| 1 |
/* |
| 2 |
** Copyright (C) 1995, 1996, 1997, 1998 Hewlett-Packard Company |
| 3 |
** Originally by Kevin Hughes, kev@kevcom.com, 3/11/94 |
| 4 |
** |
| 5 |
** This program and library is free software; you can redistribute it and/or |
| 6 |
** modify it under the terms of the GNU (Library) General Public License |
| 7 |
** as published by the Free Software Foundation; either version 2 |
| 8 |
** of the License, or any later version. |
| 9 |
** |
| 10 |
** This program is distributed in the hope that it will be useful, |
| 11 |
** but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 12 |
** MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
| 13 |
** GNU (Library) General Public License for more details. |
| 14 |
** |
| 15 |
** You should have received a copy of the GNU (Library) General Public License |
| 16 |
** along with this program; if not, write to the Free Software |
| 17 |
** Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. |
| 18 |
**-------------------------------------------------------------------------- |
| 19 |
** Config file edited by Roy Tennant 2/20/96 |
| 20 |
** Config file edited by Giulia Hill 2/27/97 to increase lenght of |
| 21 |
** words that are indexed |
| 22 |
** Added IGNORELASTCHAR |
| 23 |
** G. Hill 3/12/97 ghill@library.berkeley.edu |
| 24 |
** |
| 25 |
** Added OKNOMETA to allow no failing in case the META name is |
| 26 |
** not listed in the config.h |
| 27 |
** G. Hill 4/15/97 ghill@library.berkeley.edu |
| 28 |
** |
| 29 |
** Added IGNOREFIRSTCHAR |
| 30 |
** G.Hill 10/16/97 ghill@library.berkeley.edu |
| 31 |
**----------------------------------------------------------------------- |
| 32 |
** The following are user-definable options that you can change |
| 33 |
** to fine-tune SWISH's default options. |
| 34 |
** |
| 35 |
** 2001-03-13 rasc moved search boolean words from swish.h |
| 36 |
** |
| 37 |
** 2001-05-23 wsm added ranking weights |
| 38 |
** |
| 39 |
*/ |
| 40 |
|
| 41 |
|
| 42 |
#ifdef __VMS |
| 43 |
#define PROPFILE_EXTENSION "_prop" |
| 44 |
#define WORDDATA_EXTENSION "_wdata" |
| 45 |
#define PRESORTED_EXTENSION "_psort" |
| 46 |
#define BTREE_EXTENSION "_btree" |
| 47 |
#define ARRAY_EXTENSION "_array" |
| 48 |
#else |
| 49 |
#define PROPFILE_EXTENSION ".prop" |
| 50 |
#define WORDDATA_EXTENSION ".wdata" |
| 51 |
#define PRESORTED_EXTENSION ".psort" |
| 52 |
#define BTREE_EXTENSION ".btree" |
| 53 |
#define ARRAY_EXTENSION ".array" |
| 54 |
#endif |
| 55 |
|
| 56 |
/* MIN_PROP_COMPRESS_SIZE sets the limit for which properties are compressed |
| 57 |
* must be compiled with zlib. |
| 58 |
*/ |
| 59 |
#define MIN_PROP_COMPRESS_SIZE 100 |
| 60 |
|
| 61 |
/* This is the character used to replace UTF-8 characters that cannot be |
| 62 |
* converted to 8859-1 Latin-1 character |
| 63 |
*/ |
| 64 |
#define ENCODE_ERROR_CHAR ' ' |
| 65 |
|
| 66 |
/* Defines the file extension to use on the property file. |
| 67 |
*/ |
| 68 |
|
| 69 |
#define MAX_SORT_STRING_LEN 50 |
| 70 |
|
| 71 |
/* MAX_SORT_STRING_LEN defines the max string length to use |
| 72 |
* for sorting properties. Should be long enough to sort ALL |
| 73 |
* file paths or URLs. Useful if using StoreDescription to store |
| 74 |
* a large amount of text. |
| 75 |
*/ |
| 76 |
|
| 77 |
#define USE_DOCPATH_AS_TITLE 1 |
| 78 |
|
| 79 |
/* If USE_DOCPATH_AS_TITLE is defined then documents that do not have |
| 80 |
* a title defined (xml and txt, and HTML documents without a title) |
| 81 |
* will display the document path as the title in results. |
