« back to all changes in this revision
Viewing changes to storage/myisam/ft_static.c
- browse files at revision 76.1.1
- compare with another revision
- compare with revision 1219
- download diff
- download tarball
- view history from revision 76.1.1
- Committer: Brian Aker
- Date: 2008-07-06 08:22:57 UTC
- mto: This revision was merged to the branch mainline in revision 78.
- Revision ID: brian@tangent.org-20080706082257-gni9cj1cdjlqomz0
Final removal of fulltext core from myisam.
- files removed:
- storage/myisam/ft_boolean_search.c
- files modified:
- sql/mysqld.cc
added
removed
storage/myisam/ft_static.c
1
/* Copyright (C) 2000-2005 MySQL AB
2
3
This program is free software; you can redistribute it and/or modify
4
it under the terms of the GNU General Public License as published by
5
the Free Software Foundation; version 2 of the License.
6
7
This program is distributed in the hope that it will be useful,
8
but WITHOUT ANY WARRANTY; without even the implied warranty of
9
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
10
GNU General Public License for more details.
11
12
You should have received a copy of the GNU General Public License
13
along with this program; if not, write to the Free Software
14
Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA */
15
16
/* Written by Sergei A. Golubchik, who has a shared copyright to this code */
17
18
#include "ftdefs.h"
19
20
ulong ft_min_word_len=4;
21
ulong ft_max_word_len=HA_FT_MAXCHARLEN;
22
ulong ft_query_expansion_limit=5;
23
char ft_boolean_syntax[]="+ -><()~*:\"\"&|";
24
25
const HA_KEYSEG ft_keysegs[FT_SEGS]={
26
{
27
0, /* charset */
28
HA_FT_WLEN, /* start */
29
0, /* null_pos */
30
0, /* Bit pos */
31
HA_VAR_LENGTH_PART | HA_PACK_KEY, /* flag */
32
HA_FT_MAXBYTELEN, /* length */
33
HA_KEYTYPE_VARTEXT2, /* type */
34
63, /* language (will be overwritten) */
35
0, /* null_bit */
36
2, 0, 0 /* bit_start, bit_end, bit_length */
37
},
38
{
39
/*
40
Note, this (and the last HA_KEYTYPE_END) segment should NOT
41
be packed in any way, otherwise w_search() won't be able to
42
update key entry 'in vivo'
43
*/
44
0, 0, 0, 0, HA_NO_SORT, HA_FT_WLEN, HA_FT_WTYPE, 63, 0, 0, 0, 0
45
}
46
};
47
48
const struct _ft_vft _ft_vft_nlq = {
49
ft_nlq_read_next, ft_nlq_find_relevance, ft_nlq_close_search,
50
ft_nlq_get_relevance, ft_nlq_reinit_search
51
};
52
const struct _ft_vft _ft_vft_boolean = {
53
ft_boolean_read_next, ft_boolean_find_relevance, ft_boolean_close_search,
54
ft_boolean_get_relevance, ft_boolean_reinit_search
55
};
56
57
58
FT_INFO *ft_init_search(uint flags, void *info, uint keynr,
59
uchar *query, uint query_len, CHARSET_INFO *cs,
60
uchar *record)
61
{
62
FT_INFO *res;
63
if (flags & FT_BOOL)
64
res= ft_init_boolean_search((MI_INFO *)info, keynr, query, query_len,cs);
65
else
66
res= ft_init_nlq_search((MI_INFO *)info, keynr, query, query_len, flags,
67
record);
68
return res;
69
}
70
71
const char *ft_stopword_file = 0;
72
const char *ft_precompiled_stopwords[] = {
73
74
#ifdef COMPILE_STOPWORDS_IN
75
76
/* This particular stopword list was taken from SMART distribution
77
ftp://ftp.cs.cornell.edu/pub/smart/smart.11.0.tar.Z
78
it was slightly modified to my taste, though
79
*/
80
81
"a's",
82
"able",
83
"about",
84
"above",
85
"according",
86
"accordingly",
87
"across",
88
"actually",
89
"after",
90
"afterwards",
91
"again",
92
"against",
93
"ain't",
94
"all",
95
"allow",
96
"allows",
97
"almost",
98
"alone",
99
"along",
100
"already",
101
"also",
102
"although",
103
"always",
104
"am",
105
"among",
106
"amongst",
107
"an",
108
"and",
109
"another",
110
"any",
111
"anybody",
112
"anyhow",
113
