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