diff options
author | Daniel Baumann <daniel.baumann@progress-linux.org> | 2024-04-28 09:49:46 +0000 |
---|---|---|
committer | Daniel Baumann <daniel.baumann@progress-linux.org> | 2024-04-28 09:49:46 +0000 |
commit | 50b37d4a27d3295a29afca2286f1a5a086142cec (patch) | |
tree | 9212f763934ee090ef72d823f559f52ce387f268 /src/lib/token.c | |
parent | Initial commit. (diff) | |
download | freeradius-upstream/3.2.1+dfsg.tar.xz freeradius-upstream/3.2.1+dfsg.zip |
Adding upstream version 3.2.1+dfsg.upstream/3.2.1+dfsgupstream
Signed-off-by: Daniel Baumann <daniel.baumann@progress-linux.org>
Diffstat (limited to 'src/lib/token.c')
-rw-r--r-- | src/lib/token.c | 481 |
1 files changed, 481 insertions, 0 deletions
diff --git a/src/lib/token.c b/src/lib/token.c new file mode 100644 index 0000000..a797862 --- /dev/null +++ b/src/lib/token.c @@ -0,0 +1,481 @@ +/* + * token.c Read the next token from a string. + * Yes it's pretty primitive but effective. + * + * Version: $Id$ + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with this library; if not, write to the Free Software + * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA + * + * Copyright 2000,2006 The FreeRADIUS server project + */ + +RCSID("$Id$") + +#include <freeradius-devel/libradius.h> + +#include <ctype.h> + +const FR_NAME_NUMBER fr_tokens[] = { + { "=~", T_OP_REG_EQ, }, /* order is important! */ + { "!~", T_OP_REG_NE, }, + { "{", T_LCBRACE, }, + { "}", T_RCBRACE, }, + { "(", T_LBRACE, }, + { ")", T_RBRACE, }, + { ",", T_COMMA, }, + { "++", T_OP_INCRM, }, + { "+=", T_OP_ADD, }, + { "-=", T_OP_SUB, }, + { ":=", T_OP_SET, }, + { "=*", T_OP_CMP_TRUE, }, + { "!*", T_OP_CMP_FALSE, }, + { "==", T_OP_CMP_EQ, }, + { "^=", T_OP_PREPEND, }, + { "=", T_OP_EQ, }, + { "!=", T_OP_NE, }, + { ">=", T_OP_GE, }, + { ">", T_OP_GT, }, + { "<=", T_OP_LE, }, + { "<", T_OP_LT, }, + { "#", T_HASH, }, + { ";", T_SEMICOLON, }, + { NULL, 0, }, +}; + +const bool fr_assignment_op[] = { + false, /* invalid token */ + false, /* end of line */ + false, /* { */ + false, /* } */ + false, /* ( */ + false, /* ) 5 */ + false, /* , */ + false, /* ; */ + + true, /* ++ */ + true, /* += */ + true, /* -= 10 */ + true, /* := */ + true, /* = */ + false, /* != */ + false, /* >= */ + false, /* > 15 */ + false, /* <= */ + false, /* < */ + false, /* =~ */ + false, /* !~ */ + false, /* =* 20 */ + false, /* !* */ + false, /* == */ + true, /* ^= */ + false, /* # */ + false, /* bare word 25 */ + false, /* "foo" */ + false, /* 'foo' */ + false, /* `foo` */ + false +}; + +const bool fr_equality_op[] = { + false, /* invalid token */ + false, /* end of line */ + false, /* { */ + false, /* } */ + false, /* ( */ + false, /* ) 5 */ + false, /* , */ + false, /* ; */ + + false, /* ++ */ + false, /* += */ + false, /* -= 10 */ + false, /* := */ + false, /* = */ + true, /* != */ + true, /* >= */ + true, /* > 15 */ + true, /* <= */ + true, /* < */ + true, /* =~ */ + true, /* !~ */ + true, /* =* 20 */ + true, /* !* */ + true, /* == */ + false, /* ^= */ + false, /* # */ + false, /* bare word 25 */ + false, /* "foo" */ + false, /* 'foo' */ + false, /* `foo` */ + false +}; + +const bool fr_str_tok[] = { + false, /* invalid token */ + false, /* end of line */ + false, /* { */ + false, /* } */ + false, /* ( */ + false, /* ) 5 */ + false, /* , */ + false, /* ; */ + + false, /* ++ */ + false, /* += */ + false, /* -= 10 */ + false, /* := */ + false, /* = */ + false, /* != */ + false, /* >= */ + false, /* > 15 */ + false, /* <= */ + false, /* < */ + false, /* =~ */ + false, /* !~ */ + false, /* =* 20 */ + false, /* !* */ + false, /* == */ + false, /* ^= */ + false, /* # */ + true, /* bare word 25 */ + true, /* "foo" */ + true, /* 'foo' */ + true, /* `foo` */ + false +}; + +/* + * This works only as long as special tokens + * are max. 2 characters, but it's fast. + */ +#define TOKEN_MATCH(bptr, tptr) \ + ( (tptr)[0] == (bptr)[0] && \ + ((tptr)[1] == (bptr)[1] || (tptr)[1] == 0)) + +/* + * Read a word from a buffer and advance pointer. + * This function knows about escapes and quotes. + * + * At end-of-line, buf[0] is set to '\0'. + * Returns 0 or special token value. + */ +static FR_TOKEN getthing(char const **ptr, char *buf, int buflen, bool tok, + FR_NAME_NUMBER const *tokenlist, bool unescape) +{ + char *s; + char const *p; + char quote; + bool end = false; + unsigned int x; + FR_NAME_NUMBER const *t; + FR_TOKEN rcode; + + buf[0] = '\0'; + + /* Skip whitespace */ + p = *ptr; + + while (*p && isspace((int) *p)) p++; + + if (!