Home | History | Annotate | Line # | Download | only in stdio
      1 /*	$NetBSD: fparseln.c,v 1.10 2009/10/21 01:07:45 snj Exp $	*/
      2 
      3 /*
      4  * Copyright (c) 1997 Christos Zoulas.  All rights reserved.
      5  *
      6  * Redistribution and use in source and binary forms, with or without
      7  * modification, are permitted provided that the following conditions
      8  * are met:
      9  * 1. Redistributions of source code must retain the above copyright
     10  *    notice, this list of conditions and the following disclaimer.
     11  * 2. Redistributions in binary form must reproduce the above copyright
     12  *    notice, this list of conditions and the following disclaimer in the
     13  *    documentation and/or other materials provided with the distribution.
     14  *
     15  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
     16  * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
     17  * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
     18  * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
     19  * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
     20  * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
     21  * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
     22  * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
     23  * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
     24  * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
     25  */
     26 
     27 #include <sys/cdefs.h>
     28 #if defined(LIBC_SCCS) && !defined(lint)
     29 __RCSID("$NetBSD: fparseln.c,v 1.10 2009/10/21 01:07:45 snj Exp $");
     30 #endif /* LIBC_SCCS and not lint */
     31 
     32 #include "namespace.h"
     33 
     34 #include <assert.h>
     35 #include <errno.h>
     36 #include <stdio.h>
     37 #include <string.h>
     38 #include <stdlib.h>
     39 
     40 #ifdef __weak_alias
     41 __weak_alias(fparseln,_fparseln)
     42 #endif
     43 
     44 #if ! HAVE_FPARSELN || BROKEN_FPARSELN
     45 
     46 #ifndef HAVE_NBTOOL_CONFIG_H
     47 #include "reentrant.h"
     48 #include "local.h"
     49 #else
     50 #define FLOCKFILE(fp)
     51 #define FUNLOCKFILE(fp)
     52 #endif
     53 
     54 #if defined(_REENTRANT) && !HAVE_NBTOOL_CONFIG_H
     55 #define __fgetln(f, l) __fgetstr(f, l, '\n')
     56 #else
     57 #define __fgetln(f, l) fgetln(f, l)
     58 #endif
     59 
     60 static int isescaped(const char *, const char *, int);
     61 
     62 /* isescaped():
     63  *	Return true if the character in *p that belongs to a string
     64  *	that starts in *sp, is escaped by the escape character esc.
     65  */
     66 static int
     67 isescaped(const char *sp, const char *p, int esc)
     68 {
     69 	const char     *cp;
     70 	size_t		ne;
     71 
     72 	_DIAGASSERT(sp != NULL);
     73 	_DIAGASSERT(p != NULL);
     74 
     75 	/* No escape character */
     76 	if (esc == '\0')
     77 		return 0;
     78 
     79 	/* Count the number of escape characters that precede ours */
     80 	for (ne = 0, cp = p; --cp >= sp && *cp == esc; ne++)
     81 		continue;
     82 
     83 	/* Return true if odd number of escape characters */
     84 	return (ne & 1) != 0;
     85 }
     86 
     87 
     88 /* fparseln():
     89  *	Read a line from a file parsing continuations ending in \
     90  *	and eliminating trailing newlines, or comments starting with
     91  *	the comment char.
