ferret 0.9.6 → 0.10.0
Sign up to get free protection for your applications and to get access to all the features.
- data/MIT-LICENSE +1 -1
- data/README +12 -24
- data/Rakefile +38 -54
- data/TODO +14 -17
- data/ext/analysis.c +982 -823
- data/ext/analysis.h +133 -76
- data/ext/array.c +96 -58
- data/ext/array.h +40 -13
- data/ext/bitvector.c +476 -118
- data/ext/bitvector.h +264 -22
- data/ext/compound_io.c +217 -229
- data/ext/defines.h +49 -0
- data/ext/document.c +107 -317
- data/ext/document.h +31 -65
- data/ext/except.c +81 -36
- data/ext/except.h +117 -55
- data/ext/extconf.rb +2 -9
- data/ext/ferret.c +211 -104
- data/ext/ferret.h +22 -11
- data/ext/filter.c +97 -82
- data/ext/fs_store.c +348 -367
- data/ext/global.c +226 -188
- data/ext/global.h +44 -26
- data/ext/hash.c +474 -391
- data/ext/hash.h +441 -68
- data/ext/hashset.c +124 -96
- data/ext/hashset.h +169 -20
- data/ext/helper.c +56 -5
- data/ext/helper.h +7 -0
- data/ext/inc/lang.h +29 -49
- data/ext/inc/threading.h +31 -0
- data/ext/ind.c +288 -278
- data/ext/ind.h +68 -0
- data/ext/index.c +5688 -0
- data/ext/index.h +663 -616
- data/ext/lang.h +29 -49
- data/ext/libstemmer.c +3 -3
- data/ext/mem_pool.c +84 -0
- data/ext/mem_pool.h +35 -0
- data/ext/posh.c +1006 -0
- data/ext/posh.h +1007 -0
- data/ext/priorityqueue.c +117 -194
- data/ext/priorityqueue.h +135 -39
- data/ext/q_boolean.c +1305 -1108
- data/ext/q_const_score.c +106 -93
- data/ext/q_filtered_query.c +138 -135
- data/ext/q_fuzzy.c +206 -242
- data/ext/q_match_all.c +94 -80
- data/ext/q_multi_term.c +663 -0
- data/ext/q_parser.c +667 -593
- data/ext/q_phrase.c +992 -555
- data/ext/q_prefix.c +72 -61
- data/ext/q_range.c +235 -210
- data/ext/q_span.c +1480 -1166
- data/ext/q_term.c +273 -246
- data/ext/q_wildcard.c +127 -114
- data/ext/r_analysis.c +1720 -711
- data/ext/r_index.c +3049 -0
- data/ext/r_qparser.c +433 -146
- data/ext/r_search.c +2934 -1993
- data/ext/r_store.c +372 -143
- data/ext/r_utils.c +941 -0
- data/ext/ram_store.c +330 -326
- data/ext/search.c +1291 -668
- data/ext/search.h +403 -702
- data/ext/similarity.c +91 -113
- data/ext/similarity.h +45 -30
- data/ext/sort.c +721 -484
- data/ext/stopwords.c +361 -273
- data/ext/store.c +556 -58
- data/ext/store.h +706 -126
- data/ext/tags +3578 -2780
- data/ext/term_vectors.c +352 -0
- data/ext/threading.h +31 -0
- data/ext/win32.h +54 -0
- data/lib/ferret.rb +5 -17
- data/lib/ferret/document.rb +130 -2
- data/lib/ferret/index.rb +577 -26
- data/lib/ferret/number_tools.rb +157 -0
- data/lib/ferret_version.rb +3 -0
- data/test/test_helper.rb +5 -13
- data/test/unit/analysis/tc_analyzer.rb +513 -1
- data/test/unit/analysis/{ctc_tokenstream.rb → tc_token_stream.rb} +23 -0
- data/test/unit/index/tc_index.rb +183 -240
- data/test/unit/index/tc_index_reader.rb +312 -479
- data/test/unit/index/tc_index_writer.rb +397 -13
- data/test/unit/index/th_doc.rb +269 -206
- data/test/unit/query_parser/tc_query_parser.rb +40 -33
- data/test/unit/search/tc_filter.rb +59 -71
- data/test/unit/search/tc_fuzzy_query.rb +24 -16
- data/test/unit/search/tc_index_searcher.rb +23 -201
- data/test/unit/search/tc_multi_searcher.rb +78 -226
- data/test/unit/search/tc_search_and_sort.rb +93 -81
- data/test/unit/search/tc_sort.rb +23 -23
- data/test/unit/search/tc_sort_field.rb +7 -7
- data/test/unit/search/tc_spans.rb +51 -47
- data/test/unit/search/tm_searcher.rb +339 -0
- data/test/unit/store/tc_fs_store.rb +1 -1
- data/test/unit/store/tm_store_lock.rb +3 -3
- data/test/unit/tc_document.rb +81 -0
- data/test/unit/ts_analysis.rb +1 -1
- data/test/unit/ts_utils.rb +1 -1
- data/test/unit/utils/tc_bit_vector.rb +288 -0
- data/test/unit/utils/tc_number_tools.rb +117 -0
- data/test/unit/utils/tc_priority_queue.rb +106 -0
- metadata +140 -301
- data/CHANGELOG +0 -9
- data/ext/dummy.exe +0 -0
- data/ext/field.c +0 -408
- data/ext/frtio.h +0 -13
- data/ext/inc/except.h +0 -90
- data/ext/index_io.c +0 -382
- data/ext/index_rw.c +0 -2658
- data/ext/lang.c +0 -41
- data/ext/nix_io.c +0 -134
- data/ext/q_multi_phrase.c +0 -380
- data/ext/r_doc.c +0 -582
- data/ext/r_index_io.c +0 -1021
- data/ext/r_term.c +0 -219
- data/ext/term.c +0 -820
- data/ext/termdocs.c +0 -611
- data/ext/vector.c +0 -637
- data/ext/w32_io.c +0 -150
- data/lib/ferret/analysis.rb +0 -11
- data/lib/ferret/analysis/analyzers.rb +0 -112
- data/lib/ferret/analysis/standard_tokenizer.rb +0 -71
