jk-ferret 0.11.8.2
Sign up to get free protection for your applications and to get access to all the features.
- data/CHANGELOG +24 -0
- data/MIT-LICENSE +20 -0
- data/README +90 -0
- data/RELEASE_CHANGES +137 -0
- data/RELEASE_NOTES +60 -0
- data/Rakefile +443 -0
- data/TODO +109 -0
- data/TUTORIAL +231 -0
- data/bin/ferret-browser +79 -0
- data/ext/BZLIB_blocksort.c +1094 -0
- data/ext/BZLIB_bzlib.c +1578 -0
- data/ext/BZLIB_compress.c +672 -0
- data/ext/BZLIB_crctable.c +104 -0
- data/ext/BZLIB_decompress.c +626 -0
- data/ext/BZLIB_huffman.c +205 -0
- data/ext/BZLIB_randtable.c +84 -0
- data/ext/STEMMER_api.c +66 -0
- data/ext/STEMMER_libstemmer.c +93 -0
- data/ext/STEMMER_stem_ISO_8859_1_danish.c +337 -0
- data/ext/STEMMER_stem_ISO_8859_1_dutch.c +624 -0
- data/ext/STEMMER_stem_ISO_8859_1_english.c +1117 -0
- data/ext/STEMMER_stem_ISO_8859_1_finnish.c +762 -0
- data/ext/STEMMER_stem_ISO_8859_1_french.c +1246 -0
- data/ext/STEMMER_stem_ISO_8859_1_german.c +503 -0
- data/ext/STEMMER_stem_ISO_8859_1_hungarian.c +1230 -0
- data/ext/STEMMER_stem_ISO_8859_1_italian.c +1065 -0
- data/ext/STEMMER_stem_ISO_8859_1_norwegian.c +297 -0
- data/ext/STEMMER_stem_ISO_8859_1_porter.c +749 -0
- data/ext/STEMMER_stem_ISO_8859_1_portuguese.c +1017 -0
- data/ext/STEMMER_stem_ISO_8859_1_spanish.c +1093 -0
- data/ext/STEMMER_stem_ISO_8859_1_swedish.c +307 -0
- data/ext/STEMMER_stem_ISO_8859_2_romanian.c +998 -0
- data/ext/STEMMER_stem_KOI8_R_russian.c +700 -0
- data/ext/STEMMER_stem_UTF_8_danish.c +339 -0
- data/ext/STEMMER_stem_UTF_8_dutch.c +634 -0
- data/ext/STEMMER_stem_UTF_8_english.c +1125 -0
- data/ext/STEMMER_stem_UTF_8_finnish.c +768 -0
- data/ext/STEMMER_stem_UTF_8_french.c +1256 -0
- data/ext/STEMMER_stem_UTF_8_german.c +509 -0
- data/ext/STEMMER_stem_UTF_8_hungarian.c +1234 -0
- data/ext/STEMMER_stem_UTF_8_italian.c +1073 -0
- data/ext/STEMMER_stem_UTF_8_norwegian.c +299 -0
- data/ext/STEMMER_stem_UTF_8_porter.c +755 -0
- data/ext/STEMMER_stem_UTF_8_portuguese.c +1023 -0
- data/ext/STEMMER_stem_UTF_8_romanian.c +1004 -0
- data/ext/STEMMER_stem_UTF_8_russian.c +694 -0
- data/ext/STEMMER_stem_UTF_8_spanish.c +1097 -0
- data/ext/STEMMER_stem_UTF_8_swedish.c +309 -0
- data/ext/STEMMER_stem_UTF_8_turkish.c +2205 -0
- data/ext/STEMMER_utilities.c +478 -0
- data/ext/analysis.c +1710 -0
- data/ext/analysis.h +266 -0
- data/ext/api.h +26 -0
- data/ext/array.c +125 -0
- data/ext/array.h +62 -0
- data/ext/bitvector.c +96 -0
- data/ext/bitvector.h +594 -0
- data/ext/bzlib.h +282 -0
- data/ext/bzlib_private.h +503 -0
- data/ext/compound_io.c +384 -0
- data/ext/config.h +52 -0
- data/ext/document.c +159 -0
- data/ext/document.h +63 -0
- data/ext/except.c +102 -0
- data/ext/except.h +176 -0
- data/ext/extconf.rb +15 -0
- data/ext/ferret.c +416 -0
- data/ext/ferret.h +94 -0
- data/ext/field_index.c +262 -0
- data/ext/field_index.h +52 -0
- data/ext/filter.c +157 -0
- data/ext/fs_store.c +493 -0
- data/ext/global.c +458 -0
- data/ext/global.h +302 -0
- data/ext/hash.c +524 -0
- data/ext/hash.h +515 -0
- data/ext/hashset.c +192 -0
- data/ext/hashset.h +215 -0
- data/ext/header.h +58 -0
- data/ext/helper.c +63 -0
- data/ext/helper.h +21 -0
- data/ext/index.c +6804 -0
- data/ext/index.h +935 -0
- data/ext/internal.h +1019 -0
- data/ext/lang.c +10 -0
- data/ext/lang.h +68 -0
- data/ext/libstemmer.h +79 -0
- data/ext/mempool.c +88 -0
- data/ext/mempool.h +43 -0
- data/ext/modules.h +190 -0
- data/ext/multimapper.c +351 -0
- data/ext/multimapper.h +60 -0
- data/ext/posh.c +1006 -0
- data/ext/posh.h +973 -0
- data/ext/priorityqueue.c +149 -0
- data/ext/priorityqueue.h +155 -0
- data/ext/q_boolean.c +1621 -0
- data/ext/q_const_score.c +162 -0
- data/ext/q_filtered_query.c +212 -0
- data/ext/q_fuzzy.c +280 -0
- data/ext/q_match_all.c +149 -0
- data/ext/q_multi_term.c +673 -0
- data/ext/q_parser.c +3103 -0
- data/ext/q_phrase.c +1206 -0
- data/ext/q_prefix.c +98 -0
- data/ext/q_range.c +682 -0
- data/ext/q_span.c +2390 -0
- data/ext/q_term.c +337 -0
- data/ext/q_wildcard.c +167 -0
- data/ext/r_analysis.c +2626 -0
- data/ext/r_index.c +3468 -0
- data/ext/r_qparser.c +635 -0
- data/ext/r_search.c +4490 -0
- data/ext/r_store.c +513 -0
- data/ext/r_utils.c +1131 -0
- data/ext/ram_store.c +476 -0
- data/ext/scanner.c +895 -0
- data/ext/scanner.h +36 -0
