copyright
10*d67f4a11SSascha Wildner  *    notice, this list of conditions and the following disclaimer.
11*d67f4a11SSascha Wildner  * 2. Redistributions in binary form must reproduce the above copyright
12*d67f4a11SSascha Wildner  *    notice, this list of conditions and the following disclaimer in the
13*d67f4a11SSascha Wildner  *    documentation and/or other materials provided with the distribution.
14*d67f4a11SSascha Wildner  *
15*d67f4a11SSascha Wildner  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
16*d67f4a11SSascha Wildner  * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
17*d67f4a11SSascha Wildner  * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
18*d67f4a11SSascha Wildner  * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
19*d67f4a11SSascha Wildner  * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
20*d67f4a11SSascha Wildner  * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
21*d67f4a11SSascha Wildner  * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
22*d67f4a11SSascha Wildner  * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
23*d67f4a11SSascha Wildner  * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
24*d67f4a11SSascha Wildner  * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
25*d67f4a11SSascha Wildner  */
26*d67f4a11SSascha Wildner 
27*d67f4a11SSascha Wildner #include <sys/types.h>
28*d67f4a11SSascha Wildner #include <assert.h>
29*d67f4a11SSascha Wildner #include <errno.h>
30*d67f4a11SSascha Wildner #include <stdio.h>
31*d67f4a11SSascha Wildner #include <string.h>
32*d67f4a11SSascha Wildner #include <stdlib.h>
33*d67f4a11SSascha Wildner #include <libutil.h>
34*d67f4a11SSascha Wildner 
35*d67f4a11SSascha Wildner static int isescaped(const char *, const char *, int);
36*d67f4a11SSascha Wildner 
37*d67f4a11SSascha Wildner /* isescaped():
38*d67f4a11SSascha Wildner  *	Return true if the character in *p that belongs to a string
39*d67f4a11SSascha Wildner  *	that starts in *sp, is escaped by the escape character esc.
40*d67f4a11SSascha Wildner  */
41*d67f4a11SSascha Wildner static int
isescaped(const char * sp,const char * p,int esc)42*d67f4a11SSascha Wildner isescaped(const char *sp, const char *p, int esc)
43*d67f4a11SSascha Wildner {
44*d67f4a11SSascha Wildner 	const char     *cp;
45*d67f4a11SSascha Wildner 	size_t		ne;
46*d67f4a11SSascha Wildner 
47*d67f4a11SSascha Wildner #if 0
48*d67f4a11SSascha Wildner 	_DIAGASSERT(sp != NULL);
49*d67f4a11SSascha Wildner 	_DIAGASSERT(p != NULL);
50*d67f4a11SSascha Wildner #endif
51*d67f4a11SSascha Wildner 
52*d67f4a11SSascha Wildner 	/* No escape character */
53*d67f4a11SSascha Wildner 	if (esc == '\0')
54*d67f4a11SSascha Wildner 		return 0;
55*d67f4a11SSascha Wildner 
56*d67f4a11SSascha Wildner 	/* Count the number of escape characters that precede ours */
57*d67f4a11SSascha Wildner 	for (ne = 0, cp = p; --cp >= sp && *cp == esc; ne++)
58*d67f4a11SSascha Wildner 		continue;
59*d67f4a11SSascha Wildner 
60*d67f4a11SSascha Wildner 	/* Return true if odd number of escape characters */
61*d67f4a11SSascha Wildner 	return (ne & 1) != 0;
62*d67f4a11SSascha Wildner }
63*d67f4a11SSascha Wildner 
64*d67f4a11SSascha Wildner 
65*d67f4a11SSascha Wildner /* fparseln():
66*d67f4a11SSascha Wildner  *	Read a line from a file parsing continuations ending in \
67*d67f4a11SSascha Wildner  *	and eliminating trailing newlines, or comments starting with
68*d67f4a11SSascha Wildner  *	the comment char.