- data/lib/ferret/analysis/token.rb +0 -100
- data/lib/ferret/analysis/token_filters.rb +0 -86
- data/lib/ferret/analysis/token_stream.rb +0 -26
- data/lib/ferret/analysis/tokenizers.rb +0 -112
- data/lib/ferret/analysis/word_list_loader.rb +0 -27
- data/lib/ferret/document/document.rb +0 -152
- data/lib/ferret/document/field.rb +0 -312
- data/lib/ferret/index/compound_file_io.rb +0 -338
- data/lib/ferret/index/document_writer.rb +0 -289
- data/lib/ferret/index/field_infos.rb +0 -279
- data/lib/ferret/index/fields_io.rb +0 -181
- data/lib/ferret/index/index.rb +0 -675
- data/lib/ferret/index/index_file_names.rb +0 -33
- data/lib/ferret/index/index_reader.rb +0 -503
- data/lib/ferret/index/index_writer.rb +0 -534
- data/lib/ferret/index/multi_reader.rb +0 -377
- data/lib/ferret/index/multiple_term_doc_pos_enum.rb +0 -98
- data/lib/ferret/index/segment_infos.rb +0 -130
- data/lib/ferret/index/segment_merge_info.rb +0 -49
- data/lib/ferret/index/segment_merge_queue.rb +0 -16
- data/lib/ferret/index/segment_merger.rb +0 -358
- data/lib/ferret/index/segment_reader.rb +0 -412
- data/lib/ferret/index/segment_term_enum.rb +0 -169
- data/lib/ferret/index/segment_term_vector.rb +0 -58
- data/lib/ferret/index/term.rb +0 -53
- data/lib/ferret/index/term_buffer.rb +0 -83
- data/lib/ferret/index/term_doc_enum.rb +0 -291
- data/lib/ferret/index/term_enum.rb +0 -52
- data/lib/ferret/index/term_info.rb +0 -37
- data/lib/ferret/index/term_infos_io.rb +0 -321
- data/lib/ferret/index/term_vector_offset_info.rb +0 -20
- data/lib/ferret/index/term_vectors_io.rb +0 -553
- data/lib/ferret/query_parser.rb +0 -312
- data/lib/ferret/query_parser/query_parser.tab.rb +0 -928
- data/lib/ferret/search.rb +0 -50
- data/lib/ferret/search/boolean_clause.rb +0 -100
- data/lib/ferret/search/boolean_query.rb +0 -299
- data/lib/ferret/search/boolean_scorer.rb +0 -294
- data/lib/ferret/search/caching_wrapper_filter.rb +0 -40
- data/lib/ferret/search/conjunction_scorer.rb +0 -99
- data/lib/ferret/search/disjunction_sum_scorer.rb +0 -205
- data/lib/ferret/search/exact_phrase_scorer.rb +0 -32
- data/lib/ferret/search/explanation.rb +0 -41
- data/lib/ferret/search/field_cache.rb +0 -215
- data/lib/ferret/search/field_doc.rb +0 -31
- data/lib/ferret/search/field_sorted_hit_queue.rb +0 -184
- data/lib/ferret/search/filter.rb +0 -11
- data/lib/ferret/search/filtered_query.rb +0 -130
- data/lib/ferret/search/filtered_term_enum.rb +0 -79
- data/lib/ferret/search/fuzzy_query.rb +0 -154
- data/lib/ferret/search/fuzzy_term_enum.rb +0 -247
- data/lib/ferret/search/hit_collector.rb +0 -34
- data/lib/ferret/search/hit_queue.rb +0 -11
- data/lib/ferret/search/index_searcher.rb +0 -200
- data/lib/ferret/search/match_all_query.rb +0 -104
- data/lib/ferret/search/multi_phrase_query.rb +0 -216
- data/lib/ferret/search/multi_searcher.rb +0 -261
- data/lib/ferret/search/multi_term_query.rb +0 -65
- data/lib/ferret/search/non_matching_scorer.rb +0 -22
- data/lib/ferret/search/phrase_positions.rb +0 -55
- data/lib/ferret/search/phrase_query.rb +0 -214
- data/lib/ferret/search/phrase_scorer.rb +0 -152
- data/lib/ferret/search/prefix_query.rb +0 -54
- data/lib/ferret/search/query.rb +0 -140
- data/lib/ferret/search/query_filter.rb +0 -51
- data/lib/ferret/search/range_filter.rb +0 -103
- data/lib/ferret/search/range_query.rb +0 -139
- data/lib/ferret/search/req_excl_scorer.rb +0 -125
- data/lib/ferret/search/req_opt_sum_scorer.rb +0 -70
- data/lib/ferret/search/score_doc.rb +0 -38
- data/lib/ferret/search/score_doc_comparator.rb +0 -114
- data/lib/ferret/search/scorer.rb +0 -91
- data/lib/ferret/search/similarity.rb +0 -278
- data/lib/ferret/search/sloppy_phrase_scorer.rb +0 -47
- data/lib/ferret/search/sort.rb +0 -112
- data/lib/ferret/search/sort_comparator.rb +0 -60
- data/lib/ferret/search/sort_field.rb +0 -91
- data/lib/ferret/search/spans.rb +0 -12
- data/lib/ferret/search/spans/near_spans_enum.rb +0 -304
- data/lib/ferret/search/spans/span_first_query.rb +0 -79
- data/lib/ferret/search/spans/span_near_query.rb +0 -108
- data/lib/ferret/search/spans/span_not_query.rb +0 -130
- data/lib/ferret/search/spans/span_or_query.rb +0 -176
- data/lib/ferret/search/spans/span_query.rb +0 -25
- data/lib/ferret/search/spans/span_scorer.rb +0 -74
- data/lib/ferret/search/spans/span_term_query.rb +0 -105
- data/lib/ferret/search/spans/span_weight.rb +0 -84
