interp_parse.c revision 329183 
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 <sys/cdefs.h>
18__FBSDID("$FreeBSD: stable/11/stand/common/interp_parse.c 329183 2018-02-12 20:51:28Z kevans $");
19
20#include <stand.h>
21#include <string.h>
22#include "bootstrap.h"
23
24static void	 clean(void);
25static int	 insert(int *argcp, char *buf);
26static char	*variable_lookup(char *name);
27
28#define PARSE_BUFSIZE	1024	/* maximum size of one element */
29#define MAXARGS		20	/* maximum number of elements */
30static char		*args[MAXARGS];
31
32/*
33 * parse: accept a string of input and "parse" it for backslash
34 * substitutions and environment variable expansions (${var}),
35 * returning an argc/argv style vector of whitespace separated
36 * arguments.  Returns 0 on success, 1 on failure (ok, ok, so I
37 * wimped-out on the error codes! :).
38 *
39 * Note that the argv array returned must be freed by the caller, but
40 * we own the space allocated for arguments and will free that on next
41 * invocation.  This allows argv consumers to modify the array if
42 * required.
43 *
44 * NB: environment variables that expand to more than one whitespace
45 * separated token will be returned as a single argv[] element, not
46 * split in turn.  Expanded text is also immune to further backslash
47 * elimination or expansion since this is a one-pass, non-recursive
48 * parser.  You didn't specify more than this so if you want more, ask
49 * me. - jkh
50 */
51
52#define PARSE_FAIL(expr) \
53	if (expr) { \
54		printf("fail at line %d\n", __LINE__); \
55		clean(); \
56		free(copy); \
57		free(buf); \
58		return 1; \
59	}
60
61/* Accept the usual delimiters for a variable, returning counterpart */
62static char
63isdelim(int ch)
64{
65
66	if (ch == '{')
67		return '}';
68	else if (ch == '(')
69		return ')';
70	return '\0';
71}
72
73static int
74isquote(int ch)
75{
76
77	return (ch == '\'');
78}
79
80static int
81isdquote(int ch)
82{
83
84	return (ch == '"');
85}
86
87int
88parse(int *argc, char ***argv, const char *str)
89{
90	int ac;
91	char *val, *p, *q, *copy = NULL;
92	size_t i = 0;
93	char token, tmp, quote, dquote, *buf;
94	enum { STR, VAR, WHITE } state;
95
96	ac = *argc = 0;
97	dquote = quote = 0;
98	if (!str || (p = copy = backslash(str)) == NULL)
99		return 1;
100
101	/* Initialize vector and state */
102	clean();
103	state = STR;
104	buf = (char *)malloc(PARSE_BUFSIZE);
105	token = 0;
106
107	/* And awaaaaaaaaay we go! */
108	while (*p) {
109		switch (state) {
110		case STR:
111			if ((*p == '\\') && p[1]) {
112				p++;
113				PARSE_FAIL(i == (PARSE_BUFSIZE - 1));
114				buf[i++] = *p++;
115			} else if (isquote(*p)) {
116				quote = quote ? 0 : *p;
117				if (dquote) { /* keep quote */
118					PARSE_FAIL(i == (PARSE_BUFSIZE - 1));
119					buf[i++] = *p++;
120				} else
121					++p;
122			} else if (isdquote(*p)) {
123				dquote = dquote ? 0 : *p;
124				if (quote) { /* keep dquote */
125					PARSE_FAIL(i == (PARSE_BUFSIZE - 1));
126					buf[i++] = *p++;
127				} else
128					++p;
129			} else if (isspace(*p) && !quote && !dquote) {
130				state = WHITE;
131				if (i) {
132					buf[i] = '\0';
133					PARSE_FAIL(insert(&ac, buf));
134					i = 0;
135				}
136				++p;
137			} else if (*p == '$' && !quote) {
138				token = isdelim(*(p + 1));
139				if (token)
140					p += 2;
141				else
142					++p;
143				state = VAR;
144			} else {
145				PARSE_FAIL(i == (PARSE_BUFSIZE - 1));
146				buf[i++] = *p++;
147			}
148			break;
149
150		case WHITE:
151			if (isspace(*p))
152				++p;
153			else
154				state = STR;
155			break;
156
157		case VAR:
158			if (token) {
159				PARSE_FAIL((q = strchr(p, token)) == NULL);
160			} else {
161				q = p;
162				while (*q && !isspace(*q))
163					++q;
164			}
165			tmp = *q;
166			*q = '\0';
167			if ((val = variable_lookup(p)) != NULL) {
168				size_t len = strlen(val);
169
170				strncpy(buf + i, val, PARSE_BUFSIZE - (i + 1));
171				i += min(len, PARSE_BUFSIZE - 1);
172			}
173			*q = tmp;	/* restore value */
174			p = q + (token ? 1 : 0);
175			state = STR;
176			break;
177		}
178	}
179	/* missing terminating ' or " */
180	PARSE_FAIL(quote || dquote);
181	/* If at end of token, add it */
182	if (i && state == STR) {
183		buf[i] = '\0';
184		PARSE_FAIL(insert(&ac, buf));
185	}
186	args[ac] = NULL;
187	*argc = ac;
188	*argv = (char **)malloc((sizeof(char *) * ac + 1));
189	bcopy(args, *argv, sizeof(char *) * ac + 1);
190	free(buf);
191	free(copy);
192	return 0;
193}
194
195#define MAXARGS	20
196
197/* Clean vector space */
198static void
199clean(void)
200{
201	int		i;
202
203	for (i = 0; i < MAXARGS; i++) {
204		if (args[i] != NULL) {
205			free(args[i]);
206			args[i] = NULL;
207		}
208	}
209}
210
211static int
212insert(int *argcp, char *buf)
213{
214
215	if (*argcp >= MAXARGS)
216		return 1;
217	args[(*argcp)++] = strdup(buf);
218	return 0;
219}
220
221static char *
222variable_lookup(char *name)
223{
224
225	/* XXX search "special variable" space first? */
226	return (char *)getenv(name);
227}
228