strs.3

harvest是一个下载html网页得机器人

.\"(c) Copyright 1992, 1993 by Panagiotis Tsirigotis
.\"All rights reserved.  The file named COPYRIGHT specifies the terms 
.\"and conditions for redistribution.
.\"
.\" strs.3,v 1.1.1.1 1994/04/26 19:08:12 hardy Exp
.TH STRS 3X "10 January 1993"
.SH NAME
strs_setup, strs_match, strs_done, strs_search - string matching functions
.SH SYNOPSIS
.LP
.nf
.ft B
#include "str.h"
.LP
.ft B
strs_h strs_setup( flags, pattern )
int flags ;
char *pattern ;
.LP
.ft B
char *strs_match( handle, str, len )
strs_h handle ;
char *str ;
int len ;
.LP
.ft B
void strs_done( handle )
strs_h handle ;
.LP
.ft B
char *strs_search( flags, str, len, pattern )
int flags ;
char *str ;
int len ;
char *pattern ;
.SH DESCRIPTION
.LP
These functions perform string matching. They have been designed with
the assumption that one needs to find a certain pattern in a set of
strings. It is also possible to use them to find if a pattern occurs
in a specific string.
.LP
.B strs_setup()
is used to specify the pattern to look for. It returns a
.I handle
which is used in subsequent string matching operations against
the specified
.I pattern.
The
.I flags
argument has two parts: a search method and generic flags.
The available search methods include the following algorithms:
.RS
.TP 15
.SB STRS_BF
brute force algorithm (also called naive in the literature). 
.TP
.SB STRS_RK
Rabin-Karp algorithm (probabilistic).
.TP
.SB STRS_KMP
Knuth-Morris-Pratt algorithm.
.TP
.SB STRS_SBM
Simple Boyer-Moore (uses only the last occurrence heuristic).
.TP
.SB STRS_BMH
This is the Boyer-Moore algorithm using the last occurrence heuristic
as modified by Horspool (this is faster than the simple Boyer-Moore).
.RE
.LP
The default algorithm is the brute force method.
In practice, the fastest algorithm is the
Boyer-Moore-Horspool one.
.LP
The flags that can be specified include:
.RS
.TP 15
.SB STRS_NOMALLOC
do not allocate space for the pattern. This can be specified if
the pattern space will be available during the matching phase
(i.e. do not use this flag if the pattern space was malloc'ed and
you free it before doing any matching).
.TP
.SB STRS_IGNCASE
perform case-insensitive string matching
(the default is case-sensitive matching).
.RE
.LP
.B strs_match()
tries to match the string specified by
.I str
against the pattern identified by
.I handle.
Since the length of the string is given by the
.I len
argument the string does not need to be NUL-terminated.
.B strs_done()
should be invoked after all matching against the pattern identified by
.I handle
is done.
.LP
.B strs_search()
is equivalent to:
.LP
.PD .1v
.nf
.RS
h = strs_setup( flags, pattern ) ;
p = strs_match( h, str, len ) ;
strs_done( h ) ;
return( p ) ;
.RE
.PD
.SH "RETURN VALUES"
.LP
.B strs_setup()
returns a search handle on success or
.SM NULL
on failure.
.LP
.B strs_match()
and
.B strs_search()
return a pointer to the first occurrence of the pattern in the string or
.SM NULL
if the pattern does not occur in the string.
.SH "SEE ALSO"
Donald E. Knuth, James H. Morris, Vaughan R. Pratt.
Fast pattern matching in strings.
SIAM Journal on Computing, 6(2):323-350, 1977.
.LP
Richard M. Karp, Michael O. Rabin.
Efficient randomized pattern-matching algorithms.
Technical Report TR-31-81,
Aiken Computation Laboratory, Harvard University, 1981.
.LP
Robert S. Boyer, J. Strother Moore.
A fast string-searching algorithm.
Communications of the ACM, 20(10):762-772, 1977.
.LP
N. Horspool.
Practical fast searching in strings.
Software - Practice and Experience, 10:501-506, 1980.
.LP
Thomas H. Cormen, Charles E. Leiserson, Ronald L. Rivest.
Introduction to Algorithms.
McGraw-Hill 1990.