- data/ext/scanner_mb.c +6701 -0
- data/ext/scanner_utf8.c +4415 -0
- data/ext/search.c +1864 -0
- data/ext/search.h +953 -0
- data/ext/similarity.c +151 -0
- data/ext/similarity.h +89 -0
- data/ext/sort.c +786 -0
- data/ext/stem_ISO_8859_1_danish.h +16 -0
- data/ext/stem_ISO_8859_1_dutch.h +16 -0
- data/ext/stem_ISO_8859_1_english.h +16 -0
- data/ext/stem_ISO_8859_1_finnish.h +16 -0
- data/ext/stem_ISO_8859_1_french.h +16 -0
- data/ext/stem_ISO_8859_1_german.h +16 -0
- data/ext/stem_ISO_8859_1_hungarian.h +16 -0
- data/ext/stem_ISO_8859_1_italian.h +16 -0
- data/ext/stem_ISO_8859_1_norwegian.h +16 -0
- data/ext/stem_ISO_8859_1_porter.h +16 -0
- data/ext/stem_ISO_8859_1_portuguese.h +16 -0
- data/ext/stem_ISO_8859_1_spanish.h +16 -0
- data/ext/stem_ISO_8859_1_swedish.h +16 -0
- data/ext/stem_ISO_8859_2_romanian.h +16 -0
- data/ext/stem_KOI8_R_russian.h +16 -0
- data/ext/stem_UTF_8_danish.h +16 -0
- data/ext/stem_UTF_8_dutch.h +16 -0
- data/ext/stem_UTF_8_english.h +16 -0
- data/ext/stem_UTF_8_finnish.h +16 -0
- data/ext/stem_UTF_8_french.h +16 -0
- data/ext/stem_UTF_8_german.h +16 -0
- data/ext/stem_UTF_8_hungarian.h +16 -0
- data/ext/stem_UTF_8_italian.h +16 -0
- data/ext/stem_UTF_8_norwegian.h +16 -0
- data/ext/stem_UTF_8_porter.h +16 -0
- data/ext/stem_UTF_8_portuguese.h +16 -0
- data/ext/stem_UTF_8_romanian.h +16 -0
- data/ext/stem_UTF_8_russian.h +16 -0
- data/ext/stem_UTF_8_spanish.h +16 -0
- data/ext/stem_UTF_8_swedish.h +16 -0
- data/ext/stem_UTF_8_turkish.h +16 -0
- data/ext/stopwords.c +410 -0
- data/ext/store.c +698 -0
- data/ext/store.h +799 -0
- data/ext/symbol.c +10 -0
- data/ext/symbol.h +23 -0
- data/ext/term_vectors.c +73 -0
- data/ext/threading.h +31 -0
- data/ext/win32.h +62 -0
- data/lib/ferret.rb +30 -0
- data/lib/ferret/browser.rb +246 -0
- data/lib/ferret/browser/s/global.js +192 -0
- data/lib/ferret/browser/s/style.css +148 -0
- data/lib/ferret/browser/views/document/list.rhtml +49 -0
- data/lib/ferret/browser/views/document/show.rhtml +27 -0
- data/lib/ferret/browser/views/error/index.rhtml +7 -0
- data/lib/ferret/browser/views/help/index.rhtml +8 -0
- data/lib/ferret/browser/views/home/index.rhtml +29 -0
- data/lib/ferret/browser/views/layout.rhtml +22 -0
- data/lib/ferret/browser/views/term-vector/index.rhtml +4 -0
- data/lib/ferret/browser/views/term/index.rhtml +199 -0
- data/lib/ferret/browser/views/term/termdocs.rhtml +1 -0
- data/lib/ferret/browser/webrick.rb +14 -0
- data/lib/ferret/document.rb +130 -0
- data/lib/ferret/field_infos.rb +44 -0
- data/lib/ferret/field_symbol.rb +87 -0
- data/lib/ferret/index.rb +973 -0
- data/lib/ferret/number_tools.rb +157 -0
- data/lib/ferret/version.rb +3 -0
- data/setup.rb +1555 -0
- data/test/long_running/largefile/tc_largefile.rb +46 -0
- data/test/test_all.rb +5 -0
- data/test/test_helper.rb +29 -0
- data/test/test_installed.rb +1 -0
- data/test/threading/number_to_spoken.rb +132 -0
- data/test/threading/thread_safety_index_test.rb +88 -0
- data/test/threading/thread_safety_read_write_test.rb +73 -0
- data/test/threading/thread_safety_test.rb +133 -0
- data/test/unit/analysis/tc_analyzer.rb +550 -0
- data/test/unit/analysis/tc_token_stream.rb +653 -0
- data/test/unit/index/tc_index.rb +867 -0
- data/test/unit/index/tc_index_reader.rb +699 -0
- data/test/unit/index/tc_index_writer.rb +447 -0
- data/test/unit/index/th_doc.rb +332 -0
- data/test/unit/query_parser/tc_query_parser.rb +238 -0
- data/test/unit/search/tc_filter.rb +156 -0
- data/test/unit/search/tc_fuzzy_query.rb +147 -0
- data/test/unit/search/tc_index_searcher.rb +67 -0
- data/test/unit/search/tc_multi_searcher.rb +128 -0
- data/test/unit/search/tc_multiple_search_requests.rb +58 -0
- data/test/unit/search/tc_search_and_sort.rb +179 -0
- data/test/unit/search/tc_sort.rb +49 -0
- data/test/unit/search/tc_sort_field.rb +27 -0
- data/test/unit/search/tc_spans.rb +190 -0
- data/test/unit/search/tm_searcher.rb +436 -0
- data/test/unit/store/tc_fs_store.rb +115 -0
- data/test/unit/store/tc_ram_store.rb +35 -0
- data/test/unit/store/tm_store.rb +34 -0
- data/test/unit/store/tm_store_lock.rb +68 -0
- data/test/unit/tc_document.rb +81 -0
- data/test/unit/tc_field_symbol.rb +26 -0
- data/test/unit/ts_analysis.rb +2 -0
- data/test/unit/ts_index.rb +2 -0
- data/test/unit/ts_largefile.rb +4 -0
