1 /*
2  * argv_parse.c --- utility function for parsing a string into a
3  * 	argc, argv array.
4  *
5  * This file defines a function argv_parse() which parsing a
6  * passed-in string, handling double quotes and backslashes, and
7  * creates an allocated argv vector which can be freed using the
8  * argv_free() function.
9  *
10  * See argv_parse.h for the formal definition of the functions.
11  *
12  * Copyright 1999 by Theodore Ts'o.
13  *
14  * Permission to use, copy, modify, and distribute this software for
15  * any purpose with or without fee is hereby granted, provided that
16  * the above copyright notice and this permission notice appear in all
17  * copies.  THE SOFTWARE IS PROVIDED "AS IS" AND THEODORE TS'O (THE
18  * AUTHOR) DISCLAIMS ALL WARRANTIES WITH REGARD TO THIS SOFTWARE,
19  * INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS.
20  * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY SPECIAL, DIRECT,
21  * INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER
22  * RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN ACTION
23  * OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF OR
24  * IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.  (Isn't
25  * it sick that the U.S. culture of lawsuit-happy lawyers requires
26  * this kind of disclaimer?)
27  *
28  * Version 1.1, modified 2/27/1999
29  */
30 
31 #ifdef HAVE_STDLIB_H
32 #include <stdlib.h>
33 #endif
34 #include <ctype.h>
35 #include <string.h>
36 #include "argv_parse.h"
37 
38 #define STATE_WHITESPACE	1
39 #define STATE_TOKEN		2
40 #define STATE_QUOTED		3
41 
42 /*
43  * Returns 0 on success, -1 on failure.
44  */
argv_parse(char * in_buf,int * ret_argc,char *** ret_argv)45 int argv_parse(char *in_buf, int *ret_argc, char ***ret_argv)
46 {
47 	int	argc = 0, max_argc = 0;
48 	char 	**argv, **new_argv, *buf, ch;
49 	char	*cp = 0, *outcp = 0;
50 	int	state = STATE_WHITESPACE;
51 
52 	buf = malloc(strlen(in_buf)+1);
53 	if (!buf)
54 		return -1;
55 
56 	max_argc = 0; argc = 0; argv = 0;
57 	outcp = buf;
58 	for (cp = in_buf; (ch = *cp); cp++) {
59 		if (state == STATE_WHITESPACE) {
60 			if (isspace((int) ch))
61 				continue;
62 			/* Not whitespace, so start a new token */
63 			state = STATE_TOKEN;
64 			if (argc >= max_argc) {
65 				max_argc += 3;
66 				new_argv = realloc(argv,
67 						  (max_argc+1)*sizeof(char *));
68 				if (!new_argv) {
69 					free(argv);
70 					free(buf);
71 					return -1;
72 				}
73 				argv = new_argv;
74 			}
75 			argv[argc++] = outcp;
76 		}
77 		if (state == STATE_QUOTED) {
78 			if (ch == '"')
79 				state = STATE_TOKEN;
80 			else
81 				*outcp++ = ch;
82 			continue;
83 		}
84 		/* Must be processing characters in a word */
85 		if (isspace((int) ch)) {
86 			/*
87 			 * Terminate the current word and start
88 			 * looking for the beginning of the next word.
89 			 */
90 			*outcp++ = 0;
91 			state = STATE_WHITESPACE;
92 			continue;
93 		}
94 		if (ch == '"') {
95 			state = STATE_QUOTED;
96 			continue;
97 		}
98 		if (ch == '\\') {
99 			ch = *++cp;
100 			switch (ch) {
101 			case '\0':
102 				ch = '\\'; cp--; break;
103 			case 'n':
104 				ch = '\n'; break;
105 			case 't':
106 				ch = '\t'; break;
107 			case 'b':
108 				ch = '\b'; break;
109 			}
110 		}
111 		*outcp++ = ch;
112 	}
113 	if (state != STATE_WHITESPACE)
114 		*outcp++ = '\0';
115 	if (argv == 0) {
116 		argv = malloc(sizeof(char *));
117 		free(buf);
118 	}
119 	argv[argc] = 0;
120 	if (ret_argc)
121 		*ret_argc = argc;
122 	if (ret_argv)
123 		*ret_argv = argv;
124 	return 0;
125 }
126 
argv_free(char ** argv)127 void argv_free(char **argv)
128 {
129 	free(*argv);
130 	free(argv);
131 }
132 
133 #ifdef DEBUG
134 /*
135  * For debugging
136  */
137 
138 #include <stdio.h>
139 
main(int argc,char ** argv)140 int main(int argc, char **argv)
141 {
142 	int	ac, ret;
143 	char	**av, **cpp;
144 	char	buf[256];
145 
146 	while (!feof(stdin)) {
147 		if (fgets(buf, sizeof(buf), stdin) == NULL)
148 			break;
149 		ret = argv_parse(buf, &ac, &av);
150 		if (ret != 0) {
151 			printf("Argv_parse returned %d!\n", ret);
152 			continue;
153 		}
154 		printf("Argv_parse returned %d arguments...\n", ac);
155 		for (cpp = av; *cpp; cpp++) {
156 			if (cpp != av)
157 				printf(", ");
158 			printf("'%s'", *cpp);
159 		}
160 		printf("\n");
161 		argv_free(av);
162 	}
163 	exit(0);
164 }
165 #endif /* DEBUG */
166