     92  */
     93 char *
     94 fparseln(FILE *fp, size_t *size, size_t *lineno, const char str[3], int flags)
     95 {
     96 	static const char dstr[3] = { '\\', '\\', '#' };
     97 
     98 	size_t	s, len;
     99 	char   *buf;
    100 	char   *ptr, *cp;
    101 	int	cnt;
    102 	char	esc, con, nl, com;
    103 
    104 	_DIAGASSERT(fp != NULL);
    105 
    106 	len = 0;
    107 	buf = NULL;
    108 	cnt = 1;
    109 
    110 	if (str == NULL)
    111 		str = dstr;
    112 
    113 	esc = str[0];
    114 	con = str[1];
    115 	com = str[2];
    116 	/*
    117 	 * XXX: it would be cool to be able to specify the newline character,
    118 	 * but unfortunately, fgetln does not let us
    119 	 */
    120 	nl  = '\n';
    121 
    122 	FLOCKFILE(fp);
    123 
    124 	while (cnt) {
    125 		cnt = 0;
    126 
    127 		if (lineno)
    128 			(*lineno)++;
    129 
    130 		if ((ptr = __fgetln(fp, &s)) == NULL)
    131 			break;
    132 
    133 		if (s && com) {		/* Check and eliminate comments */
    134 			for (cp = ptr; cp < ptr + s; cp++)
    135 				if (*cp == com && !isescaped(ptr, cp, esc)) {
    136 					s = cp - ptr;
    137 					cnt = s == 0 && buf == NULL;
    138 					break;
    139 				}
    140 		}
    141 
    142 		if (s && nl) { 		/* Check and eliminate newlines */
    143 			cp = &ptr[s - 1];
    144 
    145 			if (*cp == nl)
    146 				s--;	/* forget newline */
    147 		}
    148 
    149 		if (s && con) {		/* Check and eliminate continuations */
    150 			cp = &ptr[s - 1];
    151 
    152 			if (*cp == con && !isescaped(ptr, cp, esc)) {
    153 				s--;	/* forget continuation char */
    154 				cnt = 1;
    155 			}
    156 		}
    157 
    158 		if (s == 0) {
    159 			/*
    160 			 * nothing to add, skip realloc except in case
    161 			 * we need a minimal buf to return an empty line
    162 			 */
    163 			if (cnt || buf != NULL)
    164 				continue;
    165 		}
    166 
    167 		if ((cp = realloc(buf, len + s + 1)) == NULL) {
    168 			FUNLOCKFILE(fp);
    169 			free(buf);
    170 			return NULL;
    171 		}
    172 		buf = cp;
    173 
    174 		(void) memcpy(buf + len, ptr, s);
    175 		len += s;
    176 		buf[len] = '\0';
    177 	}
    178 
    179 	FUNLOCKFILE(fp);
    180 
    181 	if ((flags & FPARSELN_UNESCALL) != 0 && esc && buf != NULL &&
    182 	    strchr(buf, esc) != NULL) {
    183 		ptr = cp = buf;
    184 		while (cp[0] != '\0') {
    185 			int skipesc;
    186 
    187 			while (cp[0] != '\0' && cp[0] != esc)
    188 				*ptr++ = *cp++;
    189 			if (cp[0] == '\0' || cp[1] == '\0')
    190 				break;
    191 
    192 			skipesc = 0;
    193 			if (cp[1] == com)
    194 				skipesc += (flags & FPARSELN_UNESCCOMM);
    195 			if (cp[1] == con)
    196 				skipesc += (flags & FPARSELN_UNESCCONT);
    197 			if (cp[1] == esc)
    198 				skipesc += (flags & FPARSELN_UNESCESC);
    199 			if (cp[1] != com && cp[1] != con && cp[1] != esc)
    200 				skipesc = (flags & FPARSELN_UNESCREST);
    201 
    202 			if (skipesc)
    203 				cp++;
    204 			else
    205 				*ptr++ = *cp++;
    206 			*ptr++ = *cp++;
    207 		}
    208 		*ptr = '\0';
    209 		len = strlen(buf);
    210 	}
    211 
    212 	if (size)
    213 		*size = len;
    214 	return buf;
    215 }
    216 
    217 #ifdef TEST
    218 
    219 int main(int, char **);
    220 
    221 int
    222 main(int argc, char **argv)
    223 {
    224 	char   *ptr;
    225 	size_t	size, line;
    226 
    227 	line = 0;
    228 	while ((ptr = fparseln(stdin, &size, &line, NULL,
    229 	    FPARSELN_UNESCALL)) != NULL)
    230 		printf("line %d (%d) |%s|\n", line, size, ptr);
    231 	return 0;
    232 }
    233 
    234 /*
    235 
    236 # This is a test
    237 line 1
    238 line 2 \
    239 line 3 # Comment
    240 line 4 \# Not comment \\\\
    241 
    242 # And a comment \
    243 line 5 \\\
    244 line 6
    245 
    246 */
    247 
    248 #endif /* TEST */
    249 #endif	/* ! HAVE_FPARSELN || BROKEN_FPARSELN */
    250