- data/lib/ferret/search/spans/spans_enum.rb +0 -44
- data/lib/ferret/search/term_query.rb +0 -128
- data/lib/ferret/search/term_scorer.rb +0 -183
- data/lib/ferret/search/top_docs.rb +0 -36
- data/lib/ferret/search/top_field_docs.rb +0 -17
- data/lib/ferret/search/weight.rb +0 -54
- data/lib/ferret/search/wildcard_query.rb +0 -26
- data/lib/ferret/search/wildcard_term_enum.rb +0 -61
- data/lib/ferret/stemmers.rb +0 -1
- data/lib/ferret/stemmers/porter_stemmer.rb +0 -218
- data/lib/ferret/store.rb +0 -5
- data/lib/ferret/store/buffered_index_io.rb +0 -190
- data/lib/ferret/store/directory.rb +0 -141
- data/lib/ferret/store/fs_store.rb +0 -381
- data/lib/ferret/store/index_io.rb +0 -245
- data/lib/ferret/store/ram_store.rb +0 -286
- data/lib/ferret/utils.rb +0 -8
- data/lib/ferret/utils/bit_vector.rb +0 -123
- data/lib/ferret/utils/date_tools.rb +0 -138
- data/lib/ferret/utils/number_tools.rb +0 -91
- data/lib/ferret/utils/parameter.rb +0 -41
- data/lib/ferret/utils/priority_queue.rb +0 -120
- data/lib/ferret/utils/string_helper.rb +0 -47
- data/lib/ferret/utils/thread_local.rb +0 -28
- data/lib/ferret/utils/weak_key_hash.rb +0 -60
- data/lib/rferret.rb +0 -37
- data/rake_utils/code_statistics.rb +0 -106
- data/test/benchmark/tb_ram_store.rb +0 -76
- data/test/benchmark/tb_rw_vint.rb +0 -26
- data/test/functional/thread_safety_index_test.rb +0 -81
- data/test/functional/thread_safety_test.rb +0 -137
- data/test/longrunning/tc_numbertools.rb +0 -60
- data/test/longrunning/tm_store.rb +0 -19
- data/test/unit/analysis/ctc_analyzer.rb +0 -532
- data/test/unit/analysis/data/wordfile +0 -6
- data/test/unit/analysis/rtc_letter_tokenizer.rb +0 -20
- data/test/unit/analysis/rtc_lower_case_filter.rb +0 -20
- data/test/unit/analysis/rtc_lower_case_tokenizer.rb +0 -27
- data/test/unit/analysis/rtc_per_field_analyzer_wrapper.rb +0 -39
- data/test/unit/analysis/rtc_porter_stem_filter.rb +0 -16
- data/test/unit/analysis/rtc_standard_analyzer.rb +0 -20
- data/test/unit/analysis/rtc_standard_tokenizer.rb +0 -20
- data/test/unit/analysis/rtc_stop_analyzer.rb +0 -20
- data/test/unit/analysis/rtc_stop_filter.rb +0 -14
- data/test/unit/analysis/rtc_white_space_analyzer.rb +0 -21
- data/test/unit/analysis/rtc_white_space_tokenizer.rb +0 -20
- data/test/unit/analysis/rtc_word_list_loader.rb +0 -32
- data/test/unit/analysis/tc_token.rb +0 -25
- data/test/unit/document/rtc_field.rb +0 -28
- data/test/unit/document/tc_document.rb +0 -47
- data/test/unit/document/tc_field.rb +0 -98
- data/test/unit/index/rtc_compound_file_io.rb +0 -107
- data/test/unit/index/rtc_field_infos.rb +0 -127
- data/test/unit/index/rtc_fields_io.rb +0 -167
- data/test/unit/index/rtc_multiple_term_doc_pos_enum.rb +0 -83
- data/test/unit/index/rtc_segment_infos.rb +0 -74
- data/test/unit/index/rtc_segment_term_docs.rb +0 -17
- data/test/unit/index/rtc_segment_term_enum.rb +0 -60
- data/test/unit/index/rtc_segment_term_vector.rb +0 -71
- data/test/unit/index/rtc_term_buffer.rb +0 -57
- data/test/unit/index/rtc_term_info.rb +0 -19
- data/test/unit/index/rtc_term_infos_io.rb +0 -192
- data/test/unit/index/rtc_term_vectors_io.rb +0 -108
- data/test/unit/index/tc_term.rb +0 -27
- data/test/unit/index/tc_term_voi.rb +0 -18
- data/test/unit/search/rtc_similarity.rb +0 -37
- data/test/unit/search/rtc_sort_field.rb +0 -14
- data/test/unit/search/tc_multi_searcher2.rb +0 -126
- data/test/unit/store/rtc_fs_store.rb +0 -62
- data/test/unit/store/rtc_ram_store.rb +0 -15
- data/test/unit/store/rtm_store.rb +0 -150
- data/test/unit/store/rtm_store_lock.rb +0 -2
- data/test/unit/ts_document.rb +0 -2
- data/test/unit/utils/rtc_bit_vector.rb +0 -73
- data/test/unit/utils/rtc_date_tools.rb +0 -50
- data/test/unit/utils/rtc_number_tools.rb +0 -59
- data/test/unit/utils/rtc_parameter.rb +0 -40
- data/test/unit/utils/rtc_priority_queue.rb +0 -62
- data/test/unit/utils/rtc_string_helper.rb +0 -21
- data/test/unit/utils/rtc_thread.rb +0 -61
- data/test/unit/utils/rtc_weak_key_hash.rb +0 -25
- data/test/utils/number_to_spoken.rb +0 -132
data/ext/store.c
CHANGED
@@ -1,86 +1,584 @@
|
|
1
1
|
#include "store.h"
|
2
|
+
#include <string.h>
|
2
3
|
|
3
|
-
|
4
|
+
#define VINT_MAX_LEN 10
|
5
|
+
#define VINT_END BUFFER_SIZE - VINT_MAX_LEN
|
4
6
|
|
5
|
-
|
6
|
-
*
|
7
|
-
* passed to +func+. If you need to pass more than one argument you should use
|
8
|
-
* a struct. When the function is finished, release the lock.