| 82 |
* Documents without a title will sort as a blank title, and not |
| 83 |
* by the document path regardless of this setting. This is a change |
| 84 |
* from versions previous to 2.2. |
| 85 |
*/ |
| 86 |
|
| 87 |
#ifdef __VMS |
| 88 |
#define USE_TEMPFILE_EXTENSION "_temp" |
| 89 |
#else |
| 90 |
#define USE_TEMPFILE_EXTENSION ".temp" |
| 91 |
#endif |
| 92 |
|
| 93 |
/* If USE_TMPFILE_EXTENSION is defined then swish will append the supplied |
| 94 |
* extension onto the index files during indexing, and when indexing is |
| 95 |
* complete will remove the extension by renaming the files. |
| 96 |
* This has two important uses when an index file already exists (and is in use): |
| 97 |
* 1) the old index can be used while indexing is running |
| 98 |
* 2) a failure during indexing will not destroy the existing index |
| 99 |
* |
| 100 |
* Note: This is used instead of a normal temporary file because possible limitation |
| 101 |
* in renaming across file systems. Therefore, the temporary index files are |
| 102 |
* stored in the same directory as the final index files. |
| 103 |
*/ |
| 104 |
|
| 105 |
#define TEMP_FILE_PREFIX "swtmp" |
| 106 |
|
| 107 |
/* TEMP_FILE_PREFIX is prepended to all temporary files. Makes them |
| 108 |
* easier to find. |
| 109 |
*/ |
| 110 |
|
| 111 |
|
| 112 |
#define ALLOW_HTTP_INDEXING_DATA_SOURCE 1 |
| 113 |
#define ALLOW_FILESYSTEM_INDEXING_DATA_SOURCE 1 |
| 114 |
#define ALLOW_EXTERNAL_PROGRAM_DATA_SOURCE 1 |
| 115 |
|
| 116 |
/* These symbols allow compile-time elimination of indexing |
| 117 |
** data sources. Any Data Source that is allowed by these |
| 118 |
** symbols can be selected for indexing from the command line. |
| 119 |
** Comment out any options you do not want to support, but |
| 120 |
** be sure to leave at least one option. |
| 121 |
*/ |
| 122 |
|
| 123 |
#define INDEXPERMS 0644 |
| 124 |
|
| 125 |
/* After SWISH generates an index file, it changes the permissions |
| 126 |
** of the file to this mode. Change to the mode you like |
| 127 |
** (note that it must be an octal number). If you don't want |
| 128 |
** permissions to be changed for you, comment out this line. |
| 129 |
*/ |
| 130 |
|
| 131 |
#define NO_PLIMIT 101 |
| 132 |
|
| 133 |
#define PLIMIT NO_PLIMIT |
| 134 |
#define FLIMIT 10000 |
| 135 |
|
| 136 |
/* SWISH uses these parameters to automatically mark words as |
| 137 |
** being too common while indexing. For instance, if I defined PLIMIT |
| 138 |
** as 80 and FLIMIT as 256, SWISH would define a common word as |
| 139 |
** a word that occurs in over 80% of all indexed files and over |
| 140 |
** 256 files. Making these numbers lower will most likely make your |
| 141 |
** index files smaller. Making PLIMIT and FLIMIT small will also |
| 142 |
** ensure that searching consumes only so much CPU resources. |
| 143 |
*/ |
| 144 |
|
| 145 |
#define VERBOSE 1 |
| 146 |
|
| 147 |
/* You can define VERBOSE to be a number from 0 to 4. 0 is totally |
| 148 |
** silent operation. The default before swish 2.2 was 3 |
| 149 |
*/ |
| 150 |
|
| 151 |
#define _AND_WORD "and" |
| 152 |
#define _OR_WORD "or" |
| 153 |
#define _NOT_WORD "not" |
| 154 |
|
| 155 |
/* |
| 156 |
** these are the default boolean operator words used by swish search |
| 157 |
*/ |
| 158 |
|
| 159 |
#define DEFAULT_RULE AND_RULE |
| 160 |
|
| 161 |
/* If a list of search words is specified without booleans, |
| 162 |
** SWISH will assume they are connected by a default rule. |
| 163 |
