reftable: add a heap-based priority queue for reftable records

This is needed to create a merged view multiple reftables Signed-off-by: Han-Wen Nienhuys <hanwen@google.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>

Han-Wen Nienhuys committed Oct 7, 2021 at 20:25 UTC 3b34f636df9e6c32c2e9b819cd5ad907497c62b7
6 files changed +224
Makefile
+2
@@ -2462,6 +2462,7 @@ REFTABLE_OBJS += reftable/block.o
2462 REFTABLE_OBJS += reftable/blocksource.o
2463 REFTABLE_OBJS += reftable/iter.o
2464 REFTABLE_OBJS += reftable/publicbasics.o
2465 +REFTABLE_OBJS += reftable/pq.o
2466 REFTABLE_OBJS += reftable/reader.o
2467 REFTABLE_OBJS += reftable/record.o
2468 REFTABLE_OBJS += reftable/refname.o
@@ -2472,6 +2473,7 @@ REFTABLE_OBJS += reftable/writer.o
2473
2474 REFTABLE_TEST_OBJS += reftable/basics_test.o
2475 REFTABLE_TEST_OBJS += reftable/block_test.o
2476 +REFTABLE_TEST_OBJS += reftable/pq_test.o
2477 REFTABLE_TEST_OBJS += reftable/record_test.o
2478 REFTABLE_TEST_OBJS += reftable/readwrite_test.o
2479 REFTABLE_TEST_OBJS += reftable/test_framework.o
reftable/pq.c new
+105
@@ -0,0 +1,105 @@
1 +/*
2 +Copyright 2020 Google LLC
3 +
4 +Use of this source code is governed by a BSD-style
5 +license that can be found in the LICENSE file or at
6 +https://developers.google.com/open-source/licenses/bsd
7 +*/
8 +
9 +#include "pq.h"
10 +
11 +#include "reftable-record.h"
12 +#include "system.h"
13 +#include "basics.h"
14 +
15 +int pq_less(struct pq_entry *a, struct pq_entry *b)
16 +{
17 + struct strbuf ak = STRBUF_INIT;
18 + struct strbuf bk = STRBUF_INIT;
19 + int cmp = 0;
20 + reftable_record_key(&a->rec, &ak);
21 + reftable_record_key(&b->rec, &bk);
22 +
23 + cmp = strbuf_cmp(&ak, &bk);
24 +
25 + strbuf_release(&ak);
26 + strbuf_release(&bk);
27 +
28 + if (cmp == 0)
29 + return a->index > b->index;
30 +
31 + return cmp < 0;
32 +}
33 +
34 +struct pq_entry merged_iter_pqueue_top(struct merged_iter_pqueue pq)
35 +{
36 + return pq.heap[0];
37 +}
38 +
39 +int merged_iter_pqueue_is_empty(struct merged_iter_pqueue pq)
40 +{
41 + return pq.len == 0;
42 +}
43 +
44 +struct pq_entry merged_iter_pqueue_remove(struct merged_iter_pqueue *pq)
45 +{
46 + int i = 0;
47 + struct pq_entry e = pq->heap[0];
48 + pq->heap[0] = pq->heap[pq->len - 1];
49 + pq->len--;
50 +
51 + i = 0;
52 + while (i < pq->len) {
53 + int min = i;
54 + int j = 2 * i + 1;
55 + int k = 2 * i + 2;
56 + if (j < pq->len && pq_less(&pq->heap[j], &pq->heap[i])) {
57 + min = j;
58 + }
59 + if (k < pq->len && pq_less(&pq->heap[k], &pq->heap[min])) {
60 + min = k;
61 + }
62 +
63 + if (min == i) {
64 + break;
65 + }
66 +
67 + SWAP(pq->heap[i], pq->heap[min]);
68 + i = min;
69 + }
70 +
71 + return e;
72 +}
73 +
74 +void merged_iter_pqueue_add(struct merged_iter_pqueue *pq, struct pq_entry e)
75 +{
76 + int i = 0;
77 + if (pq->len == pq->cap) {
78 + pq->cap = 2 * pq->cap + 1;
79 + pq->heap = reftable_realloc(pq->heap,
80 + pq->cap * sizeof(struct pq_entry));
81 + }
82 +
83 + pq->heap[pq->len++] = e;
84 + i = pq->len - 1;
85 + while (i > 0) {
86 + int j = (i - 1) / 2;
87 + if (pq_less(&pq->heap[j], &pq->heap[i])) {
88 + break;
89 + }
90 +
91 + SWAP(pq->heap[j], pq->heap[i]);
92 +
93 + i = j;
94 + }
95 +}
96 +
97 +void merged_iter_pqueue_release(struct merged_iter_pqueue *pq)
98 +{
99 + int i = 0;
100 + for (i = 0; i < pq->len; i++) {
101 + reftable_record_destroy(&pq->heap[i].rec);
102 + }
103 + FREE_AND_NULL(pq->heap);
104 + pq->len = pq->cap = 0;
105 +}
reftable/pq.h new
+33
@@ -0,0 +1,33 @@
1 +/*
2 +Copyright 2020 Google LLC
3 +
4 +Use of this source code is governed by a BSD-style
