interp_parse.c revision 38765 
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 *	$Id: interp_parse.c,v 1.1 1998/09/01 00:41:24 msmith Exp $
15 *
16 * The meat of the simple parser.
17 */
18
19#include <stand.h>
20#include <string.h>
21
22/* Forward decls */
23extern char *backslash(char *str);
24
25static void init(int *argcp, char ***argvp);
26static void clean(void);
27static int insert(int *argcp, char *buf);
28static char *variable_lookup(char *name);
29
30#define PARSE_BUFSIZE	1024	/* maximum size of one element */
31#define MAXARGS		20	/* maximum number of elements */
32static char		*args[MAXARGS];
33
34/*
35 * parse: accept a string of input and "parse" it for backslash
36 * substitutions and environment variable expansions (${var}),
37 * returning an argc/argv style vector of whitespace separated
38 * arguments.  Returns 0 on success, 1 on failure (ok, ok, so I
39 * wimped-out on the error codes! :).
40 *
41 * Note that the argv array returned must be freed by the caller, but
42 * we own the space allocated for arguments and will free that on next
43 * invocation.  This allows argv consumers to modify the array if
44 * required.
45 *
46 * NB: environment variables that expand to more than one whitespace
47 * separated token will be returned as a single argv[] element, not
48 * split in turn.  Expanded text is also immune to further backslash
49 * elimination or expansion since this is a one-pass, non-recursive
50 * parser.  You didn't specify more than this so if you want more, ask
51 * me. - jkh
52 */
53
54#define PARSE_FAIL(expr) \
55if (expr) { \
56    printf("fail at line %d\n", __LINE__); \
57    clean(); \
58    free(copy); \
59    free(buf); \
60    return 1; \
61}
62
63/* Accept the usual delimiters for a variable, returning counterpart */
64static char
65isdelim(char ch)
66{
67    if (ch == '{')
68	return '}';
69    else if (ch == '(')
70	return ')';
71    return '\0';
72}
73
74static int
75isquote(char ch)
76{
77    return (ch == '\'' || ch == '"');
78}
79
80int
81parse(int *argc, char ***argv, char *str)
82{
83    int ac;
84    char *val, *p, *q, *copy = NULL;
85    int i = 0;
86    char token, tmp, quote, *buf;
87    enum { STR, VAR, WHITE } state;
88
89    ac = *argc = 0;
90    quote = 0;
91    if (!str || (p = copy = backslash(str)) == NULL)
92	return 1;
93
94    /* Initialize vector and state */
95    clean();
96    state = STR;
97    buf = (char *)malloc(PARSE_BUFSIZE);
98    token = 0;
99
100    /* And awaaaaaaaaay we go! */
101    while (*p) {
102	switch (state) {
103	case STR:
104	    if (isquote(*p)) {
105		quote = quote ? 0 : *p;
106		++p;
107	    }
108	    else if (isspace(*p) && !quote) {
109		state = WHITE;
110		if (i) {
111		    buf[i] = '\0';
112		    PARSE_FAIL(insert(&ac, buf));
113		    i = 0;
114		}
115		++p;
116	    } else if (*p == '$') {
117		token = isdelim(*(p + 1));
118		if (token)
119		    p += 2;
120		else
121		    ++p;
122		state = VAR;
123	    } else {
124		PARSE_FAIL(i == (PARSE_BUFSIZE - 1));
125		buf[i++] = *p++;
126	    }
127	    break;
128
129	case WHITE:
130	    if (isspace(*p))
131		++p;
132	    else
133		state = STR;
134	    break;
135
136	case VAR:
137	    if (token) {
138		PARSE_FAIL((q = index(p, token)) == NULL);
139	    } else {
140		q = p;
141		while (*q && !isspace(*q))
142		    ++q;
143	    }
144	    tmp = *q;
145	    *q = '\0';
146	    if ((val = variable_lookup(p)) != NULL) {
147		int len = strlen(val);
148
149		strncpy(buf + i, val, PARSE_BUFSIZE - (i + 1));
150		i += min(len, sizeof(buf) - 1);
151	    }
152	    *q = tmp;	/* restore value */
153	    p = q + (token ? 1 : 0);
154	    state = STR;
155	    break;
156	}
157    }
158    /* If at end of token, add it */
159    if (i && state == STR) {
160	buf[i] = '\0';
161	PARSE_FAIL(insert(&ac, buf));
162    }
163    args[ac] = NULL;
164    *argc = ac;
165    *argv = (char **)malloc((sizeof(char *) * ac + 1));
166    bcopy(args, *argv, sizeof(char *) * ac + 1);
167    free(copy);
168    return 0;
169}
170
171#define MAXARGS	20
172
173/* Clean vector space */
174static void
175clean(void)
176{
177    int		i;
178
179    for (i = 0; i < MAXARGS; i++) {
180	if (args[i] != NULL) {
181	    free(args[i]);
182	    args[i] = NULL;
183	}
184    }
185}
186
187static int
188insert(int *argcp, char *buf)
189{
190    if (*argcp >= MAXARGS)
191	return 1;
192    args[(*argcp)++] = strdup(buf);
193    return 0;
194}
195
196static char *
197variable_lookup(char *name)
198{
199    /* XXX search "special variable" space first? */
200    return (char *)getenv(name);
201}
202