swish_config.h

harvest是一个下载html网页得机器人

/*
** SWISH Default Configuration File
**
** Kevin Hughes, [e-mail] 
** 3/11/94
**
** The following are user-definable options that you can change
** to fine-tune SWISH's default options.
*/

/* #define NEXTSTEP */

/* You may need to define this if compiling on a NeXTstep machine.
*/

#define INDEXPERMS 0644

/* After SWISH generates an index file, it changes the permissions
** of the file to this mode. Change to the mode you like
** (note that it must be an octal number). If you don't want
** permissions to be changed for you, comment out this line.
*/

#define PLIMIT 80
#define FLIMIT 10000

/* SWISH uses these parameters to automatically mark words as
** being too common while indexing. For instance, if I defined PLIMIT
** as 80 and FLIMIT as 256, SWISH would define a common word as
** a word that occurs in over 80% of all indexed files and over
** 256 files. Making these numbers lower will most likely make your
** index files smaller. Making PLIMIT and FLIMIT small will also
** ensure that searching consumes only so much CPU resources.
*/

#define VERBOSE 1

/* You can define VERBOSE to be a number from 0 to 3. 0 is totally
** silent operation; 3 is very verbose.
*/

#define MAXHITS 5000

/* MAXHITS is the maximum number of results to return from a search.
*/

#define DEFAULT_RULE AND_RULE

/* If a list of search words is specified without booleans,
** SWISH will assume they are connected by a default rule.
** This can be AND_RULE or OR_RULE.
*/

#define MINWORDLIMIT 3

/* This is the minimum length of a word. Anything shorter will not
** be indexed.
*/

#define MAXWORDLIMIT 40

/* This is the maximum length of a word. Anything longer will not
** be indexed.
*/

#define ASCIIENTITIES 1

/* If defined as 1, all entities in search words and indexed
** words will be converted to an ASCII equivalent. For instance,
** with this feature you can index the word "resumé" or
** "resumé" and it will be indexed as the word "resume".
** If defined as 0, only numerical entities will be converted
** to named entities, if they exist.
*/

#define IGNOREALLV 0
#define IGNOREALLC 0
#define IGNOREALLN 0

/* If IGNOREALLV is 1, words containing all vowels won't be indexed.
** If IGNOREALLC is 1, words containing all consonants won't be indexed.
** If IGNOREALLN is 1, words containing all digits won't be indexed.
** Define as 0 to allow words with consistent characters.
** Vowels are defined as "aeiou", digits are "0123456789".
*/

#define IGNOREROWV 60
#define IGNOREROWC 60
#define IGNOREROWN 60

/* IGNOREROWV is the maximum number of consecutive vowels a word can have.
** IGNOREROWC is the maximum number of consecutive consonants a word can have.
** IGNOREROWN is the maximum number of consecutive digits a word can have.
** Vowels are defined as "aeiou", digits are "0123456789".
*/

#define IGNORESAME 100

/* IGNORESAME is the maximum times a character can repeat in a word.
*/

#define WORDCHARS "abcdefghijklmnopqrstuvwxyz懒旅呐魄壬仕掏蜗醒矣哉重寠仝圮轃捺徕沅彐玷殛腱眍镳耱篝貊鴾氝