|
9
|
-
*
|
10
|
-
* @param lock lock to be locked while func is called
|
11
|
-
* @param func function to call with the lock locked
|
12
|
-
* @param arg argument to pass to the function
|
13
|
-
* @throws IO_ERROR if the lock is already locked
|
14
|
-
* @see with_lock_name
|
7
|
+
/*
|
8
|
+
* TODO: add try finally
|
15
9
|
*/
|
16
10
|
void with_lock(Lock *lock, void (*func)(void *arg), void *arg)
|
17
11
|
{
|
18
|
-
|
19
|
-
|
20
|
-
|
21
|
-
|
22
|
-
|
12
|
+
if (!lock->obtain(lock)) {
|
13
|
+
RAISE(IO_ERROR, "couldn't obtain lock \"%s\"", lock->name);
|
14
|
+
}
|
15
|
+
func(arg);
|
16
|
+
lock->release(lock);
|
23
17
|
}
|
24
18
|
|
25
|
-
|
26
|
-
*
|
27
|
-
* +func+ with the lock locked. The argument +arg+ will be passed to +func+.
|
28
|
-
* If you need to pass more than one argument you should use a struct. When
|
29
|
-
* the function is finished, release and destroy the lock.
|
30
|
-
*
|
31
|
-
* @param store store to open the lock in
|
32
|
-
* @param lock_name name of the lock to open
|
33
|
-
* @param func function to call with the lock locked
|
34
|
-
* @param arg argument to pass to the function
|
35
|
-
* @throws IO_ERROR if the lock is already locked
|
36
|
-
* @see with_lock
|
19
|
+
/*
|
20
|
+
* TODO: add try finally
|
37
21
|
*/
|
38
22
|
void with_lock_name(Store *store, char *lock_name,
|
39
|
-
|
23
|
+
void (*func)(void *arg), void *arg)
|
40
24
|
{
|
41
|
-
|
42
|
-
|
43
|
-
|
44
|
-
|
45
|
-
|
46
|
-
|
47
|
-
|
25
|
+
Lock *lock = store->open_lock(store, lock_name);
|
26
|
+
if (!lock->obtain(lock)) {
|
27
|
+
RAISE(LOCK_ERROR, "couldn't obtain lock \"%s\"", lock->name);
|
28
|
+
}
|
29
|
+
func(arg);
|
30
|
+
lock->release(lock);
|
31
|
+
store->close_lock(lock);
|
48
32
|
}
|
49
33
|
|
50
|
-
/**
|
51
|
-
* Remove a reference to the store. If the reference count gets to zero free
|
52
|
-
* all resources used by the store.
|
53
|
-
*
|
54
|
-
* @param store the store to be dereferenced
|
55
|
-
*/
|
56
34
|
void store_deref(Store *store)
|
57
35
|
{
|
58
|
-
|
59
|
-
|
60
|
-
|
61
|
-
|
62
|
-
|
63
|
-
|
36
|
+
mutex_lock(&store->mutex_i);
|
37
|
+
if (--store->ref_cnt == 0) {
|
38
|
+
store->close_i(store);
|
39
|
+
}
|
40
|
+
else {
|
41
|
+
mutex_unlock(&store->mutex_i);
|
42
|
+
}
|
64
43
|
}
|
65
44
|
|
66
45
|
/**
|
67
46
|
* Create a store struct initializing the mutex.
|
68
47
|
*/
|
69
|
-
Store *
|
48
|
+
Store *store_new()
|
70
49
|
{
|
71
|
-
|
72
|
-
|
73
|
-
|
74
|
-
|
75
|
-
|
50
|
+
Store *store = ALLOC(Store);
|
51
|
+
store->ref_cnt = 1;
|
52
|
+
mutex_init(&store->mutex_i, NULL);
|
53
|
+
mutex_init(&store->mutex, NULL);
|
54
|
+
return store;
|
76
55
|
}
|
77
56
|
|
78
57
|
/**
|
79
|
-
* Destroy the store
|
58
|
+
* Destroy the store freeing allocated resources
|
59
|
+
*
|
60
|
+
* @param store the store struct to free
|
80
61
|
*/
|
81
62
|
void store_destroy(Store *store)
|
82
63
|
{
|
83
|
-
|
84
|
-
|
85
|
-
|
64
|
+
mutex_destroy(&store->mutex_i);
|
65
|
+
mutex_destroy(&store->mutex);
|
66
|
+
free(store);
|
67
|
+
}
|
68
|
+
|
69
|
+
/**
|
70
|
+
* Create a newly allocated and initialized OutStream object
|
71
|
+
*
|
72
|
+
* @return a newly allocated and initialized OutStream object
|
73
|
+
*/
|
74
|
+
OutStream *os_new()
|
75
|
+
{
|
76
|
+
OutStream *os = ALLOC(OutStream);
|
77
|
+
os->buf.start = 0;
|
78
|
+
os->buf.pos = 0;
|
79
|
+
os->buf.len = 0;
|
80
|
+
return os;
|
81
|
+
}
|
82
|
+
|
83
|
+
/**
|
84
|
+
* Flush the countents of the OutStream's buffers
|
85
|
+
*
|
86
|
+
* @param the OutStream to flush
|
87
|
+
*/
|
88
|
+
inline void os_flush(OutStream *os)
|
89
|
+
{
|
90
|
+
os->m->flush_i(os, os->buf.buf, os->buf.pos);
|
91
|
+
os->buf.start += os->buf.pos;
|
92
|
+
os->buf.pos = 0;
|
93
|
+
}
|
94
|
+
|
95
|
+
void os_close(OutStream *os)
|
96
|
+
{
|
97
|
+
os_flush(os);
|
98
|
+
os->m->close_i(os);
|
99
|
+
free(os);
|
100
|
+
}
|
101
|
+
|
102
|
+
off_t os_pos(OutStream *os)
|
103
|
+
{
|
104
|
+
return os->buf.start + os->buf.pos;
|
105
|
+
}
|
106
|
+
|
107
|
+
void os_seek(OutStream *os, off_t new_pos)
|
108
|
+
{
|
109
|
+
os_flush(os);
|
110
|
+
os->buf.start = new_pos;
|
111
|
+
os->m->seek_i(os, new_pos);
|
112
|
+
}
|
113
|
+
|
114
|
+
/**
|
115
|
+
* Unsafe alternative to os_write_byte. Only use this method if you know there
|
116
|
+
* is no chance of buffer overflow.