** This can be AND_RULE or OR_RULE. |
| 164 |
*/ |
| 165 |
|
| 166 |
#define TITLETOPLINES 12 |
| 167 |
|
| 168 |
/* This is how many lines deep SWISH will look into an HTML file to |
| 169 |
** attempt to find a <TITLE> tag. This has no effect when using the libxml2 parser. |
| 170 |
*/ |
| 171 |
|
| 172 |
|
| 173 |
#define MINWORDLIMIT 1 |
| 174 |
|
| 175 |
/* This is the minimum length of a word. Anything shorter will not |
| 176 |
** be indexed. |
| 177 |
** Do not change it here. Use MinWordLimit in config file |
| 178 |
*/ |
| 179 |
|
| 180 |
#define MAXWORDLIMIT 40 |
| 181 |
|
| 182 |
/* This is the maximum length of a word. Anything longer will not |
| 183 |
** be indexed. |
| 184 |
** Do not change it here. Use MaxWordLimit in config file |
| 185 |
*/ |
| 186 |
|
| 187 |
#define CONVERTHTMLENTITIES 1 |
| 188 |
|
| 189 |
/* If defined as 1, all entities in indexed |
| 190 |
** words will be converted to an ASCII equivalent. For instance, |
| 191 |
** with this feature you can index the word "resumé" or |
| 192 |
** "resumé" and it will be indexed as the word "resume". |
| 193 |
** 2001-01 Do not change it here. Use ConvertHTMLEtities Yes/No in |
| 194 |
** config file |
| 195 |
*/ |
| 196 |
|
| 197 |
#define IGNOREALLV 0 |
| 198 |
#define IGNOREALLC 0 |
| 199 |
#define IGNOREALLN 0 |
| 200 |
|
| 201 |
/* If IGNOREALLV is 1, words containing all vowels won't be indexed. |
| 202 |
** If IGNOREALLC is 1, words containing all consonants won't be indexed. |
| 203 |
** If IGNOREALLN is 1, words containing all digits won't be indexed. |
| 204 |
** Define as 0 to allow words with consistent characters. |
| 205 |
** Vowels are defined as "aeiou", digits are "0123456789". |
| 206 |
*/ |
| 207 |
|
| 208 |
#define IGNOREROWV 60 |
| 209 |
#define IGNOREROWC 60 |
| 210 |
#define IGNOREROWN 60 |
| 211 |
|
| 212 |
/* IGNOREROWV is the maximum number of consecutive vowels a word can have. |
| 213 |
** IGNOREROWC is the maximum number of consecutive consonants a word can have. |
| 214 |
** IGNOREROWN is the maximum number of consecutive digits a word can have. |
| 215 |
** Vowels are defined as "aeiou", digits are "0123456789". |
| 216 |
*/ |
| 217 |
|
| 218 |
#define IGNORESAME 100 |
| 219 |
|
| 220 |
/* IGNORESAME is the maximum times a character can repeat in a word. |
| 221 |
*/ |
| 222 |
/* Dec 6, 2001 - Grabbed "letters" from /usr/local/share/aspell/iso8859-1.dat (http://aspell.sf.net) - moseley */ |
| 223 |
#define WORDCHARS "0123456789abcdefghijklmnopqrstuvwxyzªµºÀÁÂÃÄÅÆÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖØÙÚÛÜÝÞßàáâãäåæçèéêëìíîïðñòóôõöøùúûüýþÿ" |
| 224 |
|
| 225 |
/* |
| 226 |
#define WORDCHARS "abcdefghijklmnopqrstuvwxyzÁÂÃÈýÊËÌÐÝÞÍðÎÏÒÓÔÕØÙÛîèãõßø£ÜíÀ0123456789" |
| 227 |
*/ |
| 228 |
|
| 229 |
/* WORDCHARS is a string of characters which SWISH permits to |
| 230 |
** be in words. Words are defined by these characters. |
| 231 |
** |
| 232 |
** Also note that if you specify the backslash character (\) or |
| 233 |
** double quote (") you need to type a backslash before them to |
| 234 |
** make the compiler understand them. |
| 235 |
*/ |
| 236 |
|
| 237 |
#define BEGINCHARS "0123456789abcdefghijklmnopqrstuvwxyzªµºÀÁÂÃÄÅÆÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖØÙÚÛÜÝÞßàáâãäåæçèéêëìíîïðñòóôõöøùúûüýþÿ" |
| 238 |
|
| 239 |
/* Of the characters that you decide can go into words, this is |
| 240 |
** a list of characters that words can begin with. It should be |
| 241 |
** a subset of (or equal to) WORDCHARS. |
| 242 |
*/ |
| 243 |
|
| 244 |