69*d67f4a11SSascha Wildner  */
70*d67f4a11SSascha Wildner char *
fparseln(FILE * fp,size_t * size,size_t * lineno,const char str[3],int flags)71*d67f4a11SSascha Wildner fparseln(FILE *fp, size_t *size, size_t *lineno, const char str[3], int flags)
72*d67f4a11SSascha Wildner {
73*d67f4a11SSascha Wildner 	static const char dstr[3] = { '\\', '\\', '#' };
74*d67f4a11SSascha Wildner 
75*d67f4a11SSascha Wildner 	size_t	s, len;
76*d67f4a11SSascha Wildner 	char   *buf;
77*d67f4a11SSascha Wildner 	char   *ptr, *cp;
78*d67f4a11SSascha Wildner 	int	cnt;
79*d67f4a11SSascha Wildner 	char	esc, con, nl, com;
80*d67f4a11SSascha Wildner 
81*d67f4a11SSascha Wildner #if 0
82*d67f4a11SSascha Wildner 	_DIAGASSERT(fp != NULL);
83*d67f4a11SSascha Wildner #endif
84*d67f4a11SSascha Wildner 
85*d67f4a11SSascha Wildner 	len = 0;
86*d67f4a11SSascha Wildner 	buf = NULL;
87*d67f4a11SSascha Wildner 	cnt = 1;
88*d67f4a11SSascha Wildner 
89*d67f4a11SSascha Wildner 	if (str == NULL)
90*d67f4a11SSascha Wildner 		str = dstr;
91*d67f4a11SSascha Wildner 
92*d67f4a11SSascha Wildner 	esc = str[0];
93*d67f4a11SSascha Wildner 	con = str[1];
94*d67f4a11SSascha Wildner 	com = str[2];
95*d67f4a11SSascha Wildner 	/*
96*d67f4a11SSascha Wildner 	 * XXX: it would be cool to be able to specify the newline character,
97*d67f4a11SSascha Wildner 	 * but unfortunately, fgetln does not let us
98*d67f4a11SSascha Wildner 	 */
99*d67f4a11SSascha Wildner 	nl  = '\n';
100*d67f4a11SSascha Wildner 
101*d67f4a11SSascha Wildner 	while (cnt) {
102*d67f4a11SSascha Wildner 		cnt = 0;
103*d67f4a11SSascha Wildner 
104*d67f4a11SSascha Wildner 		if (lineno)
105*d67f4a11SSascha Wildner 			(*lineno)++;
106*d67f4a11SSascha Wildner 
107*d67f4a11SSascha Wildner 		if ((ptr = fgetln(fp, &s)) == NULL)
108*d67f4a11SSascha Wildner 			break;
109*d67f4a11SSascha Wildner 
110*d67f4a11SSascha Wildner 		if (s && com) {		/* Check and eliminate comments */
111*d67f4a11SSascha Wildner 			for (cp = ptr; cp < ptr + s; cp++)
112*d67f4a11SSascha Wildner 				if (*cp == com && !isescaped(ptr, cp, esc)) {
113*d67f4a11SSascha Wildner 					s = cp - ptr;
114*d67f4a11SSascha Wildner 					cnt = s == 0 && buf == NULL;
115*d67f4a11SSascha Wildner 					break;
116*d67f4a11SSascha Wildner 				}
117*d67f4a11SSascha Wildner 		}
118*d67f4a11SSascha Wildner 
119*d67f4a11SSascha Wildner 		if (s && nl) { 		/* Check and eliminate newlines */
120*d67f4a11SSascha Wildner 			cp = &ptr[s - 1];
121*d67f4a11SSascha Wildner 
122*d67f4a11SSascha Wildner 			if (*cp == nl)
123*d67f4a11SSascha Wildner 				s--;	/* forget newline */
124*d67f4a11SSascha Wildner 		}
125*d67f4a11SSascha Wildner 
126*d67f4a11SSascha Wildner 		if (s && con) {		/* Check and eliminate continuations */
127*d67f4a11SSascha Wildner 			cp = &ptr[s - 1];
128*d67f4a11SSascha Wildner 
129*d67f4a11SSascha Wildner 			if (*cp == con && !isescaped(ptr, cp, esc)) {
130*d67f4a11SSascha Wildner 				s--;	/* forget continuation char */
131*d67f4a11SSascha Wildner 				cnt = 1;
132*d67f4a11SSascha Wildner 			}
133*d67f4a11SSascha Wildner 		}
134*d67f4a11SSascha Wildner 
135*d67f4a11SSascha Wildner 		if (s == 0) {
136*d67f4a11SSascha Wildner 			/*
137*d67f4a11SSascha Wildner 			 * nothing to add, skip realloc except in case
138*d67f4a11SSascha Wildner 			 * we need a minimal buf to return an empty line
139*d67f4a11SSascha Wildner 			 */
140*d67f4a11SSascha Wildner 			if (cnt || buf != NULL)
141*d67f4a11SSascha Wildner 				continue;
142*d67f4a11SSascha Wildner 		}
143*d67f4a11SSascha Wildner 