|
117
|
+
*/
|
118
|
+
#define write_byte(os, b) os->buf.buf[os->buf.pos++] = (uchar)b
|
119
|
+
|
120
|
+
/**
|
121
|
+
* Write a single byte +b+ to the OutStream +os+
|
122
|
+
*
|
123
|
+
* @param os the OutStream to write to
|
124
|
+
* @param b the byte to write
|
125
|
+
* @raise IO_ERROR if there is an IO error writing to the filesystem
|
126
|
+
*/
|
127
|
+
inline void os_write_byte(OutStream *os, uchar b)
|
128
|
+
{
|
129
|
+
if (os->buf.pos >= BUFFER_SIZE) {
|
130
|
+
os_flush(os);
|
131
|
+
}
|
132
|
+
write_byte(os, b);
|
133
|
+
}
|
134
|
+
|
135
|
+
void os_write_bytes(OutStream *os, uchar *buf, int len)
|
136
|
+
{
|
137
|
+
if (os->buf.pos > 0) { /* flush buffer */
|
138
|
+
os_flush(os);
|
139
|
+
}
|
140
|
+
|
141
|
+
if (len < BUFFER_SIZE) {
|
142
|
+
os->m->flush_i(os, buf, len);
|
143
|
+
os->buf.start += len;
|
144
|
+
}
|
145
|
+
else {
|
146
|
+
int pos = 0;
|
147
|
+
int size;
|
148
|
+
while (pos < len) {
|
149
|
+
if (len - pos < BUFFER_SIZE) {
|
150
|
+
size = len - pos;
|
151
|
+
}
|
152
|
+
else {
|
153
|
+
size = BUFFER_SIZE;
|
154
|
+
}
|
155
|
+
os->m->flush_i(os, buf + pos, size);
|
156
|
+
pos += size;
|
157
|
+
os->buf.start += size;
|
158
|
+
}
|
159
|
+
}
|
160
|
+
}
|
161
|
+
|
162
|
+
/**
|
163
|
+
* Create a newly allocated and initialized InStream
|
164
|
+
*
|
165
|
+
* @return a newly allocated and initialized InStream
|
166
|
+
*/
|
167
|
+
InStream *is_new()
|
168
|
+
{
|
169
|
+
InStream *is = ALLOC(InStream);
|
170
|
+
is->buf.start = 0;
|
171
|
+
is->buf.pos = 0;
|
172
|
+
is->buf.len = 0;
|
173
|
+
is->ref_cnt_ptr = ALLOC_AND_ZERO(int);
|
174
|
+
return is;
|
175
|
+
}
|
176
|
+
|
177
|
+
/**
|
178
|
+
* Refill the InStream's buffer from the store source (filesystem or memory).
|
179
|
+
*
|
180
|
+
* @param is the InStream to refill
|
181
|
+
* @raise IO_ERROR if there is a error reading from the filesystem
|
182
|
+
* @raise EOF_ERROR if there is an attempt to read past the end of the file
|
183
|
+
*/
|
184
|
+
void is_refill(InStream *is)
|
185
|
+
{
|
186
|
+
off_t start = is->buf.start + is->buf.pos;
|
187
|
+
off_t last = start + BUFFER_SIZE;
|
188
|
+
off_t flen = is->m->length_i(is);
|
189
|
+
|
190
|
+
if (last > flen) { /* don't read past EOF */
|
191
|
+
last = flen;
|
192
|
+
}
|
193
|
+
|
194
|
+
is->buf.len = last - start;
|
195
|
+
if (is->buf.len <= 0) {
|
196
|
+
RAISE(EOF_ERROR, "current pos = %"F_OFF_T_PFX"d, "
|
197
|
+
"file length = %"F_OFF_T_PFX"d", start, flen);
|
198
|
+
}
|
199
|
+
|
200
|
+
is->m->read_i(is, is->buf.buf, is->buf.len);
|
201
|
+
|
202
|
+
is->buf.start = start;
|
203
|
+
is->buf.pos = 0;
|
204
|
+
}
|
205
|
+
|
206
|
+
/**
|
207
|
+
* Unsafe alternative to is_read_byte. Only use this method when you know
|
208
|
+
* there is no chance that you will read past the end of the InStream's
|
209
|
+
* buffer.
|
210
|
+
*/
|
211
|
+
#define read_byte(is) is->buf.buf[is->buf.pos++]
|
212
|
+
|
213
|
+
/**
|
214
|
+
* Read a singly byte (unsigned char) from the InStream +is+.
