1/*-
2 * Redistribution and use in source and binary forms, with or without
3 * modification, are permitted provided that the following conditions
4 * are met:
5 * 1. Redistributions of source code must retain the above copyright
6 *    notice, this list of conditions and the following disclaimer.
7 * 2. Redistributions in binary form must reproduce the above copyright
8 *    notice, this list of conditions and the following disclaimer in the
9 *    documentation and/or other materials provided with the distribution.
10 *
11 * Jordan K. Hubbard
12 * 29 August 1998
13 *
14 * The meat of the simple parser.
15 */
16
17#include <stand.h>
18#include <string.h>
19#include "bootstrap.h"
20
21static void	 clean(void);
22static int	 insert(int *argcp, char *buf);
23static char	*variable_lookup(char *name);
24
25#define PARSE_BUFSIZE	1024	/* maximum size of one element */
26#define MAXARGS		20	/* maximum number of elements */
27static char		*args[MAXARGS];
28
29/*
30 * parse: accept a string of input and "parse" it for backslash
31 * substitutions and environment variable expansions (${var}),
32 * returning an argc/argv style vector of whitespace separated
33 * arguments.  Returns 0 on success, 1 on failure (ok, ok, so I
34 * wimped-out on the error codes! :).
35 *
36 * Note that the argv array returned must be freed by the caller, but
37 * we own the space allocated for arguments and will free that on next
38 * invocation.  This allows argv consumers to modify the array if
39 * required.
40 *
41 * NB: environment variables that expand to more than one whitespace
42 * separated token will be returned as a single argv[] element, not
43 * split in turn.  Expanded text is also immune to further backslash
44 * elimination or expansion since this is a one-pass, non-recursive
45 * parser.  You didn't specify more than this so if you want more, ask
46 * me. - jkh
47 */
48
49#define PARSE_FAIL(expr) \
50	if (expr) { \
51		printf("fail at line %d\n", __LINE__); \
52		clean(); \
53		free(copy); \
54		free(buf); \
55		return 1; \
56	}
57
58/* Accept the usual delimiters for a variable, returning counterpart */
59static char
60isdelim(int ch)
61{
62
63	if (ch == '{')
64		return '}';
65	else if (ch == '(')
66		return ')';
67	return '\0';
68}
69
70static int
71isquote(int ch)
72{
73
74	return (ch == '\'');
75}
76
77static int
78isdquote(int ch)
79{
80
81	return (ch == '"');
82}
83
84int
85parse(int *argc, char ***argv, const char *str)
86{
87	int ac;
88	char *val, *p, *q, *copy = NULL;
89	size_t i = 0;
90	char token, tmp, quote, dquote, *buf;
91	enum { STR, VAR, WHITE } state;
92
93	ac = *argc = 0;
94	dquote = quote = 0;
95	if (!str || (p = copy = backslash(str)) == NULL)
96		return 1;
97
98	/* Initialize vector and state */
99	clean();
100	state = STR;
101	buf = (char *)malloc(PARSE_BUFSIZE);
102	token = 0;
103
104	/* And awaaaaaaaaay we go! */
105	while (*p) {
106		switch (state) {
107		case STR:
108			if ((*p == '\\') && p[1]) {
109				p++;
110				PARSE_FAIL(i == (PARSE_BUFSIZE - 1));
111				buf[i++] = *p++;
112			} else if (isquote(*p)) {
113				quote = quote ? 0 : *p;
114				if (dquote) { /* keep quote */
115					PARSE_FAIL(i == (PARSE_BUFSIZE - 1));
116					buf[i++] = *p++;
117				} else
118					++p;
119			} else if (isdquote(*p)) {
120				dquote = dquote ? 0 : *p;
121				if (quote) { /* keep dquote */
122					PARSE_FAIL(i == (PARSE_BUFSIZE - 1));
123					buf[i++] = *p++;
124				} else
125					++p;
126			} else if (isspace(*p) && !quote && !dquote) {
127				state = WHITE;
128				if (i) {
129					buf[i] = '\0';
130					PARSE_FAIL(insert(&ac, buf));
131					i = 0;
132				}
133				++p;
134			} else if (*p == '$' && !quote) {
135				token = isdelim(*(p + 1));
136				if (token)
137					p += 2;
138				else
139					++p;
140				state = VAR;
141			} else {
142				PARSE_FAIL(i == (PARSE_BUFSIZE - 1));
143				buf[i++] = *p++;
144			}
145			break;
146
147		case WHITE:
148			if (isspace(*p))
149				++p;
150			else
151				state = STR;
152			break;
153
154		case VAR:
155			if (token) {
156				PARSE_FAIL((q = strchr(p, token)) == NULL);
157			} else {
158				q = p;
159				while (*q && !isspace(*q))
160					++q;
161			}
162			tmp = *q;
163			*q = '\0';
164			if ((val = variable_lookup(p)) != NULL) {
165				size_t len = strlen(val);
166
167				strncpy(buf + i, val, PARSE_BUFSIZE - (i + 1));
168				i += min(len, PARSE_BUFSIZE - 1);
169			}
170			*q = tmp;	/* restore value */
171			p = q + (token ? 1 : 0);
172			state = STR;
173			break;
174		}
175	}
176	/* missing terminating ' or " */
177	PARSE_FAIL(quote || dquote);
178	/* If at end of token, add it */
179	if (i && state == STR) {
180		buf[i] = '\0';
181		PARSE_FAIL(insert(&ac, buf));
182	}
183	args[ac] = NULL;
184	*argc = ac;
185	*argv = (char **)malloc((sizeof(char *) * ac + 1));
186	bcopy(args, *argv, sizeof(char *) * ac + 1);
187	free(buf);
188	free(copy);
189	return 0;
190}
191
192#define MAXARGS	20
193
194/* Clean vector space */
195static void
196clean(void)
197{
198	int		i;
199
200	for (i = 0; i < MAXARGS; i++) {
201		if (args[i] != NULL) {
202			free(args[i]);
203			args[i] = NULL;
204		}
205	}
206}
207
208static int
209insert(int *argcp, char *buf)
210{
211
212	if (*argcp >= MAXARGS)
213		return 1;
214	args[(*argcp)++] = strdup(buf);
215	return 0;
216}
217
218static char *
219variable_lookup(char *name)
220{
221
222	/* XXX search "special variable" space first? */
223	return (char *)getenv(name);
224}
225