BOOTSTRAP (kutils) functions
LIBRARY ROUTINE
kparse_string_scan - scan a data string for a specific section
LIBRARY CALL
char *kparse_string_scan(
char *data,
char *key1,
char *key2,
int mode,
char **key1_format,
char **key2_format,
int *status)
INPUT
- data - the data string to search through
- key1 - the regular expression begin key to search for
- key2 - the regular expression end key to search for
- mode - tells the parser which mode to work in:
KIGNORE_CASE Ignore case on a-z and A-Z
KLITERAL Use case information on
a-z and A-Z
OUTPUT
- key1_format - the address of a pointer to hold the returned
begin key that was matched. Sufficient space for
the returned string will be allocated if you
pass in a valid pointer. Note that if this
parameter is passed in as NULL, this routine
will ignore it, and the string that key1 matched
will not be returned.
- key2_format - the address of a pointer to hold the returned
end key that was matched. Sufficient space for
the returned string will be allocated if you
pass in a valid pointer. Note that if this
parameter is passed in as NULL, this routine
will ignore it, and the string that key2 matched
will not be returned.
- status - error status of the search. It can be one of
the following:
KPARSE_OK (parse ok, return data
valid),
KPARSE_NOKEY (couldn't find begin key)
KPARSE_NOEND (couldn't find end key)
KPARSE_DATAERR (data string was invalid)
KPARSE_PARTKEY (data ended with partial
match)
KPARSE_PARTEND (same as above, but on
end key)
KPARSE_NULLKEY (key was NULL)
KPARSE_NULLEND (end key was NULL)
KPARSE_SYNTAXKEY (key had an illegal
syntax)
KPARSE_SYNTAXEND (end key had an illegal
syntax)
RETURN VALUE
This routine returns a pointer to a malloc'ed string
containing the text between the two matched keys.
If an error occured during the search, it will return NULL,
and the error status is set appropriately.
DESCRIPTION
This routine finds a section of data, marked by
begin and end keys, out of a data string. When the begin
and end keys are matched, this routine allocates a string big
enough to hold the data between the keys, and then copies the
data into that space. A pointer to this string is then
returned by this routine. This routine makes calls
to kparse_string_search() which handles the regular
expression searching for the begin and end match keys.
ADDITIONAL INFORMATION
The regular expression syntax supported by the parser
is: '.', '*', '[..]', '[^..]', '$', '^', and '\\'. The explainations
for each syntax is explained below.
.
Match any single character except newline
*
Match the preceding character or ranger of characters 0 or more times. This
matching includes items specified within a [...] construct.
[...] or [^...]
Matches any one character contained within the brackets. If the first
character after the is a ']', then it is included as a character to match.
If the first character after the '[' is a '^', then it will match all characters
NOT included within the []. Finally, the '-' will indicate a range of
characters. For example, [a-z] specifies all characters between and
including the ascii values 'a' and 'z'. If the '-' follows the '[', or is
right before the ']', then it is interpreted literally.
^
If it is the first character of the regular expression, it matches the
beginning of the line.
$
If it is the last character of the regular expression, it matches the end of
line. However, it does NOT put the \\n in your key when the matched key is
returned.
\e
This escapes the meaning of a special character.
EXAMPLES
none
SIDE EFFECTS
This routine mallocs data and sets the value of the key1_format
and key2_format parameters. Thus, the user should pass in
addresses of an unused character pointers for them. The
calling routine is responsible for freeing the space
malloc'ed for the key1_format and key2_format parameters.
This routine mallocs the space for the return string;
and hence, is responsible for freeing the that space
via kfree_and_NULL() when they are done with it.
RESTRICTIONS
It does not support the following regular expression
constructs: or'ing '|', grouping of regular expressions '()',
match one or more times '+', or match n to m times '\\{n,m\\}'.
Finally, the '\\number' and '\\(\\)' constructs have no meaning
for these routines, so they are not supported either.
MODIFICATION
none
FILES
$BOOTSTRAP/objects/library/kutils/src/parse.c
SEE ALSO
kparse_string_search(kutils), kfree(kutils)
COPYRIGHT
Copyright (C) 1993 - 1997, Khoral Research, Inc. ("KRI") All rights reserved.