|
215
|
+
*
|
216
|
+
* @param is the Instream to read from
|
217
|
+
* @return a single unsigned char read from the InStream +is+
|
218
|
+
* @raise IO_ERROR if there is a error reading from the filesystem
|
219
|
+
* @raise EOF_ERROR if there is an attempt to read past the end of the file
|
220
|
+
*/
|
221
|
+
inline uchar is_read_byte(InStream *is)
|
222
|
+
{
|
223
|
+
if (is->buf.pos >= is->buf.len) {
|
224
|
+
is_refill(is);
|
225
|
+
}
|
226
|
+
|
227
|
+
return read_byte(is);
|
228
|
+
}
|
229
|
+
|
230
|
+
off_t is_pos(InStream *is)
|
231
|
+
{
|
232
|
+
return is->buf.start + is->buf.pos;
|
233
|
+
}
|
234
|
+
|
235
|
+
uchar *is_read_bytes(InStream *is, uchar *buf, int len)
|
236
|
+
{
|
237
|
+
int i;
|
238
|
+
off_t start;
|
239
|
+
|
240
|
+
if ((is->buf.pos + len) < is->buf.len) {
|
241
|
+
for (i = 0; i < len; i++) {
|
242
|
+
buf[i] = read_byte(is);
|
243
|
+
}
|
244
|
+
}
|
245
|
+
else { /* read all-at-once */
|
246
|
+
start = is_pos(is);
|
247
|
+
is->m->seek_i(is, start);
|
248
|
+
is->m->read_i(is, buf, len);
|
249
|
+
|
250
|
+
is->buf.start = start + len; /* adjust stream variables */
|
251
|
+
is->buf.pos = 0;
|
252
|
+
is->buf.len = 0; /* trigger refill on read */
|
253
|
+
}
|
254
|
+
return buf;
|
255
|
+
}
|
256
|
+
|
257
|
+
void is_seek(InStream *is, off_t pos)
|
258
|
+
{
|
259
|
+
if (pos >= is->buf.start && pos < (is->buf.start + is->buf.len)) {
|
260
|
+
is->buf.pos = pos - is->buf.start; /* seek within buffer */
|
261
|
+
}
|
262
|
+
else {
|
263
|
+
is->buf.start = pos;
|
264
|
+
is->buf.pos = 0;
|
265
|
+
is->buf.len = 0; /* trigger refill() on read() */
|
266
|
+
is->m->seek_i(is, pos);
|
267
|
+
}
|
268
|
+
}
|
269
|
+
|
270
|
+
void is_close(InStream *is)
|
271
|
+
{
|
272
|
+
if (--(*(is->ref_cnt_ptr)) < 0) {
|
273
|
+
is->m->close_i(is);
|
274
|
+
free(is->ref_cnt_ptr);
|
275
|
+
}
|
276
|
+
free(is);
|
277
|
+
}
|
278
|
+
|
279
|
+
InStream *is_clone(InStream *is)
|
280
|
+
{
|
281
|
+
InStream *new_index_i = ALLOC(InStream);
|
282
|
+
memcpy(new_index_i, is, sizeof(InStream));
|
283
|
+
(*(new_index_i->ref_cnt_ptr))++;
|
284
|
+
return new_index_i;
|
285
|
+
}
|
286
|
+
|
287
|
+
f_i32 is_read_i32(InStream *is)
|
288
|
+
{
|
289
|
+
return ((f_i32)is_read_byte(is) << 24) |
|
290
|
+
((f_i32)is_read_byte(is) << 16) |
|
291
|
+
((f_i32)is_read_byte(is) << 8) |
|
292
|
+
((f_i32)is_read_byte(is));
|
86
293
|
}
|
294
|
+
|
295
|
+
f_i64 is_read_i64(InStream *is)
|
296
|
+
{
|
297
|
+
return ((f_i64)is_read_byte(is) << 56) |
|
298
|
+
((f_i64)is_read_byte(is) << 48) |
|
299
|
+
((f_i64)is_read_byte(is) << 40) |
|
300
|
+
((f_i64)is_read_byte(is) << 32) |
|
301
|
+
((f_i64)is_read_byte(is) << 24) |
|
302
|
+
((f_i64)is_read_byte(is) << 16) |
|
303
|
+
((f_i64)is_read_byte(is) << 8) |
|
304
|
+
((f_i64)is_read_byte(is));
|
305
|
+
}
|
306
|
+
|
307
|
+
f_u32 is_read_u32(InStream *is)
|
308
|
+
{
|
309
|
+
return ((f_u32)is_read_byte(is) << 24) |
|
310
|
+
((f_u32)is_read_byte(is) << 16) |
|
311
|
+
((f_u32)is_read_byte(is) << 8) |
|
312
|
+
((f_u32)is_read_byte(is));
|
313
|
+
}
|
314
|
+
|
315
|
+
f_u64 is_read_u64(InStream *is)
|
316
|
+
{
|
317
|
+
return ((f_u64)is_read_byte(is) << 56) |
|
318
|
+
((f_u64)is_read_byte(is) << 48) |
|
319
|
+
((f_u64)is_read_byte(is) << 40) |
|
320
|
+
((f_u64)is_read_byte(is) << 32) |
|
321
|
+
((f_u64)is_read_byte(is) << 24) |
|
322
|
+
((f_u64)is_read_byte(is) << 16) |
|
323
|
+
((f_u64)is_read_byte(is) << 8) |
|
324
|
+
((f_u64)is_read_byte(is));
|
325
|
+
}
|
326
|
+
|
327
|
+
/* optimized to use unchecked read_byte if there is definitely space */
|
328
|
+
inline unsigned int is_read_vint(InStream *is)
|
329
|
+
{
|
330
|
+
register unsigned int res, b;
|
331
|
+
register int shift = 7;
|
332
|
+
|
333
|
+
if (is->buf.pos > (is->buf.len - VINT_MAX_LEN)) {
|
334
|
+
b = is_read_byte(is);
|
335
|
+
res = b & 0x7F; /* 0x7F = 0b01111111 */
|
336
|
+
|
337