5 +license that can be found in the LICENSE file or at
6 +https://developers.google.com/open-source/licenses/bsd
7 +*/
8 +
9 +#ifndef PQ_H
10 +#define PQ_H
11 +
12 +#include "record.h"
13 +
14 +struct pq_entry {
15 + int index;
16 + struct reftable_record rec;
17 +};
18 +
19 +struct merged_iter_pqueue {
20 + struct pq_entry *heap;
21 + size_t len;
22 + size_t cap;
23 +};
24 +
25 +struct pq_entry merged_iter_pqueue_top(struct merged_iter_pqueue pq);
26 +int merged_iter_pqueue_is_empty(struct merged_iter_pqueue pq);
27 +void merged_iter_pqueue_check(struct merged_iter_pqueue pq);
28 +struct pq_entry merged_iter_pqueue_remove(struct merged_iter_pqueue *pq);
29 +void merged_iter_pqueue_add(struct merged_iter_pqueue *pq, struct pq_entry e);
30 +void merged_iter_pqueue_release(struct merged_iter_pqueue *pq);
31 +int pq_less(struct pq_entry *a, struct pq_entry *b);
32 +
33 +#endif
reftable/pq_test.c new
+82
@@ -0,0 +1,82 @@
1 +/*
2 +Copyright 2020 Google LLC
3 +
4 +Use of this source code is governed by a BSD-style
5 +license that can be found in the LICENSE file or at
6 +https://developers.google.com/open-source/licenses/bsd
7 +*/
8 +
9 +#include "system.h"
10 +
11 +#include "basics.h"
12 +#include "constants.h"
13 +#include "pq.h"
14 +#include "record.h"
15 +#include "reftable-tests.h"
16 +#include "test_framework.h"
17 +
18 +void merged_iter_pqueue_check(struct merged_iter_pqueue pq)
19 +{
20 + int i;
21 + for (i = 1; i < pq.len; i++) {
22 + int parent = (i - 1) / 2;
23 +
24 + EXPECT(pq_less(&pq.heap[parent], &pq.heap[i]));
25 + }
26 +}
27 +
28 +static void test_pq(void)
29 +{
30 + char *names[54] = { NULL };
31 + int N = ARRAY_SIZE(names) - 1;
32 +
33 + struct merged_iter_pqueue pq = { NULL };
34 + const char *last = NULL;
35 +
36 + int i = 0;
37 + for (i = 0; i < N; i++) {
38 + char name[100];
39 + snprintf(name, sizeof(name), "%02d", i);
40 + names[i] = xstrdup(name);
41 + }
42 +
43 + i = 1;
44 + do {
45 + struct reftable_record rec =
46 + reftable_new_record(BLOCK_TYPE_REF);
47 + struct pq_entry e = { 0 };
48 +
49 + reftable_record_as_ref(&rec)->refname = names[i];
50 + e.rec = rec;
51 + merged_iter_pqueue_add(&pq, e);
52 + merged_iter_pqueue_check(pq);
53 + i = (i * 7) % N;
54 + } while (i != 1);
55 +
56 + while (!merged_iter_pqueue_is_empty(pq)) {
57 + struct pq_entry e = merged_iter_pqueue_remove(&pq);
58 + struct reftable_ref_record *ref =
59 + reftable_record_as_ref(&e.rec);
60 +
61 + merged_iter_pqueue_check(pq);
62 +
63 + if (last) {
64 + EXPECT(strcmp(last, ref->refname) < 0);
65 + }
66 + last = ref->refname;
67 + ref->refname = NULL;
68 + reftable_free(ref);
69 + }
70 +
71 + for (i = 0; i < N; i++) {
72 + reftable_free(names[i]);
73 + }
74 +
75 + merged_iter_pqueue_release(&pq);
76 +}
77 +
78 +int pq_test_main(int argc, const char *argv[])
79 +{
80 + RUN_TEST(test_pq);
81 + return 0;
82 +}
reftable/reftable-tests.h
+1
@@ -12,6 +12,7 @@ https://developers.google.com/open-source/licenses/bsd
12 int basics_test_main(int argc, const char **argv);
13 int block_test_main(int argc, const char **argv);
14 int merged_test_main(int argc, const char **argv);
15 +int pq_test_main(int argc, const char **argv);
16 int record_test_main(int argc, const char **argv);
17 int refname_test_main(int argc, const char **argv);
18 int readwrite_test_main(int argc, const char **argv);
t/helper/test-reftable.c
+1
@@ -5,6 +5,7 @@ int cmd__reftable(int argc, const char **argv)
5 {
6 basics_test_main(argc, argv);
7 block_test_main(argc, argv);
8 + pq_test_main(argc, argv);
9 record_test_main(argc, argv);
10 readwrite_test_main(argc, argv);
11 tree_test_main(argc, argv);