- data/test/unit/ts_query_parser.rb +2 -0
- data/test/unit/ts_search.rb +2 -0
- data/test/unit/ts_store.rb +2 -0
- data/test/unit/ts_utils.rb +2 -0
- data/test/unit/utils/tc_bit_vector.rb +295 -0
- data/test/unit/utils/tc_number_tools.rb +117 -0
- data/test/unit/utils/tc_priority_queue.rb +106 -0
- data/test/utils/content_generator.rb +226 -0
- metadata +319 -0
data/ext/bitvector.c
ADDED
@@ -0,0 +1,96 @@
|
|
1
|
+
#include "bitvector.h"
|
2
|
+
#include "internal.h"
|
3
|
+
#include <string.h>
|
4
|
+
|
5
|
+
BitVector *bv_new_capa(int capa)
|
6
|
+
{
|
7
|
+
BitVector *bv = ALLOC_AND_ZERO(BitVector);
|
8
|
+
|
9
|
+
/* The capacity passed by the user is number of bits allowed, however we
|
10
|
+
* store capacity as the number of words (U32) allocated. */
|
11
|
+
bv->capa = max2(TO_WORD(capa), 4);
|
12
|
+
bv->bits = ALLOC_AND_ZERO_N(u32, bv->capa);
|
13
|
+
bv->curr_bit = -1;
|
14
|
+
bv->ref_cnt = 1;
|
15
|
+
return bv;
|
16
|
+
}
|
17
|
+
|
18
|
+
BitVector *bv_new()
|
19
|
+
{
|
20
|
+
return bv_new_capa(BV_INIT_CAPA);
|
21
|
+
}
|
22
|
+
|
23
|
+
void bv_destroy(BitVector *bv)
|
24
|
+
{
|
25
|
+
if (--(bv->ref_cnt) == 0) {
|
26
|
+
free(bv->bits);
|
27
|
+
free(bv);
|
28
|
+
}
|
29
|
+
}
|
30
|
+
|
31
|
+
void bv_clear(BitVector *bv)
|
32
|
+
{
|
33
|
+
memset(bv->bits, 0, bv->capa * sizeof(u32));
|
34
|
+
bv->extends_as_ones = 0;
|
35
|
+
bv->count = 0;
|
36
|
+
bv->size = 0;
|
37
|
+
}
|
38
|
+
|
39
|
+
void bv_scan_reset(BitVector *bv)
|
40
|
+
{
|
41
|
+
bv->curr_bit = -1;
|
42
|
+
}
|
43
|
+
|
44
|
+
int bv_eq(BitVector *bv1, BitVector *bv2)
|
45
|
+
{
|
46
|
+
if (bv1 == bv2) {
|
47
|
+
return true;
|
48
|
+
}
|
49
|
+
|
50
|
+
if (bv1->extends_as_ones != bv2->extends_as_ones) {
|
51
|
+
return false;
|
52
|
+
}
|
53
|
+
|
54
|
+
u32 *bits = bv1->bits;
|
55
|
+
u32 *bits2 = bv2->bits;
|
56
|
+
int min_size = min2(bv1->size, bv2->size);
|
57
|
+
int word_size = TO_WORD(min_size);
|
58
|
+
int ext_word_size = 0;
|
59
|
+
int i;
|
60
|
+
|
61
|
+
for (i = 0; i < word_size; i++) {
|
62
|
+
if (bits[i] != bits2[i]) {
|
63
|
+
return false;
|
64
|
+
}
|
65
|
+
}
|
66
|
+
if (bv1->size > min_size) {
|
67
|
+
bits = bv1->bits;
|
68
|
+
ext_word_size = TO_WORD(bv1->size);
|
69
|
+
}
|
70
|
+
else if (bv2->size > min_size) {
|
71
|
+
bits = bv2->bits;
|
72
|
+
ext_word_size = TO_WORD(bv2->size);
|
73
|
+
}
|
74
|
+
if (ext_word_size) {
|
75
|
+
const u32 expected = (bv1->extends_as_ones ? 0xFFFFFFFF : 0);
|
76
|
+
for (i = word_size; i < ext_word_size; i++) {
|
77
|
+
if (bits[i] != expected) {
|
78
|
+
return false;
|
79
|
+
}
|
80
|
+
}
|
81
|
+
}
|
82
|
+
return true;
|
83
|
+
}
|
84
|
+
|
85
|
+
unsigned long bv_hash(BitVector *bv)
|
86
|
+
{
|
87
|
+
unsigned long hash = 0;
|
88
|
+
const u32 empty_word = bv->extends_as_ones ? 0xFFFFFFFF : 0;
|
89
|
+
int i;
|
90
|
+
for (i = TO_WORD(bv->size) - 1; i >= 0; i--) {
|
91
|
+
const u32 word = bv->bits[i];
|
92
|
+
if (word != empty_word)
|
93
|
+
hash = (hash << 1) ^ word;
|
94
|
+
}
|
95
|
+
return (hash << 1) | bv->extends_as_ones;
|
96
|
+
}
|
data/ext/bitvector.h
ADDED
@@ -0,0 +1,594 @@
|
|
1
|
+
#ifndef FRT_BIT_VECTOR_H
|
2
|
+
#define FRT_BIT_VECTOR_H
|
3
|
+
|
4
|
+
#ifdef __cplusplus
|
5
|
+
extern "C" {
|
6
|
+
#endif
|
7
|
+
|
8
|
+
#include "global.h"
|
9
|
+
|
10
|
+
#define FRT_BV_INIT_CAPA 256
|
11
|
+
|
12
|
+
typedef struct FrtBitVector
|
13
|
+
{
|
14
|
+
/** The bits are held in an array of 32-bit integers */
|
15
|
+
frt_u32 *bits;
|
16
|
+
|
17
|
+
/** size is equal to 1 + the highest order bit set */
|
18
|
+
int size;
|
19
|
+
|
20
|
+
/** capa is the number of words (U32) allocated for the bits */
|
21
|
+
int capa;
|
22
|
+
|
23
|
+
/** count is the running count of bits set. This is kept up to
|
24
|
+
* date by frt_bv_set and frt_bv_unset. You can reset this value
|
25
|
+
* by calling frt_bv_recount */
|
26
|
+
int count;
|
27
|
+
|
28
|
+
/** curr_bit is used by scan_next to record the previously scanned bit */
|
29
|
+
int curr_bit;
|
30
|
+
|
31
|
+
bool extends_as_ones : 1;
|
32
|
+
int ref_cnt;
|
33
|
+
} FrtBitVector;
|
34
|
+
|
35
|
+
/**
|
36
|
+
* Create a new FrtBitVector with a capacity of
|
37
|
+
* +FRT_BV_INIT_CAPA+. Note that the FrtBitVector is growable and will
|
38
|
+
* adjust it's capacity when you use frt_bv_set.