|
+
while ((b & 0x80) != 0) { /* 0x80 = 0b10000000 */
|
338
|
+
b = is_read_byte(is);
|
339
|
+
res |= (b & 0x7F) << shift;
|
340
|
+
shift += 7;
|
341
|
+
}
|
342
|
+
}
|
343
|
+
else { /* unchecked optimization */
|
344
|
+
b = read_byte(is);
|
345
|
+
res = b & 0x7F; /* 0x7F = 0b01111111 */
|
346
|
+
|
347
|
+
while ((b & 0x80) != 0) { /* 0x80 = 0b10000000 */
|
348
|
+
b = read_byte(is);
|
349
|
+
res |= (b & 0x7F) << shift;
|
350
|
+
shift += 7;
|
351
|
+
}
|
352
|
+
}
|
353
|
+
|
354
|
+
return res;
|
355
|
+
}
|
356
|
+
|
357
|
+
/* optimized to use unchecked read_byte if there is definitely space */
|
358
|
+
inline off_t is_read_voff_t(InStream *is)
|
359
|
+
{
|
360
|
+
register off_t res, b;
|
361
|
+
register int shift = 7;
|
362
|
+
|
363
|
+
if (is->buf.pos > (is->buf.len - VINT_MAX_LEN)) {
|
364
|
+
b = is_read_byte(is);
|
365
|
+
res = b & 0x7F; /* 0x7F = 0b01111111 */
|
366
|
+
|
367
|
+
while ((b & 0x80) != 0) { /* 0x80 = 0b10000000 */
|
368
|
+
b = is_read_byte(is);
|
369
|
+
res |= (b & 0x7F) << shift;
|
370
|
+
shift += 7;
|
371
|
+
}
|
372
|
+
}
|
373
|
+
else { /* unchecked optimization */
|
374
|
+
b = read_byte(is);
|
375
|
+
res = b & 0x7F; /* 0x7F = 0b01111111 */
|
376
|
+
|
377
|
+
while ((b & 0x80) != 0) { /* 0x80 = 0b10000000 */
|
378
|
+
b = read_byte(is);
|
379
|
+
res |= (b & 0x7F) << shift;
|
380
|
+
shift += 7;
|
381
|
+
}
|
382
|
+
}
|
383
|
+
|
384
|
+
return res;
|
385
|
+
}
|
386
|
+
|
387
|
+
inline void is_skip_vints(InStream *is, register int cnt)
|
388
|
+
{
|
389
|
+
for (; cnt > 0; cnt--) {
|
390
|
+
while ((is_read_byte(is) & 0x80) != 0) {
|
391
|
+
}
|
392
|
+
}
|
393
|
+
}
|
394
|
+
|
395
|
+
inline void is_read_chars(InStream *is, char *buffer,
|
396
|
+
int off, int len)
|
397
|
+
{
|
398
|
+
int end, i;
|
399
|
+
|
400
|
+
end = off + len;
|
401
|
+
|
402
|
+
for (i = off; i < end; i++) {
|
403
|
+
buffer[i] = is_read_byte(is);
|
404
|
+
}
|
405
|
+
}
|
406
|
+
|
407
|
+
char *is_read_string(InStream *is)
|
408
|
+
{
|
409
|
+
register int length = (int) is_read_vint(is);
|
410
|
+
char *str = ALLOC_N(char, length + 1);
|
411
|
+
str[length] = '\0';
|
412
|
+
|
413
|
+
if (is->buf.pos > (is->buf.len - length)) {
|
414
|
+
register int i;
|
415
|
+
for (i = 0; i < length; i++) {
|
416
|
+
str[i] = is_read_byte(is);
|
417
|
+
}
|
418
|
+
}
|
419
|
+
else { /* unchecked optimization */
|
420
|
+
memcpy(str, is->buf.buf + is->buf.pos, length);
|
421
|
+
is->buf.pos += length;
|
422
|
+
}
|
423
|
+
|
424
|
+
return str;
|
425
|
+
}
|
426
|
+
|
427
|
+
void os_write_i32(OutStream *os, f_i32 num)
|
428
|
+
{
|
429
|
+
os_write_byte(os, (uchar)((num >> 24) & 0xFF));
|
430
|
+
os_write_byte(os, (uchar)((num >> 16) & 0xFF));
|
431
|
+
os_write_byte(os, (uchar)((num >> 8) & 0xFF));
|
432
|
+
os_write_byte(os, (uchar)(num & 0xFF));
|
433
|
+
}
|
434
|
+
|
435
|
+
void os_write_i64(OutStream *os, f_i64 num)
|
436
|
+
{
|
437
|
+
os_write_byte(os, (uchar)((num >> 56) & 0xFF));
|
438
|
+
os_write_byte(os, (uchar)((num >> 48) & 0xFF));
|
439
|
+
os_write_byte(os, (uchar)((num >> 40) & 0xFF));
|
440
|
+
os_write_byte(os, (uchar)((num >> 32) & 0xFF));
|
441
|
+
os_write_byte(os, (uchar)((num >> 24) & 0xFF));
|
442
|
+
os_write_byte(os, (uchar)((num >> 16) & 0xFF));
|
443
|
+
os_write_byte(os, (uchar)((num >> 8) & 0xFF));
|
444
|
+
os_write_byte(os, (uchar)(num & 0xFF));
|
445
|
+
}
|
446
|
+
|
447
|
+
void os_write_u32(OutStream *os, f_u32 num)
|
448
|
+
{
|
449
|
+
os_write_byte(os, (uchar)((num >> 24) & 0xFF));
|
450
|
+
os_write_byte(os, (uchar)((num >> 16) & 0xFF));
|
451
|
+
os_write_byte(os, (uchar)((num >> 8) & 0xFF));
|
452
|
+
os_write_byte(os, (uchar)(num & 0xFF));
|
453
|
+
}
|
454
|
+
|
455
|
+
void os_write_u64(OutStream *os, f_u64 num)
|
456
|
+
{
|
457
|
+
os_write_byte(os, (uchar)((num >> 56) & 0xFF));
|
458
|
+
os_write_byte(os, (uchar)((num >> 48) & 0xFF));
|
459
|
+
