diff options
Diffstat (limited to 'src/libstrongswan/utils/lexparser.c')
-rw-r--r-- | src/libstrongswan/utils/lexparser.c | 137 |
1 files changed, 137 insertions, 0 deletions
diff --git a/src/libstrongswan/utils/lexparser.c b/src/libstrongswan/utils/lexparser.c new file mode 100644 index 000000000..9d3f06593 --- /dev/null +++ b/src/libstrongswan/utils/lexparser.c @@ -0,0 +1,137 @@ +/** + * @file lexparser.c + * + * @brief lexical parser for text-based configuration files + * + */ + +/* + * Copyright (C) 2001-2006 Andreas Steffen, Zuercher Hochschule Winterthur + * + * This program is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License as published by the + * Free Software Foundation; either version 2 of the License, or (at your + * option) any later version. See <http://www.fsf.org/copyleft/gpl.txt>. + * + * This program 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 General Public License + * for more details. + */ + +#include <string.h> + +#include "lexparser.h" + + +/** + * eat whitespace + */ +bool eat_whitespace(chunk_t *src) +{ + while (src->len > 0 && (*src->ptr == ' ' || *src->ptr == '\t')) + { + src->ptr++; src->len--; + } + return src->len > 0 && *src->ptr != '#'; +} + +/** + * compare string with chunk + */ +bool match(const char *pattern, const chunk_t *ch) +{ + return ch->len == strlen(pattern) && strncmp(pattern, ch->ptr, ch->len) == 0; +} + +/** + * extracts a token ending with a given termination symbol + */ +bool extract_token(chunk_t *token, const char termination, chunk_t *src) +{ + u_char *eot = memchr(src->ptr, termination, src->len); + + /* initialize empty token */ + *token = chunk_empty; + + if (eot == NULL) /* termination symbol not found */ + { + return FALSE; + } + + /* extract token */ + token->ptr = src->ptr; + token->len = (u_int)(eot - src->ptr); + + /* advance src pointer after termination symbol */ + src->ptr = eot + 1; + src->len -= (token->len + 1); + + return TRUE; +} + +/** + * fetches a new line terminated by \n or \r\n + */ +bool fetchline(chunk_t *src, chunk_t *line) +{ + if (src->len == 0) /* end of src reached */ + return FALSE; + + if (extract_token(line, '\n', src)) + { + if (line->len > 0 && *(line->ptr + line->len -1) == '\r') + line->len--; /* remove optional \r */ + } + else /*last line ends without newline */ + { + *line = *src; + src->ptr += src->len; + src->len = 0; + } + return TRUE; +} + +err_t extract_value(chunk_t *value, chunk_t *line) +{ + char delimiter = ' '; + + if (!eat_whitespace(line)) + { + *value = chunk_empty; + return NULL; + } + if (*line->ptr == '\'' || *line->ptr == '"') + { + delimiter = *line->ptr; + line->ptr++; line->len--; + } + if (!extract_token(value, delimiter, line)) + { + if (delimiter == ' ') + { + *value = *line; + line->len = 0; + } + else + { + return "missing second delimiter"; + } + } + return NULL; +} + +/** + * extracts a parameter: value pair + */ +err_t extract_parameter_value(chunk_t *name, chunk_t *value, chunk_t *line) +{ + /* extract name */ + if (!extract_token(name,':', line)) + { + return "missing ':'"; + } + + /* extract value */ + return extract_value(value, line); +} |