|
39
|
+
*
|
40
|
+
* @return FrtBitVector with a capacity of +FRT_BV_INIT_CAPA+.
|
41
|
+
*/
|
42
|
+
extern FRT_ATTR_MALLOC
|
43
|
+
FrtBitVector *frt_bv_new();
|
44
|
+
|
45
|
+
/**
|
46
|
+
* Create a new FrtBitVector with a capacity of +capa+. Note that the
|
47
|
+
* FrtBitVector is growable and will adjust it's capacity when you use
|
48
|
+
* frt_bv_set.
|
49
|
+
*
|
50
|
+
* @param capa the initial capacity of the FrtBitVector
|
51
|
+
* @return FrtBitVector with a capacity of +capa+.
|
52
|
+
*/
|
53
|
+
extern FRT_ATTR_MALLOC
|
54
|
+
FrtBitVector *frt_bv_new_capa(int capa);
|
55
|
+
|
56
|
+
/**
|
57
|
+
* Destroy a FrtBitVector, freeing all memory allocated to that
|
58
|
+
* FrtBitVector
|
59
|
+
*
|
60
|
+
* @param bv FrtBitVector to destroy
|
61
|
+
*/
|
62
|
+
extern void frt_bv_destroy(FrtBitVector *bv);
|
63
|
+
|
64
|
+
/**
|
65
|
+
* Set the bit at position +index+ with +value+. If +index+ is outside
|
66
|
+
* of the range of the FrtBitVector, that is >= FrtBitVector.size,
|
67
|
+
* FrtBitVector.size will be set to +index+ + 1. If it is greater than
|
68
|
+
* the capacity of the FrtBitVector, the capacity will be expanded to
|
69
|
+
* accomodate.
|
70
|
+
*
|
71
|
+
* @param bv the FrtBitVector to set the bit in
|
72
|
+
* @param index the index of the bit to set
|
73
|
+
* @param value the boolean value
|
74
|
+
*/
|
75
|
+
|
76
|
+
/*
|
77
|
+
* FIXME: if the top set bit is unset, size is not adjusted. This will not
|
78
|
+
* cause any bugs in this code but could cause problems if users are relying
|
79
|
+
* on the fact that size is accurate.
|
80
|
+
*/
|
81
|
+
static FRT_ATTR_ALWAYS_INLINE
|
82
|
+
void frt_bv_set_value(FrtBitVector *bv, int bit, bool value)
|
83
|
+
{
|
84
|
+
frt_u32 *word_p;
|
85
|
+
int word = bit >> 5;
|
86
|
+
frt_u32 bitmask = 1 << (bit & 31);
|
87
|
+
|
88
|
+
/* Check to see if we need to grow the BitVector */
|
89
|
+
if (unlikely(bit >= bv->size)) {
|
90
|
+
bv->size = bit + 1; /* size is max range of bits set */
|
91
|
+
if (word >= bv->capa) {
|
92
|
+
int capa = bv->capa << 1;
|
93
|
+
while (capa <= word) {
|
94
|
+
capa <<= 1;
|
95
|
+
}
|
96
|
+
FRT_REALLOC_N(bv->bits, frt_u32, capa);
|
97
|
+
memset(bv->bits + bv->capa, (bv->extends_as_ones ? 0xFF : 0),
|
98
|
+
sizeof(frt_u32) * (capa - bv->capa));
|
99
|
+
bv->capa = capa;
|
100
|
+
}
|
101
|
+
}
|
102
|
+
|
103
|
+
/* Set the required bit */
|
104
|
+
word_p = &(bv->bits[word]);
|
105
|
+
if ((!!(bitmask & *word_p)) != value) {
|
106
|
+
if (value) {
|
107
|
+
bv->count++;
|
108
|
+
*word_p |= bitmask;
|
109
|
+
}
|
110
|
+
else {
|
111
|
+
bv->count--;
|
112
|
+
*word_p &= ~bitmask;
|
113
|
+
}
|
114
|
+
}
|
115
|
+
}
|
116
|
+
|
117
|
+
/**
|
118
|
+
* Set the bit at position +index+. If +index+ is outside of the range
|
119
|
+
* of the FrtBitVector, that is >= FrtBitVector.size,
|
120
|
+
* FrtBitVector.size will be set to +index+ + 1. If it is greater than
|
121
|
+
* the capacity of the FrtBitVector, the capacity will be expanded to
|
122
|
+
* accomodate.
|
123
|
+
*
|
124
|
+
* @param bv the FrtBitVector to set the bit in
|
125
|
+
* @param index the index of the bit to set
|
126
|
+
*/
|
127
|
+
static FRT_ATTR_ALWAYS_INLINE
|
128
|
+
void frt_bv_set(FrtBitVector *bv, int bit)
|
129
|
+
{
|
130
|
+
frt_bv_set_value(bv, bit, 1);
|
131
|
+
}
|
132
|
+
|
133
|
+
/**
|
134
|
+
* Unsafely set the bit at position +index+. If you choose to use this
|
135
|
+
* function you must create the FrtBitVector with a large enough
|
136
|
+
* capacity to accomodate all of the frt_bv_set_fast operations. You
|
137
|
+
* must also set bits in order and only one time per bit. Otherwise,
|
138
|
+
* use the safe frt_bv_set function.