144*d67f4a11SSascha Wildner 		if ((cp = realloc(buf, len + s + 1)) == NULL) {
145*d67f4a11SSascha Wildner 			free(buf);
146*d67f4a11SSascha Wildner 			return NULL;
147*d67f4a11SSascha Wildner 		}
148*d67f4a11SSascha Wildner 		buf = cp;
149*d67f4a11SSascha Wildner 
150*d67f4a11SSascha Wildner 		(void) memcpy(buf + len, ptr, s);
151*d67f4a11SSascha Wildner 		len += s;
152*d67f4a11SSascha Wildner 		buf[len] = '\0';
153*d67f4a11SSascha Wildner 	}
154*d67f4a11SSascha Wildner 
155*d67f4a11SSascha Wildner 	if ((flags & FPARSELN_UNESCALL) != 0 && esc && buf != NULL &&
156*d67f4a11SSascha Wildner 	    strchr(buf, esc) != NULL) {
157*d67f4a11SSascha Wildner 		ptr = cp = buf;
158*d67f4a11SSascha Wildner 		while (cp[0] != '\0') {
159*d67f4a11SSascha Wildner 			int skipesc;
160*d67f4a11SSascha Wildner 
161*d67f4a11SSascha Wildner 			while (cp[0] != '\0' && cp[0] != esc)
162*d67f4a11SSascha Wildner 				*ptr++ = *cp++;
163*d67f4a11SSascha Wildner 			if (cp[0] == '\0' || cp[1] == '\0')
164*d67f4a11SSascha Wildner 				break;
165*d67f4a11SSascha Wildner 
166*d67f4a11SSascha Wildner 			skipesc = 0;
167*d67f4a11SSascha Wildner 			if (cp[1] == com)
168*d67f4a11SSascha Wildner 				skipesc += (flags & FPARSELN_UNESCCOMM);
169*d67f4a11SSascha Wildner 			if (cp[1] == con)
170*d67f4a11SSascha Wildner 				skipesc += (flags & FPARSELN_UNESCCONT);
171*d67f4a11SSascha Wildner 			if (cp[1] == esc)
172*d67f4a11SSascha Wildner 				skipesc += (flags & FPARSELN_UNESCESC);
173*d67f4a11SSascha Wildner 			if (cp[1] != com && cp[1] != con && cp[1] != esc)
174*d67f4a11SSascha Wildner 				skipesc = (flags & FPARSELN_UNESCREST);
175*d67f4a11SSascha Wildner 
176*d67f4a11SSascha Wildner 			if (skipesc)
177*d67f4a11SSascha Wildner 				cp++;
178*d67f4a11SSascha Wildner 			else
179*d67f4a11SSascha Wildner 				*ptr++ = *cp++;
180*d67f4a11SSascha Wildner 			*ptr++ = *cp++;
181*d67f4a11SSascha Wildner 		}
182*d67f4a11SSascha Wildner 		*ptr = '\0';
183*d67f4a11SSascha Wildner 		len = strlen(buf);
184*d67f4a11SSascha Wildner 	}
185*d67f4a11SSascha Wildner 
186*d67f4a11SSascha Wildner 	if (size)
187*d67f4a11SSascha Wildner 		*size = len;
188*d67f4a11SSascha Wildner 	return buf;
189*d67f4a11SSascha Wildner }
190*d67f4a11SSascha Wildner 
191*d67f4a11SSascha Wildner #ifdef TEST
192*d67f4a11SSascha Wildner 
193*d67f4a11SSascha Wildner int
main(int argc,char * argv[])194*d67f4a11SSascha Wildner main(int argc, char *argv[])
195*d67f4a11SSascha Wildner {
196*d67f4a11SSascha Wildner 	char   *ptr;
197*d67f4a11SSascha Wildner 	size_t	size, line;
198*d67f4a11SSascha Wildner 
199*d67f4a11SSascha Wildner 	line = 0;
200*d67f4a11SSascha Wildner 	while ((ptr = fparseln(stdin, &size, &line, NULL,
201*d67f4a11SSascha Wildner 	    FPARSELN_UNESCALL)) != NULL)
202*d67f4a11SSascha Wildner 		printf("line %d (%d) |%s|\n", line, size, ptr);
203*d67f4a11SSascha Wildner 	return 0;
204*d67f4a11SSascha Wildner }
205*d67f4a11SSascha Wildner 
206*d67f4a11SSascha Wildner /*
207*d67f4a11SSascha Wildner 
208*d67f4a11SSascha Wildner # This is a test
209*d67f4a11SSascha Wildner line 1
210*d67f4a11SSascha Wildner line 2 \
211*d67f4a11SSascha Wildner line 3 # Comment
212*d67f4a11SSascha Wildner line 4 \# Not comment \\\\
213*d67f4a11SSascha Wildner 
214*d67f4a11SSascha Wildner # And a comment \
215*d67f4a11SSascha Wildner line 5 \\\
216*d67f4a11SSascha Wildner line 6
217*d67f4a11SSascha Wildner 
218*d67f4a11SSascha Wildner */
219*d67f4a11SSascha Wildner 
220*d67f4a11SSascha Wildner #endif /* TEST */
221