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("$FreeBSD$"); 19 20 #include <stand.h> 21 #include <string.h> 22 #include "bootstrap.h" 23 24 static void clean(void); 25 static int insert(int *argcp, char *buf); 26 static char *variable_lookup(char *name); 27 28 #define PARSE_BUFSIZE 1024 /* maximum size of one element */ 29 #define MAXARGS 20 /* maximum number of elements */ 30 static 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 */ 62 static char 63 isdelim(int ch) 64 { 65 if (ch == '{') 66 return '}'; 67 else if (ch == '(') 68 return ')'; 69 return '\0'; 70 } 71 72 static int 73 isquote(int ch) 74 { 75 return (ch == '\''); 76 } 77 78 static int 79 isdquote(int ch) 80 { 81 return (ch == '"'); 82 } 83 84 int 85 parse(int *argc, char ***argv, 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 } 120 } else if (isdquote(*p)) { 121 dquote = dquote ? 0 : *p; 122 if (quote) { /* keep dquote */ 123 PARSE_FAIL(i == (PARSE_BUFSIZE - 1)); 124 buf[i++] = *p++; 125 } else { 126 ++p; 127 } 128 } else if (isspace(*p) && !quote && !dquote) { 129 state = WHITE; 130 if (i) { 131 buf[i] = '\0'; 132 PARSE_FAIL(insert(&ac, buf)); 133 i = 0; 134 } 135 ++p; 136 } else if (*p == '$' && !quote) { 137 token = isdelim(*(p + 1)); 138 if (token) 139 p += 2; 140 else 141 ++p; 142 state = VAR; 143 } else { 144 PARSE_FAIL(i == (PARSE_BUFSIZE - 1)); 145 buf[i++] = *p++; 146 } 147 break; 148 149 case WHITE: 150 if (isspace(*p)) 151 ++p; 152 else 153 state = STR; 154 break; 155 156 case VAR: 157 if (token) { 158 PARSE_FAIL((q = strchr(p, token)) == NULL); 159 } else { 160 q = p; 161 while (*q && !isspace(*q)) 162 ++q; 163 } 164 tmp = *q; 165 *q = '\0'; 166 if ((val = variable_lookup(p)) != NULL) { 167 size_t len = strlen(val); 168 169 strncpy(buf + i, val, PARSE_BUFSIZE - (i + 1)); 170 i += min(len, PARSE_BUFSIZE - 1); 171 } 172 *q = tmp; /* restore value */ 173 p = q + (token ? 1 : 0); 174 state = STR; 175 break; 176 } 177 } 178 /* missing terminating ' or " */ 179 PARSE_FAIL(quote || dquote); 180 /* If at end of token, add it */ 181 if (i && state == STR) { 182 buf[i] = '\0'; 183 PARSE_FAIL(insert(&ac, buf)); 184 } 185 args[ac] = NULL; 186 *argc = ac; 187 *argv = (char **)malloc((sizeof(char *) * ac + 1)); 188 bcopy(args, *argv, sizeof(char *) * ac + 1); 189 free(buf); 190 free(copy); 191 return 0; 192 } 193 194 #define MAXARGS 20 195 196 /* Clean vector space */ 197 static void 198 clean(void) 199 { 200 int i; 201 202 for (i = 0; i < MAXARGS; i++) { 203 if (args[i] != NULL) { 204 free(args[i]); 205 args[i] = NULL; 206 } 207 } 208 } 209 210 static int 211 insert(int *argcp, char *buf) 212 { 213 if (*argcp >= MAXARGS) 214 return 1; 215 args[(*argcp)++] = strdup(buf); 216 return 0; 217 } 218 219 static char * 220 variable_lookup(char *name) 221 { 222 /* XXX search "special variable" space first? */ 223 return (char *)getenv(name); 224 } 225