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 #include <stand.h> 19 #include <string.h> 20 #include "bootstrap.h" 21 22 static void clean(void); 23 static int insert(int *argcp, char *buf); 24 static char *variable_lookup(char *name); 25 26 #define PARSE_BUFSIZE 1024 /* maximum size of one element */ 27 #define MAXARGS 20 /* maximum number of elements */ 28 static char *args[MAXARGS]; 29 30 /* 31 * parse: accept a string of input and "parse" it for backslash 32 * substitutions and environment variable expansions (${var}), 33 * returning an argc/argv style vector of whitespace separated 34 * arguments. Returns 0 on success, 1 on failure (ok, ok, so I 35 * wimped-out on the error codes! :). 36 * 37 * Note that the argv array returned must be freed by the caller, but 38 * we own the space allocated for arguments and will free that on next 39 * invocation. This allows argv consumers to modify the array if 40 * required. 41 * 42 * NB: environment variables that expand to more than one whitespace 43 * separated token will be returned as a single argv[] element, not 44 * split in turn. Expanded text is also immune to further backslash 45 * elimination or expansion since this is a one-pass, non-recursive 46 * parser. You didn't specify more than this so if you want more, ask 47 * me. - jkh 48 */ 49 50 #define PARSE_FAIL(expr) \ 51 if (expr) { \ 52 printf("fail at line %d\n", __LINE__); \ 53 clean(); \ 54 free(copy); \ 55 free(buf); \ 56 return 1; \ 57 } 58 59 /* Accept the usual delimiters for a variable, returning counterpart */ 60 static char 61 isdelim(int ch) 62 { 63 64 if (ch == '{') 65 return '}'; 66 else if (ch == '(') 67 return ')'; 68 return '\0'; 69 } 70 71 static int 72 isquote(int ch) 73 { 74 75 return (ch == '\''); 76 } 77 78 static int 79 isdquote(int ch) 80 { 81 82 return (ch == '"'); 83 } 84 85 int 86 parse(int *argc, char ***argv, const char *str) 87 { 88 int ac; 89 char *val, *p, *q, *copy = NULL; 90 size_t i = 0; 91 char token, tmp, quote, dquote, *buf; 92 enum { STR, VAR, WHITE } state; 93 94 ac = *argc = 0; 95 dquote = quote = 0; 96 if (!str || (p = copy = backslash(str)) == NULL) 97 return 1; 98 99 /* Initialize vector and state */ 100 clean(); 101 state = STR; 102 buf = (char *)malloc(PARSE_BUFSIZE); 103 token = 0; 104 105 /* And awaaaaaaaaay we go! */ 106 while (*p) { 107 switch (state) { 108 case STR: 109 if ((*p == '\\') && p[1]) { 110 p++; 111 PARSE_FAIL(i == (PARSE_BUFSIZE - 1)); 112 buf[i++] = *p++; 113 } else if (isquote(*p)) { 114 quote = quote ? 0 : *p; 115 if (dquote) { /* keep quote */ 116 PARSE_FAIL(i == (PARSE_BUFSIZE - 1)); 117 buf[i++] = *p++; 118 } else 119 ++p; 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 } 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 = (char **)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 197 clean(void) 198 { 199 int i; 200 201 for (i = 0; i < MAXARGS; i++) { 202 if (args[i] != NULL) { 203 free(args[i]); 204 args[i] = NULL; 205 } 206 } 207 } 208 209 static int 210 insert(int *argcp, char *buf) 211 { 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 { 222 223 /* XXX search "special variable" space first? */ 224 return (char *)getenv(name); 225 } 226