"anyone",
114
"anything",
115
"anyway",
116
"anyways",
117
"anywhere",
118
"apart",
119
"appear",
120
"appreciate",
121
"appropriate",
122
"are",
123
"aren't",
124
"around",
125
"as",
126
"aside",
127
"ask",
128
"asking",
129
"associated",
130
"at",
131
"available",
132
"away",
133
"awfully",
134
"be",
135
"became",
136
"because",
137
"become",
138
"becomes",
139
"becoming",
140
"been",
141
"before",
142
"beforehand",
143
"behind",
144
"being",
145
"believe",
146
"below",
147
"beside",
148
"besides",
149
"best",
150
"better",
151
"between",
152
"beyond",
153
"both",
154
"brief",
155
"but",
156
"by",
157
"c'mon",
158
"c's",
159
"came",
160
"can",
161
"can't",
162
"cannot",
163
"cant",
164
"cause",
165
"causes",
166
"certain",
167
"certainly",
168
"changes",
169
"clearly",
170
"co",
171
"com",
172
"come",
173
"comes",
174
"concerning",
175
"consequently",
176
"consider",
177
"considering",
178
"contain",
179
"containing",
180
"contains",
181
"corresponding",
182
"could",
183
"couldn't",
184
"course",
185
"currently",
186
"definitely",
187
"described",
188
"despite",
189
"did",
190
"didn't",
191
"different",
192
"do",
193
"does",
194
"doesn't",
195
"doing",
196
"don't",
197
"done",
198
"down",
199
"downwards",
200
"during",
201
"each",
202
"edu",
203
"eg",
204
"eight",
205
"either",
206
"else",
207
"elsewhere",
208
"enough",
209
"entirely",
210
"especially",
211
"et",
212
"etc",
213
"even",
214
"ever",
215
"every",
216
"everybody",
217
"everyone",
218
"everything",
219
"everywhere",
220
"ex",
221
"exactly",
222
"example",
223
"except",
224
"far",
225
"few",
226
"fifth",
227
"first",
228
"five",
229
"followed",
230
"following",
231
"follows",
232
"for",
233
"former",
234
"formerly",
235
"forth",
236
"four",
237
"from",
238
"further",
239
"furthermore",
240
"get",
241
"gets",
242
"getting",
243
"given",
244
"gives",
245
"go",
246
"goes",
247
"going",
248
"gone",
249
"got",
250
"gotten",
251
"greetings",
252
"had",
253
"hadn't",
254
"happens",
255
"hardly",
256
"has",
257
"hasn't",
258
"have",
259
"haven't",
260
"having",
261
"he",
262
"he's",
263
"hello",
264
"help",
265
"hence",
266
"her",
267
"here",
268
"here's",
269
"hereafter",
270
"hereby",
271
"herein",
272
"hereupon",
273
"hers",
274
"herself",
275
"hi",
276
"him",
277
"himself",
278
"his",
279
"hither",
280
"hopefully",
281
"how",
282
"howbeit",
283
"however",
284
"i'd",
285
"i'll",
286
"i'm",
287
"i've",
288
"ie",
289
"if",
290
"ignored",
291
"immediate",
292
"in",
293
"inasmuch",
294
"inc",
295
"indeed",
296
"indicate",
297
"indicated",
298
"indicates",
299
"inner",
300
"insofar",
301
"instead",
302
"into",
303
"inward",
304
"is",
305
"isn't",
306
"it",
307
"it'd",
308
"it'll",
309
"it's",
310
"its",
311
"itself",
312
"just",
313
"keep",
314
"keeps",
315
"kept",
316
"know",
317
"knows",
318
"known",
319
"last",
320
"lately",
321
"later",
322
"latter",
323
"latterly",
324
"least",
325
"less",
326
"lest",
327
"let",
328
"let's",
329
"like",
330
"liked",
331
"likely",
332
"little",
333
"look",
334
"looking",
335
"looks",
336
"ltd",
337
"mainly",
338
"many",
339
"may",
340
"maybe",
341
"me",
342
"mean",
343
"meanwhile",
344
"merely",
345
"might",
346
"more",
347
"moreover",
348
"most",
349
"mostly",
350
"much",
351
"must",
352
"my",
353
"myself",
354
"name",
355
"namely",
356
"nd",
357
"near",
358
"nearly",
359
"necessary",
360
"need",
361
"needs",
362
"neither",
363
"never",
364
"nevertheless",
365
"new",
366
"next",
367
"nine",
368
"no",
369
"nobody",
370
"non",
371
"none",
372
"noone",
373
"nor",
374
"normally",
375
"not",
376
"nothing",
377
"novel",
378
"now",
379
"nowhere",
380
"obviously",
381
"of",
382
"off",