|
139
|
+
*
|
140
|
+
* So this is ok;
|
141
|
+
* <pre>
|
142
|
+
* FrtBitVector *bv = frt_bv_new_capa(1000);
|
143
|
+
* frt_bv_set_fast(bv, 900);
|
144
|
+
* frt_bv_set_fast(bv, 920);
|
145
|
+
* frt_bv_set_fast(bv, 999);
|
146
|
+
* </pre>
|
147
|
+
*
|
148
|
+
* While these are not ok;
|
149
|
+
* <pre>
|
150
|
+
* FrtBitVector *bv = frt_bv_new_capa(90);
|
151
|
+
* frt_bv_set_fast(bv, 80);
|
152
|
+
* frt_bv_set_fast(bv, 79); // <= Bad: Out of Order
|
153
|
+
* frt_bv_set_fast(bv, 80); // <= Bad: Already set
|
154
|
+
* frt_bv_set_fast(bv, 90); // <= Bad: Out of Range. index must be < capa
|
155
|
+
* </pre>
|
156
|
+
*
|
157
|
+
* @param bv the FrtBitVector to set the bit in
|
158
|
+
* @param index the index of the bit to set
|
159
|
+
*/
|
160
|
+
static FRT_ATTR_ALWAYS_INLINE
|
161
|
+
void frt_bv_set_fast(FrtBitVector *bv, int bit)
|
162
|
+
{
|
163
|
+
bv->count++;
|
164
|
+
bv->size = bit + 1;
|
165
|
+
bv->bits[bit >> 5] |= (1 << (bit & 31));
|
166
|
+
}
|
167
|
+
|
168
|
+
/**
|
169
|
+
* Return 1 if the bit at +index+ was set or 0 otherwise. If +index+
|
170
|
+
* is out of range, that is greater then the BitVectors capacity, it
|
171
|
+
* will also return 0.
|
172
|
+
*
|
173
|
+
* @param bv the FrtBitVector to check in
|
174
|
+
* @param index the index of the bit to check
|
175
|
+
* @return 1 if the bit was set, 0 otherwise
|
176
|
+
*/
|
177
|
+
static FRT_ATTR_ALWAYS_INLINE
|
178
|
+
int frt_bv_get(FrtBitVector *bv, int bit)
|
179
|
+
{
|
180
|
+
/* out of range so return 0 because it can't have been set */
|
181
|
+
if (unlikely(bit >= bv->size)) {
|
182
|
+
return bv->extends_as_ones;
|
183
|
+
}
|
184
|
+
return (bv->bits[bit >> 5] >> (bit & 31)) & 0x01;
|
185
|
+
}
|
186
|
+
|
187
|
+
/**
|
188
|
+
* Unset the bit at position +index+. If the +index+ was out of range,
|
189
|
+
* that is greater than the BitVectors capacity then do
|
190
|
+
* nothing. (frt_bv_get will return 0 in this case anyway).
|
191
|
+
*
|
192
|
+
* @param bv the FrtBitVector to unset the bit in
|
193
|
+
* @param index the index of the bit to unset
|
194
|
+
*/
|
195
|
+
static FRT_ATTR_ALWAYS_INLINE
|
196
|
+
void frt_bv_unset(FrtBitVector *bv, int bit)
|
197
|
+
{
|
198
|
+
frt_bv_set_value(bv, bit, 0);
|
199
|
+
}
|
200
|
+
|
201
|
+
/**
|
202
|
+
* Clear all set bits. This function will set all set bits to 0.
|
203
|
+
*
|
204
|
+
* @param bv the FrtBitVector to clear
|
205
|
+
*/
|
206
|
+
extern void frt_bv_clear(FrtBitVector *bv);
|
207
|
+
|
208
|
+
/**
|
209
|
+
* Resets the set bit count by running through the whole FrtBitVector
|
210
|
+
* and counting all set bits. A running count of the bits is kept by
|
211
|
+
* frt_bv_set, *frt_bv_get and frt_bv_set_fast so this function is
|
212
|
+
* only necessary if the count could have been corrupted somehow or if
|
213
|
+
* the FrtBitVector has been constructed in a different way (for
|
214
|
+
* example being read from the file_system).
|
215
|
+
*
|
216
|
+
* @param bv the FrtBitVector to count the bits in
|
217
|
+
* @return the number of set bits in the FrtBitVector. FrtBitVector.count is also
|
218
|
+
* set
|
219
|
+
*/
|
220
|
+
static FRT_ATTR_ALWAYS_INLINE
|
221
|
+
int frt_bv_recount(FrtBitVector *bv)
|
222
|
+
{
|
223
|
+
unsigned int extra = ((bv->size & 31) >> 3) + 1;
|
224
|
+
unsigned int len = bv->size >> 5;
|
225
|
+
unsigned int idx, count = 0;
|
226
|
+
|
227
|
+
if (bv->extends_as_ones) {
|
228
|
+
for (idx = 0; idx < len; ++idx) {
|
229
|
+
count += frt_count_zeros(bv->bits[idx]);
|
230
|
+
}
|
231
|
+
switch (extra) {
|
232
|
+
case 4: count += frt_count_zeros(bv->bits[idx] | 0x00ffffff);
|
233
|
+
case 3: count += frt_count_zeros(bv->bits[idx] | 0xff00ffff);
|
234
|
+
case 2: count += frt_count_zeros(bv->bits[idx] | 0xffff00ff);
|
235
|
+
case 1: count += frt_count_zeros(bv->bits[idx] | 0xffffff00);
|
236
|
+
}
|
237
|
+
}
|
238
|
+
else {
|
239
|
+
for (idx = 0; idx < len; ++idx) {
|
240
|
+
count += frt_count_ones(bv->bits[idx]);
|
241
|
+
}
|
242
|
+
switch (extra) {
|
243
|
+
case 4: count += frt_count_ones(bv->bits[idx] & 0xff000000);
|
244
|
+
case 3: count += frt_count_ones(bv->bits[idx] & 0x00ff0000);
|
245
|
+
case 2: count += frt_count_ones(bv->bits[idx] & 0x0000ff00);
|
246
|
+
case 1: count += frt_count_ones(bv->bits[idx] & 0x000000ff);
|
247
|
+
}
|
248
|
+
}
|
249
|
+
return bv->count = count;
|
250
|
+
}
|
251
|
+
|
252
|
+
/**
|
253
|
+
* Reset the FrtBitVector for scanning. This function should be called
|
254
|
+
* before using frt_bv_scan_next to scan through all set bits in the
|
255
|
+
* FrtBitVector. This is not necessary when using
|
256
|
+
* frt_bv_scan_next_from.