#define ENDCHARS "0123456789abcdefghijklmnopqrstuvwxyzªµºÀÁÂÃÄÅÆÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖØÙÚÛÜÝÞßàáâãäåæçèéêëìíîïðñòóôõöøùúûüýþÿ" |
| 245 |
|
| 246 |
/* This is the same as BEGINCHARS, except you're testing for |
| 247 |
** valid characters at the ends of words. |
| 248 |
*/ |
| 249 |
|
| 250 |
#define IGNORELASTCHAR "" |
| 251 |
|
| 252 |
/* Array that contains the char that, if considered valid in the middle of |
| 253 |
** a word need to be disreguarded when at the end. It is important to also |
| 254 |
** set the given char's in the ENDCHARS array, otherwise the word will not |
| 255 |
** be indexed because considered invalid. |
| 256 |
** If none just leave the empty list "". Do not erase the line. |
| 257 |
*/ |
| 258 |
|
| 259 |
#define IGNOREFIRSTCHAR "" |
| 260 |
|
| 261 |
/* Array that contains the char that, if considered valid in the middle of |
| 262 |
** a word need to be disreguarded when at the beginning. It is important to also |
| 263 |
** set the given char's in the BEGINCHARS array, otherwise the word will not |
| 264 |
** be indexed because considered invalid. |
| 265 |
** If none just leave the empty list "". Do not erase the line. |
| 266 |
*/ |
| 267 |
|
| 268 |
#define IGNORE_STOPWORDS_IN_QUERY 1 |
| 269 |
|
| 270 |
/* Added JM 1/10/98. Setting this to 0 (default) causes a stopword in |
| 271 |
** an AND_RULE search to create an empty result. Setting it to 1 simply |
| 272 |
** ignores the stopwords and does a search on the remaining words. |
| 273 |
*/ |
| 274 |
|
| 275 |
#define INDEXTAGS 0 |
| 276 |
|
| 277 |
/* Normally, all data in tags in HTML files (except for words in |
| 278 |
** comments or meta tags) is ignored. If you want to index HTML files with the |
| 279 |
** text within tags and all, define this to be 1 and not 0. |
| 280 |
** NOTE: if you set it to 1 you will not be able to do context nor |
| 281 |
** metaNames searches, as tags are just plain text with no specific |
| 282 |
** meaning. |
| 283 |
*/ |
| 284 |
|
| 285 |
// #define BLANK_PROP_VALUE " *BLANK*" |
| 286 |
|
| 287 |
/* This effects how blank properties are stored |
| 288 |
** Normally, blank properties are treated as if they were not even contained int |
| 289 |
** the document. That is: |
| 290 |
** <meta name="author" content=""> |
| 291 |
** is ignored, and no "author" property is stored for that docment. |
| 292 |
** If BLANK_PROP_VALUE is set, then blank properties will be stored |
| 293 |
** but using the string provided as the property value. |
| 294 |
** If you use a leading space, then these properties will sort |
| 295 |
** before other properties (since leading whitespace is removed from |
| 296 |
** properties), and after documents that do not include the property |
| 297 |
*/ |
| 298 |
|
| 299 |
#define RANK_TITLE 4 |
| 300 |
#define RANK_HEADER 3 |
| 301 |
#define RANK_META 3 |
| 302 |
#define RANK_COMMENTS 1 |
| 303 |
#define RANK_EMPHASIZED 0 |
| 304 |
|
| 305 |
/* This symbols affect the weights applied during ranking. Note that they are added |
| 306 |
** together and added to a base rank of 1.0 -- thus defining a rank with a value of |
| 307 |
** 2.0 really means it is ranked (1.0 + 2.0) times greater than normal. |
| 308 |
** A value of 0.0 applies no additional ranking boost. Note that RANK_COMMENTS only |
| 309 |
** applies if you are indexing comment. Be sure you understand how these interact |
| 310 |
** in getrank; don't just go changing these values! |
| 311 |
*/ |
| 312 |
|
| 313 |