*p) { + *ptr = p; + return T_EOL; + } + + /* + * Might be a 1 or 2 character token. + */ + if (tok) for (t = tokenlist; t->name; t++) { + if (TOKEN_MATCH(p, t->name)) { + strcpy(buf, t->name); + p += strlen(t->name); + + rcode = t->number; + goto done; + } + } + + /* Read word. */ + quote = '\0'; + switch (*p) { + default: + rcode = T_BARE_WORD; + break; + + case '\'': + rcode = T_SINGLE_QUOTED_STRING; + break; + + case '"': + rcode = T_DOUBLE_QUOTED_STRING; + break; + + case '`': + rcode = T_BACK_QUOTED_STRING; + break; + } + + if (rcode != T_BARE_WORD) { + quote = *p; + end = false; + p++; + } + s = buf; + + while (*p && buflen-- > 1) { + /* + * We're looking for strings. Stop on spaces, or + * (if given a token list), on a token, or on a + * comma. + */ + if (!quote) { + if (isspace((int) *p)) { + break; + } + + if (tok) { + for (t = tokenlist; t->name; t++) { + if (TOKEN_MATCH(p, t->name)) { + *s++ = 0; + goto done; + } + } + } + if (*p == ',') break; + + /* + * Copy the character over. + */ + *s++ = *p++; + continue; + } /* else there was a quotation character */ + + /* + * Un-escaped quote character. We're done. + */ + if (*p == quote) { + end = true; + p++; + break; + } + + /* + * Everything but backslash gets copied over. + */ + if (*p != '\\') { + *s++ = *p++; + continue; + } + + /* + * There's nothing after the backslash, it's an error. + */ + if (!p[1]) { + fr_strerror_printf("Unterminated string"); + return T_INVALID; + } + + if (unescape) { + p++; + + switch (*p) { + case 'r': + *s++ = '\r'; + break; + case 'n': + *s++ = '\n'; + break; + case 't': + *s++ = '\t'; + break; + + default: + if (*p >= '0' && *p <= '9' && + sscanf(p, "%3o", &x) == 1) { + *s++ = x; + p += 2; + } else + *s++ = *p; + break; + } + p++; + + } else { + /* + * Convert backslash-quote to quote, but + * leave everything else alone. + */ + if (p[1] == quote) { /* convert '\'' --> ' */ + p++; + } else { + if (buflen < 2) { + fr_strerror_printf("Truncated input"); + return T_INVALID; + } + + *(s++) = *(p++); + } + *(s++) = *(p++); + } + } + + *s++ = 0; + + if (quote && !end) { + fr_strerror_printf("Unterminated string"); + return T_INVALID; + } + +done: + /* Skip whitespace again. */ + while (*p && isspace((int) *p)) p++; + + *ptr = p; + + return rcode; +} + +/* + * Read a "word" - this means we don't honor + * tokens as delimiters. + */ +int getword(char const **ptr, char *buf, int buflen, bool unescape) +{ + return getthing(ptr, buf, buflen, false, fr_tokens, unescape) == T_EOL ? 0 : 1; +} + + +/* + * Read the next word, use tokens as delimiters. + */ +FR_TOKEN gettoken(char const **ptr, char *buf, int buflen, bool unescape) +{ + return getthing(ptr, buf, buflen, true, fr_tokens, unescape); +} + +/* + * Expect an operator. + */ +FR_TOKEN getop(char const **ptr) +{ + char op[3]; + FR_TOKEN rcode; + + rcode = getthing(ptr, op, sizeof(op), true, fr_tokens, false); + if (!fr_assignment_op[rcode] && !fr_equality_op[rcode]) { + fr_strerror_printf("Expected operator"); + return T_INVALID; + } + return rcode; +} + +/* + * Expect a string. + */ +FR_TOKEN getstring(char const **ptr, char *buf, int buflen, bool unescape) +{ + char const *p; + + if (!ptr || !*ptr || !buf) return T_INVALID; + + p = *ptr; + + while (*p && (isspace((int)*p))) p++; + + *ptr = p; + + if ((*p == '"') || (*p == '\'') || (*p == '`')) { + return gettoken(ptr, buf, buflen, unescape); + } + + return getthing(ptr, buf, buflen, false, fr_tokens, unescape); +} + +/* + * Convert a string to an integer + */ +int fr_str2int(FR_NAME_NUMBER const *table, char const *name, int def) +{ + FR_NAME_NUMBER const *this; + + if (!name) { + return def; + } + + for (this = table; this->name != NULL; this++) { + if (strcasecmp(this->name, name) == 0) { + return this->number; + } + } + + return def; +} + +/* + * Convert a string matching part of name to an integer. + */ +int fr_substr2int(FR_NAME_NUMBER const *table, char const *name, int def, int len) +{ + FR_NAME_NUMBER const *this; + size_t max; + + if (!name) { + return def; + } + + for (this = table; this->name != NULL; this++) { + size_t tlen; + + tlen = strlen(this->name); + + /* + * Don't match "request" to user input "req". + */ + if ((len > 0) && (len < (int) tlen)) continue; + + /* + * Match up to the length of the table entry if len is < 0. + */ + max = (len < 0) ? tlen : (unsigned)len; + + if (strncasecmp(this->name, name, max) == 0) { + return this->number; + } + } + + return def; +} + +/* + * Convert an integer to a string. + */ +char const *fr_int2str(FR_NAME_NUMBER const *table, int number, + char const *def) +{ + FR_NAME_NUMBER const *this; + + for (this = table; this->name != NULL; this++) { + if (this->number == number) { + return this->name; + } + } + + return def; +} + +char const *fr_token_name(int token) +{ + return fr_int2str(fr_tokens, token, "???"); +} |