|
257
|
+
*
|
258
|
+
* @param bv the FrtBitVector to reset for scanning
|
259
|
+
*/
|
260
|
+
extern void frt_bv_scan_reset(FrtBitVector *bv);
|
261
|
+
|
262
|
+
/**
|
263
|
+
* Scan the FrtBitVector for the next set bit after +from+. If no more
|
264
|
+
* bits are set then return -1, otherwise return the index of teh next
|
265
|
+
* set bit.
|
266
|
+
*
|
267
|
+
* @param bv the FrtBitVector to scan
|
268
|
+
* @return the next set bit's index or -1 if no more bits are set
|
269
|
+
*/
|
270
|
+
static FRT_ATTR_ALWAYS_INLINE
|
271
|
+
int frt_bv_scan_next_from(FrtBitVector *bv, const int bit)
|
272
|
+
{
|
273
|
+
frt_u32 pos = bit >> 5;
|
274
|
+
frt_u32 word = bv->bits[pos];
|
275
|
+
|
276
|
+
if (bit >= bv->size)
|
277
|
+
return -1;
|
278
|
+
|
279
|
+
/* Keep only the bits above this position */
|
280
|
+
word &= ~0 << (bit & 31);
|
281
|
+
if (word) {
|
282
|
+
goto done;
|
283
|
+
}
|
284
|
+
else {
|
285
|
+
frt_u32 word_size = FRT_TO_WORD(bv->size);
|
286
|
+
for (pos++; pos < word_size; ++pos)
|
287
|
+
{
|
288
|
+
if ( (word = bv->bits[pos]) )
|
289
|
+
goto done;
|
290
|
+
}
|
291
|
+
}
|
292
|
+
return -1;
|
293
|
+
done:
|
294
|
+
return bv->curr_bit = (pos << 5) + frt_count_trailing_zeros(word);
|
295
|
+
}
|
296
|
+
|
297
|
+
/**
|
298
|
+
* Scan the FrtBitVector for the next set bit. Before using this
|
299
|
+
* function you should reset the FrtBitVector for scanning using
|
300
|
+
* +frt_bv_scan_reset+. You can the repeatedly call frt_bv_scan_next
|
301
|
+
* to get each set bit until it finally returns -1.
|
302
|
+
*
|
303
|
+
* @param bv the FrtBitVector to scan
|
304
|
+
* @return the next set bits index or -1 if no more bits are set
|
305
|
+
*/
|
306
|
+
static FRT_ATTR_ALWAYS_INLINE
|
307
|
+
int frt_bv_scan_next(FrtBitVector *bv)
|
308
|
+
{
|
309
|
+
return frt_bv_scan_next_from(bv, bv->curr_bit + 1);
|
310
|
+
}
|
311
|
+
|
312
|
+
/**
|
313
|
+
* Scan the FrtBitVector for the next unset bit after +from+. If no
|
314
|
+
* more bits are unset then return -1, otherwise return the index of
|
315
|
+
* teh next unset bit.
|
316
|
+
*
|
317
|
+
* @param bv the FrtBitVector to scan
|
318
|
+
* @return the next unset bit's index or -1 if no more bits are unset
|
319
|
+
*/
|
320
|
+
static FRT_ATTR_ALWAYS_INLINE
|
321
|
+
int frt_bv_scan_next_unset_from(FrtBitVector *bv, const int bit)
|
322
|
+
{
|
323
|
+
frt_u32 pos = bit >> 5;
|
324
|
+
frt_u32 word = bv->bits[pos];
|
325
|
+
|
326
|
+
if (bit >= bv->size)
|
327
|
+
return -1;
|
328
|
+
|
329
|
+
/* Set all of the bits below this position */
|
330
|
+
word |= (1 << (bit & 31)) - 1;
|
331
|
+
if (~word) {
|
332
|
+
goto done;
|
333
|
+
}
|
334
|
+
else {
|
335
|
+
frt_u32 word_size = FRT_TO_WORD(bv->size);
|
336
|
+
for (pos++; pos < word_size; ++pos)
|
337
|
+
{
|
338
|
+
if ( ~(word = bv->bits[pos]) )
|
339
|
+
goto done;
|
340
|
+
}
|
341
|
+
}
|
342
|
+
return -1;
|
343
|
+
done:
|
344
|
+
return bv->curr_bit = (pos << 5) + frt_count_trailing_ones(word);
|
345
|
+
}
|
346
|
+
|
347
|
+
/**
|
348
|
+
* Scan the FrtBitVector for the next unset bit. Before using this
|
349
|
+
* function you should reset the FrtBitVector for scanning using
|
350
|
+
* +frt_bv_scan_reset+. You can the repeated call frt_bv_scan_next to
|
351
|
+
* get each unset bit until it finally returns -1.
|
352
|
+
*
|
353
|
+
* @param bv the FrtBitVector to scan
|
354
|
+
* @return the next unset bits index or -1 if no more bits are unset
|
355
|
+
*/
|
356
|
+
static FRT_ATTR_ALWAYS_INLINE
|
357
|
+
int frt_bv_scan_next_unset(FrtBitVector *bv)
|
358
|
+
{
|
359
|
+
return frt_bv_scan_next_unset_from(bv, bv->curr_bit + 1);
|
360
|
+
}
|
361
|
+
|
362
|
+
/**
|
363
|
+
* Check whether the two BitVectors have the same bits set.