#define SPIDERDIRECTORY "./" |
| 314 |
|
| 315 |
#define SWAP_LOC_DEFAULT 0 |
| 316 |
|
| 317 |
/* 2001/08 jmruiz -- Default chunk size - Index will work with blocks of files. This number specifies when to coalesce locations to save memory */ |
| 318 |
#define INDEX_DEFAULT_CHUNK_SIZE 10000 |
| 319 |
|
| 320 |
/* 2001/08 jmruiz -- Default optimal zone size for temporal storage of locations */ |
| 321 |
/* 1<<23 is 8 MB */ |
| 322 |
#define INDEX_DEFAULT_OPTIMAL_CHUNK_ZONE_SIZE_FOR_LOCATIONS 1<<23 |
| 323 |
|
| 324 |
/* 2002/06 Number of swap loc files (-e) */ |
| 325 |
#define MAX_LOC_SWAP_FILES 377 |
| 326 |
|
| 327 |
/* 2001/08 jmruiz -- To avoid emalloc/erealloc in some routines some stack arrays have been added. This is their default size */ |
| 328 |
#define MAX_STACK_POSITIONS 1024 |
| 329 |
|
| 330 |
/* 2001/08 jmruiz -- Do not change this (it must be a unsigned number) */ |
| 331 |
/* This is the maximum size of a block of coalesced locations */ |
| 332 |
#define COALESCE_BUFFER_MAX_SIZE 1<<18 /* (256 KB) */ |
| 333 |
|
| 334 |
/* 2001/08 jmruiz -- File System sort flag - 0 means that filenames |
| 335 |
** will not be indexed - 1 means that filenames will be indexed */ |
| 336 |
#define SORT_FILENAMES 0 |
| 337 |
|
| 338 |
/* 2001/10 jmruiz -- Added BTREE schema to store words */ |
| 339 |
|
| 340 |
//#define USE_BTREE |
| 341 |
|
| 342 |
|
| 343 |
/* 09/00 Jose Ruiz. When set to 1 part of the info is swapped to disk |
| 344 |
** to save memory in the index proccess |
| 345 |
** Do not change it. You can activate this option through the command |
| 346 |
** line (option -e) |
| 347 |
*/ |
| 348 |
|
| 349 |
/* Set this to 1 if you are compiling under Win32 |
| 350 |
define _WIN32 1 |
| 351 |
*/ |
| 352 |
|
| 353 |
/* --- BEGIN PORTING-RELATED SYMBOLS --- */ |
| 354 |
|
| 355 |
#ifdef _WIN32 |
| 356 |
#define NO_SYMBOLIC_FILE_LINKS /* Win32 has no symbolic links */ |
| 357 |
#endif |
| 358 |
|
| 359 |
#ifdef __VMS |
| 360 |
#define NO_SYMBOLIC_FILE_LINKS /* VMS has no symbolic links */ |
| 361 |
#endif |
| 362 |
|
| 363 |
#ifdef _WIN32 |
| 364 |
#undef INDEXPERMS /* Win32 version doesn't use chmod() */ |
| 365 |
#endif |
| 366 |
|
| 367 |
#ifdef _WIN32 |
| 368 |
typedef int pid_t; /* process ID */ |
| 369 |
#endif |
| 370 |
|
| 371 |
//#ifdef _WIN32 |
| 372 |
//#define TMPDIR "c:\\windows\\temp" |
| 373 |
//#elif defined(__VMS) |
| 374 |
//#define TMPDIR "sys$scratch:" |
| 375 |
//#else |
| 376 |
//#define TMPDIR "/var/tmp" |
| 377 |
//#endif |
| 378 |
|
| 379 |
|
| 380 |
/* Default Delimiter of phrase search */ |
| 381 |
#define PHRASE_DELIMITER_CHAR '"' |
| 382 |
|
| 383 |
|
| 384 |
/* |
| 385 |
* Binary files must be open with the "b" option under Win32, so all |
| 386 |
* fopen() calls to index files have to go through these routines to |
| 387 |
* keep the code portable. |
| 388 |
* Note: text files should be opened normally, without the "b" option, |
| 389 |
* otherwise end-of-line processing is not done correctly (on Win32). |
| 390 |
*/ |
| 391 |
#define F_READ_BINARY "rb" |
| 392 |
#define F_WRITE_BINARY "wb" |
| 393 |
#define F_READWRITE_BINARY "rb+" |
| 394 |
|
| 395 |
#define F_READ_TEXT "r" |
| 396 |
#define F_WRITE_TEXT "w" |
| 397 |
#define F_READWRITE_TEXT "r+" |
| 398 |
|
| 399 |
|
| 400 |
|
| 401 |
/* #define NEXTSTEP */ |
| 402 |
|
| 403 |
/* You may need to define this if compiling on a NeXTstep machine. |
| 404 |
*/ |
| 405 |
|
| 406 |
/* --- END PORTING-RELATED SYMBOLS --- */ |
| 407 |
|
Parent Directory
| 
Revision Log
| 
Revision Graph