os_write_byte(os, (uchar)((num >> 40) & 0xFF));
|
460
|
+
os_write_byte(os, (uchar)((num >> 32) & 0xFF));
|
461
|
+
os_write_byte(os, (uchar)((num >> 24) & 0xFF));
|
462
|
+
os_write_byte(os, (uchar)((num >> 16) & 0xFF));
|
463
|
+
os_write_byte(os, (uchar)((num >> 8) & 0xFF));
|
464
|
+
os_write_byte(os, (uchar)(num & 0xFF));
|
465
|
+
}
|
466
|
+
|
467
|
+
/* optimized to use an unchecked write if there is space */
|
468
|
+
inline void os_write_vint(OutStream *os, register unsigned int num)
|
469
|
+
{
|
470
|
+
if (os->buf.pos > VINT_END) {
|
471
|
+
while (num > 127) {
|
472
|
+
os_write_byte(os, (uchar)((num & 0x7f) | 0x80));
|
473
|
+
num >>= 7;
|
474
|
+
}
|
475
|
+
os_write_byte(os, (uchar)(num));
|
476
|
+
}
|
477
|
+
else {
|
478
|
+
while (num > 127) {
|
479
|
+
write_byte(os, (uchar)((num & 0x7f) | 0x80));
|
480
|
+
num >>= 7;
|
481
|
+
}
|
482
|
+
write_byte(os, (uchar)(num));
|
483
|
+
}
|
484
|
+
}
|
485
|
+
|
486
|
+
/* optimized to use an unchecked write if there is space */
|
487
|
+
inline void os_write_voff_t(OutStream *os, register off_t num)
|
488
|
+
{
|
489
|
+
if (os->buf.pos > VINT_END) {
|
490
|
+
while (num > 127) {
|
491
|
+
os_write_byte(os, (uchar)((num & 0x7f) | 0x80));
|
492
|
+
num >>= 7;
|
493
|
+
}
|
494
|
+
os_write_byte(os, (uchar)num);
|
495
|
+
}
|
496
|
+
else {
|
497
|
+
while (num > 127) {
|
498
|
+
write_byte(os, (uchar)((num & 0x7f) | 0x80));
|
499
|
+
num >>= 7;
|
500
|
+
}
|
501
|
+
write_byte(os, (uchar)num);
|
502
|
+
}
|
503
|
+
}
|
504
|
+
|
505
|
+
void os_write_string(OutStream *os, char *str)
|
506
|
+
{
|
507
|
+
int len = (int)strlen(str);
|
508
|
+
os_write_vint(os, len);
|
509
|
+
|
510
|
+
os_write_bytes(os, (uchar *)str, len);
|
511
|
+
}
|
512
|
+
|
513
|
+
/**
|
514
|
+
* Determine if the filename is the name of a lock file. Return 1 if it is, 0
|
515
|
+
* otherwise.
|
516
|
+
*
|
517
|
+
* @param filename the name of the file to check
|
518
|
+
* @return 1 (true) if the file is a lock file, 0 (false) otherwise
|
519
|
+
*/
|
520
|
+
int file_is_lock(char *filename)
|
521
|
+
{
|
522
|
+
int start = (int) strlen(filename) - 4;
|
523
|
+
return ((start > 0) && (strcmp(LOCK_EXT, &filename[start]) == 0));
|
524
|
+
}
|
525
|
+
|
526
|
+
void is2os_copy_bytes(InStream *is, OutStream *os, int cnt)
|
527
|
+
{
|
528
|
+
int len;
|
529
|
+
uchar buf[BUFFER_SIZE];
|
530
|
+
|
531
|
+
for (; cnt > 0; cnt -= BUFFER_SIZE) {
|
532
|
+
len = ((cnt > BUFFER_SIZE) ? BUFFER_SIZE : cnt);
|
533
|
+
is_read_bytes(is, buf, len);
|
534
|
+
os_write_bytes(os, buf, len);
|
535
|
+
}
|
536
|
+
}
|
537
|
+
|
538
|
+
void is2os_copy_vints(InStream *is, OutStream *os, int cnt)
|
539
|
+
{
|
540
|
+
uchar b;
|
541
|
+
for (; cnt > 0; cnt--) {
|
542
|
+
while (((b = is_read_byte(is)) & 0x80) != 0) {
|
543
|
+
os_write_byte(os, b);
|
544
|
+
}
|
545
|
+
os_write_byte(os, b);
|
546
|
+
}
|
547
|
+
}
|
548
|
+
|
549
|
+
/**
|
550
|
+
* Test argument used to test the store->each function
|
551
|
+
*/
|
552
|
+
struct FileNameConcatArg
|
553
|
+
{
|
554
|
+
char *p;
|
555
|
+
char *end;
|
556
|
+
};
|
557
|
+
|
558
|
+
/**
|
559
|
+
* Test function used to test store->each function
|
560
|
+
*/
|
561
|
+
static void concat_filenames(char *fname, void *arg)
|
562
|
+
{
|
563
|
+
struct FileNameConcatArg *fnca = (struct FileNameConcatArg *)arg;
|
564
|
+
if (fnca->p + strlen(fname) + 2 < fnca->end) {
|
565
|
+
strcpy(fnca->p, fname);
|
566
|
+
fnca->p += strlen(fname);
|
567
|
+
*(fnca->p++) = ',';
|
568
|
+
*(fnca->p++) = ' ';
|
569
|
+
}
|
570
|
+
}
|
571
|
+
|
572
|
+
char *store_to_s(Store *store, char *buf, int buf_size)
|
573
|
+
{
|
574
|
+
struct FileNameConcatArg fnca;
|
575
|
+
|
576
|
+
fnca.p = buf;
|
577
|
+
fnca.end = buf + buf_size;
|
578
|
+
store->each(store, &concat_filenames, &fnca);
|
579
|
+
if (fnca.p > buf + 2) {
|
580
|
+
fnca.p[-2] = '\0';
|
581
|
+
}
|
582
|
+
return buf;
|
583
|
+
}
|
584
|
+
|