|
364
|
+
*
|
365
|
+
* @param bv1 first FrtBitVector to compare
|
366
|
+
* @param bv2 second BitVectors to compare
|
367
|
+
* @return true if bv1 == bv2
|
368
|
+
*/
|
369
|
+
extern int frt_bv_eq(FrtBitVector *bv1, FrtBitVector *bv2);
|
370
|
+
|
371
|
+
/**
|
372
|
+
* Determines a hash value for the FrtBitVector
|
373
|
+
*
|
374
|
+
* @param bv the FrtBitVector to hash
|
375
|
+
* @return A hash value for the FrtBitVector
|
376
|
+
*/
|
377
|
+
extern unsigned long frt_bv_hash(FrtBitVector *bv);
|
378
|
+
|
379
|
+
static FRT_ATTR_ALWAYS_INLINE
|
380
|
+
void frt_bv_capa(FrtBitVector *bv, int capa, int size)
|
381
|
+
{
|
382
|
+
int word_size = FRT_TO_WORD(size);
|
383
|
+
if (bv->capa < capa)
|
384
|
+
{
|
385
|
+
FRT_REALLOC_N(bv->bits, frt_u32, capa);
|
386
|
+
bv->capa = capa;
|
387
|
+
memset(bv->bits + word_size, (bv->extends_as_ones ? 0xFF : 0),
|
388
|
+
sizeof(frt_u32) * (capa - word_size));
|
389
|
+
}
|
390
|
+
bv->size = size;
|
391
|
+
}
|
392
|
+
|
393
|
+
#define frt_bv_and_ext(dest, src, extends_as_ones, i, max) do { \
|
394
|
+
if (extends_as_ones) \
|
395
|
+
memcpy(&dest[i], &src[i], sizeof(*dest)*(max - i)); \
|
396
|
+
else memset(&dest[i], 0x00 , sizeof(*dest)*(max - i)); \
|
397
|
+
} while(0)
|
398
|
+
|
399
|
+
#define frt_bv_or_ext(dest, src, extends_as_ones, i, max) do { \
|
400
|
+
if (extends_as_ones) \
|
401
|
+
memset(&dest[i], 0xFF , sizeof(*dest)*(max - i)); \
|
402
|
+
else memcpy(&dest[i], &src[i], sizeof(*dest)*(max - i)); \
|
403
|
+
} while(0)
|
404
|
+
|
405
|
+
#define frt_bv_xor_ext(dest, src, extends_as_ones, i, max) do { \
|
406
|
+
frt_u32 n = (extends_as_ones ? 0xffffffff : 0); \
|
407
|
+
for (; i < max; ++i) \
|
408
|
+
dest[i] = src[i] ^ n; \
|
409
|
+
} while(0)
|
410
|
+
|
411
|
+
#define FRT_BV_OP(bv, a, b, op, ext_cb) do { \
|
412
|
+
int i; \
|
413
|
+
int a_wsz = FRT_TO_WORD(a->size); \
|
414
|
+
int b_wsz = FRT_TO_WORD(b->size); \
|
415
|
+
int max_size = frt_max2(a->size, b->size); \
|
416
|
+
int min_size = frt_min2(a->size, b->size); \
|
417
|
+
int max_word_size = FRT_TO_WORD(max_size); \
|
418
|
+
int min_word_size = FRT_TO_WORD(min_size); \
|
419
|
+
int capa = frt_max2(frt_round2(max_word_size), 4); \
|
420
|
+
\
|
421
|
+
bv->extends_as_ones = (a->extends_as_ones op b->extends_as_ones); \
|
422
|
+
frt_bv_capa(bv, capa, max_size); \
|
423
|
+
\
|
424
|
+
for (i = 0; i < min_word_size; ++i) \
|
425
|
+
bv->bits[i] = a->bits[i] op b->bits[i]; \
|
426
|
+
\
|
427
|
+
if (a_wsz != b_wsz) { \
|
428
|
+
frt_u32 *bits = a->bits; \
|
429
|
+
bool extends_as_ones = b->extends_as_ones; \
|
430
|
+
if (a_wsz < b_wsz) { \
|
431
|
+
bits = b->bits; \
|
432
|
+
extends_as_ones = a->extends_as_ones; \
|
433
|
+
} \
|
434
|
+
ext_cb(bv->bits, bits, extends_as_ones, i, max_word_size); \
|
435
|
+
} \
|
436
|
+
frt_bv_recount(bv); \
|
437
|
+
} while(0)
|
438
|
+
|
439
|
+
static FRT_ATTR_ALWAYS_INLINE
|
440
|
+
FrtBitVector *frt_bv_and_i(FrtBitVector *bv,
|
441
|
+
FrtBitVector *a, FrtBitVector *b)
|
442
|
+
{
|
443
|
+
FRT_BV_OP(bv, a, b, &, frt_bv_and_ext);
|
444
|
+
return bv;
|
445
|
+
}
|
446
|
+
|
447
|
+
static FRT_ATTR_ALWAYS_INLINE
|
448
|
+
FrtBitVector *frt_bv_or_i(FrtBitVector *bv,
|
449
|
+
FrtBitVector *a, FrtBitVector *b)
|
450
|
+
{
|
451
|
+
FRT_BV_OP(bv, a, b, |, frt_bv_or_ext);
|
452
|
+
return bv;
|
453
|
+
}
|
454
|
+
|
455
|
+
static FRT_ATTR_ALWAYS_INLINE
|
456
|
+
FrtBitVector *frt_bv_xor_i(FrtBitVector *bv,
|
457
|
+
FrtBitVector *a, FrtBitVector *b)
|
458
|
+
{
|
459
|
+
FRT_BV_OP(bv, a, b, ^, frt_bv_xor_ext);
|
460
|
+
return bv;
|
461
|
+
}
|
462
|
+
|
463
|
+
static FRT_ATTR_ALWAYS_INLINE
|
464
|
+
FrtBitVector *frt_bv_not_i(FrtBitVector *bv, FrtBitVector *bv1)
|
465
|
+
{
|
466
|
+
int i;
|
467
|
+
int word_size = FRT_TO_WORD(bv1->size);
|
468
|
+
int capa = frt_max2(frt_round2(word_size), 4);
|
469
|
+
|
470
|
+
bv->extends_as_ones = !bv1->extends_as_ones;
|
471
|
+
frt_bv_capa(bv, capa, bv1->size);
|
472
|
+
|
473
|
+
for (i = 0; i < word_size; i++)
|
474
|
+
bv->bits[i] = ~(bv1->bits[i]);