383
"often",
384
"oh",
385
"ok",
386
"okay",
387
"old",
388
"on",
389
"once",
390
"one",
391
"ones",
392
"only",
393
"onto",
394
"or",
395
"other",
396
"others",
397
"otherwise",
398
"ought",
399
"our",
400
"ours",
401
"ourselves",
402
"out",
403
"outside",
404
"over",
405
"overall",
406
"own",
407
"particular",
408
"particularly",
409
"per",
410
"perhaps",
411
"placed",
412
"please",
413
"plus",
414
"possible",
415
"presumably",
416
"probably",
417
"provides",
418
"que",
419
"quite",
420
"qv",
421
"rather",
422
"rd",
423
"re",
424
"really",
425
"reasonably",
426
"regarding",
427
"regardless",
428
"regards",
429
"relatively",
430
"respectively",
431
"right",
432
"said",
433
"same",
434
"saw",
435
"say",
436
"saying",
437
"says",
438
"second",
439
"secondly",
440
"see",
441
"seeing",
442
"seem",
443
"seemed",
444
"seeming",
445
"seems",
446
"seen",
447
"self",
448
"selves",
449
"sensible",
450
"sent",
451
"serious",
452
"seriously",
453
"seven",
454
"several",
455
"shall",
456
"she",
457
"should",
458
"shouldn't",
459
"since",
460
"six",
461
"so",
462
"some",
463
"somebody",
464
"somehow",
465
"someone",
466
"something",
467
"sometime",
468
"sometimes",
469
"somewhat",
470
"somewhere",
471
"soon",
472
"sorry",
473
"specified",
474
"specify",
475
"specifying",
476
"still",
477
"sub",
478
"such",
479
"sup",
480
"sure",
481
"t's",
482
"take",
483
"taken",
484
"tell",
485
"tends",
486
"th",
487
"than",
488
"thank",
489
"thanks",
490
"thanx",
491
"that",
492
"that's",
493
"thats",
494
"the",
495
"their",
496
"theirs",
497
"them",
498
"themselves",
499
"then",
500
"thence",
501
"there",
502
"there's",
503
"thereafter",
504
"thereby",
505
"therefore",
506
"therein",
507
"theres",
508
"thereupon",
509
"these",
510
"they",
511
"they'd",
512
"they'll",
513
"they're",
514
"they've",
515
"think",
516
"third",
517
"this",
518
"thorough",
519
"thoroughly",
520
"those",
521
"though",
522
"three",
523
"through",
524
"throughout",
525
"thru",
526
"thus",
527
"to",
528
"together",
529
"too",
530
"took",
531
"toward",
532
"towards",
533
"tried",
534
"tries",
535
"truly",
536
"try",
537
"trying",
538
"twice",
539
"two",
540
"un",
541
"under",
542
"unfortunately",
543
"unless",
544
"unlikely",
545
"until",
546
"unto",
547
"up",
548
"upon",
549
"us",
550
"use",
551
"used",
552
"useful",
553
"uses",
554
"using",
555
"usually",
556
"value",
557
"various",
558
"very",
559
"via",
560
"viz",
561
"vs",
562
"want",
563
"wants",
564
"was",
565
"wasn't",
566
"way",
567
"we",
568
"we'd",
569
"we'll",
570
"we're",
571
"we've",
572
"welcome",
573
"well",
574
"went",
575
"were",
576
"weren't",
577
"what",
578
"what's",
579
"whatever",
580
"when",
581
"whence",
582
"whenever",
583
"where",
584
"where's",
585
"whereafter",
586
"whereas",
587
"whereby",
588
"wherein",
589
"whereupon",
590
"wherever",
591
"whether",
592
"which",
593
"while",
594
"whither",
595
"who",
596
"who's",
597
"whoever",
598
"whole",
599
"whom",
600
"whose",
601
"why",
602
"will",
603
"willing",
604
"wish",
605
"with",
606
"within",
607
"without",
608
"won't",
609
"wonder",
610
"would",
611
"wouldn't",
612
"yes",
613
"yet",
614
"you",
615
"you'd",
616
"you'll",
617
"you're",
618
"you've",
619
"your",
620
"yours",
621
"yourself",
622
"yourselves",
623
"zero",
624
#endif
625
626
NULL };
627
628
static int ft_default_parser_parse(MYSQL_FTPARSER_PARAM *param)
629
{
630
return param->mysql_parse(param, param->doc, param->length);
631
}
632
633
struct st_mysql_ftparser ft_default_parser=
634
{
635
MYSQL_FTPARSER_INTERFACE_VERSION, ft_default_parser_parse, 0, 0
636
};
637
Loggerhead is a web-based interface for Breezy
Version: 2.0.1