|
475
|
+
|
476
|
+
memset(bv->bits + word_size, (bv->extends_as_ones ? 0xFF : 0),
|
477
|
+
sizeof(frt_u32) * (bv->capa - word_size));
|
478
|
+
|
479
|
+
frt_bv_recount(bv);
|
480
|
+
return bv;
|
481
|
+
}
|
482
|
+
|
483
|
+
/**
|
484
|
+
* ANDs two BitVectors (+bv1+ and +bv2+) together and return the resultant
|
485
|
+
* FrtBitVector
|
486
|
+
*
|
487
|
+
* @param bv1 first FrtBitVector to AND
|
488
|
+
* @param bv2 second FrtBitVector to AND
|
489
|
+
* @return A FrtBitVector with all bits set that are set in both bv1 and bv2
|
490
|
+
*/
|
491
|
+
static FRT_ATTR_ALWAYS_INLINE
|
492
|
+
FrtBitVector *frt_bv_and(FrtBitVector *bv1, FrtBitVector *bv2)
|
493
|
+
{
|
494
|
+
return frt_bv_and_i(frt_bv_new(), bv1, bv2);
|
495
|
+
}
|
496
|
+
|
497
|
+
/**
|
498
|
+
* ORs two BitVectors (+bv1+ and +bv2+) together and return the resultant
|
499
|
+
* FrtBitVector
|
500
|
+
*
|
501
|
+
* @param bv1 first FrtBitVector to OR
|
502
|
+
* @param bv2 second FrtBitVector to OR
|
503
|
+
* @return A FrtBitVector with all bits set that are set in both bv1 and bv2
|
504
|
+
*/
|
505
|
+
static FRT_ATTR_ALWAYS_INLINE
|
506
|
+
FrtBitVector *frt_bv_or(FrtBitVector *bv1, FrtBitVector *bv2)
|
507
|
+
{
|
508
|
+
return frt_bv_or_i(frt_bv_new(), bv1, bv2);
|
509
|
+
}
|
510
|
+
|
511
|
+
|
512
|
+
/**
|
513
|
+
* XORs two BitVectors (+bv1+ and +bv2+) together and return the resultant
|
514
|
+
* FrtBitVector
|
515
|
+
*
|
516
|
+
* @param bv1 first FrtBitVector to XOR
|
517
|
+
* @param bv2 second FrtBitVector to XOR
|
518
|
+
* @return A FrtBitVector with all bits set that are equal in bv1 and bv2
|
519
|
+
*/
|
520
|
+
static FRT_ATTR_ALWAYS_INLINE
|
521
|
+
FrtBitVector *frt_bv_xor(FrtBitVector *bv1, FrtBitVector *bv2)
|
522
|
+
{
|
523
|
+
return frt_bv_xor_i(frt_bv_new(), bv1, bv2);
|
524
|
+
}
|
525
|
+
|
526
|
+
/**
|
527
|
+
* Returns FrtBitVector with all of +bv+'s bits flipped
|
528
|
+
*
|
529
|
+
* @param bv FrtBitVector to flip
|
530
|
+
* @return A FrtBitVector with all bits set that are set in both bv1 and bv2
|
531
|
+
*/
|
532
|
+
static FRT_ATTR_ALWAYS_INLINE
|
533
|
+
FrtBitVector *frt_bv_not(FrtBitVector *bv)
|
534
|
+
{
|
535
|
+
return frt_bv_not_i(frt_bv_new(), bv);
|
536
|
+
}
|
537
|
+
|
538
|
+
/**
|
539
|
+
* ANDs two BitVectors together +bv1+ and +bv2+ in place of +bv1+
|
540
|
+
*
|
541
|
+
* @param bv1 first FrtBitVector to AND
|
542
|
+
* @param bv2 second FrtBitVector to AND
|
543
|
+
* @return A FrtBitVector
|
544
|
+
* @return bv1 with all bits set that where set in both bv1 and bv2
|
545
|
+
*/
|
546
|
+
static FRT_ATTR_ALWAYS_INLINE
|
547
|
+
FrtBitVector *frt_bv_and_x(FrtBitVector *bv1, FrtBitVector *bv2)
|
548
|
+
{
|
549
|
+
return frt_bv_and_i(bv1, bv1, bv2);
|
550
|
+
}
|
551
|
+
|
552
|
+
/**
|
553
|
+
* ORs two BitVectors together
|
554
|
+
*
|
555
|
+
* @param bv1 first FrtBitVector to OR
|
556
|
+
* @param bv2 second FrtBitVector to OR
|
557
|
+
* @return bv1
|
558
|
+
*/
|
559
|
+
static FRT_ATTR_ALWAYS_INLINE
|
560
|
+
FrtBitVector *frt_bv_or_x(FrtBitVector *bv1, FrtBitVector *bv2)
|
561
|
+
{
|
562
|
+
return frt_bv_or_i(bv1, bv1, bv2);
|
563
|
+
}
|
564
|
+
|
565
|
+
/**
|
566
|
+
* XORs two BitVectors together +bv1+ and +bv2+ in place of +bv1+
|
567
|
+
*
|
568
|
+
* @param bv1 first FrtBitVector to XOR
|
569
|
+
* @param bv2 second FrtBitVector to XOR
|
570
|
+
* @return bv1
|
571
|
+
*/
|
572
|
+
static FRT_ATTR_ALWAYS_INLINE
|
573
|
+
FrtBitVector *frt_bv_xor_x(FrtBitVector *bv1, FrtBitVector *bv2)
|
574
|
+
{
|
575
|
+
return frt_bv_xor_i(bv1, bv1, bv2);
|
576
|
+
}
|
577
|
+
|
578
|
+
/**
|
579
|
+
* Flips all bits in the FrtBitVector +bv+
|
580
|
+
*
|
581
|
+
* @param bv FrtBitVector to flip
|
582
|
+
* @return A +bv+ with all it's bits flipped
|
583
|
+
*/
|
584
|
+
static FRT_ATTR_ALWAYS_INLINE
|
585
|
+
FrtBitVector *frt_bv_not_x(FrtBitVector *bv)
|
586
|
+
{
|
587
|
+
return frt_bv_not_i(bv, bv);
|
588
|
+
}
|
589
|
+
|
590
|
+
#ifdef __cplusplus
|
591
|
+
} // extern "C"
|
592
|
+
#endif
|
593
|
+
|
594
|
+
#endif
|