Mercurial > hg4j
annotate test/org/tmatesoft/hg/test/MapTagsToFileRevisions.java @ 438:0d128e09d70f
Yet another alternative to collect tags for a file
author | Artem Tikhomirov <tikhomirov.artem@gmail.com> |
---|---|
date | Thu, 19 Apr 2012 19:16:45 +0200 |
parents | 9265f7f4b49b |
children | 6526d8adbc0f |
rev | line source |
---|---|
254
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
1 package org.tmatesoft.hg.test; |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
2 |
324
283b294d1079
Explore alternatives to access file-changelog combined history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
311
diff
changeset
|
3 import static org.tmatesoft.hg.repo.HgRepository.TIP; |
283b294d1079
Explore alternatives to access file-changelog combined history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
311
diff
changeset
|
4 |
255
5a6ab50b4cbf
Improve memory footprint of tag collection (about 14 Mb saved for cpython repo without HashMap.Entry and Entry[])
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
254
diff
changeset
|
5 import java.io.File; |
5a6ab50b4cbf
Improve memory footprint of tag collection (about 14 Mb saved for cpython repo without HashMap.Entry and Entry[])
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
254
diff
changeset
|
6 import java.util.ArrayList; |
256
b61ed0f2c4da
Yet another slight perf/mem improvement for the tags per file sample
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
255
diff
changeset
|
7 import java.util.Arrays; |
438
0d128e09d70f
Yet another alternative to collect tags for a file
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
436
diff
changeset
|
8 import java.util.Collections; |
255
5a6ab50b4cbf
Improve memory footprint of tag collection (about 14 Mb saved for cpython repo without HashMap.Entry and Entry[])
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
254
diff
changeset
|
9 import java.util.HashMap; |
5a6ab50b4cbf
Improve memory footprint of tag collection (about 14 Mb saved for cpython repo without HashMap.Entry and Entry[])
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
254
diff
changeset
|
10 import java.util.LinkedList; |
5a6ab50b4cbf
Improve memory footprint of tag collection (about 14 Mb saved for cpython repo without HashMap.Entry and Entry[])
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
254
diff
changeset
|
11 import java.util.List; |
5a6ab50b4cbf
Improve memory footprint of tag collection (about 14 Mb saved for cpython repo without HashMap.Entry and Entry[])
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
254
diff
changeset
|
12 import java.util.Map; |
254
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
13 |
324
283b294d1079
Explore alternatives to access file-changelog combined history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
311
diff
changeset
|
14 import org.junit.Assert; |
423
9c9c442b5f2e
Major refactoring of exception handling. Low-level API uses RuntimeExceptions, while checked are left for higher level
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
368
diff
changeset
|
15 import org.tmatesoft.hg.core.HgCallbackTargetException; |
255
5a6ab50b4cbf
Improve memory footprint of tag collection (about 14 Mb saved for cpython repo without HashMap.Entry and Entry[])
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
254
diff
changeset
|
16 import org.tmatesoft.hg.core.HgChangeset; |
5a6ab50b4cbf
Improve memory footprint of tag collection (about 14 Mb saved for cpython repo without HashMap.Entry and Entry[])
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
254
diff
changeset
|
17 import org.tmatesoft.hg.core.HgChangesetHandler; |
5a6ab50b4cbf
Improve memory footprint of tag collection (about 14 Mb saved for cpython repo without HashMap.Entry and Entry[])
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
254
diff
changeset
|
18 import org.tmatesoft.hg.core.HgException; |
5a6ab50b4cbf
Improve memory footprint of tag collection (about 14 Mb saved for cpython repo without HashMap.Entry and Entry[])
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
254
diff
changeset
|
19 import org.tmatesoft.hg.core.HgLogCommand; |
5a6ab50b4cbf
Improve memory footprint of tag collection (about 14 Mb saved for cpython repo without HashMap.Entry and Entry[])
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
254
diff
changeset
|
20 import org.tmatesoft.hg.core.Nodeid; |
326
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
21 import org.tmatesoft.hg.internal.IntMap; |
255
5a6ab50b4cbf
Improve memory footprint of tag collection (about 14 Mb saved for cpython repo without HashMap.Entry and Entry[])
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
254
diff
changeset
|
22 import org.tmatesoft.hg.repo.HgChangelog; |
326
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
23 import org.tmatesoft.hg.repo.HgChangelog.RawChangeset; |
255
5a6ab50b4cbf
Improve memory footprint of tag collection (about 14 Mb saved for cpython repo without HashMap.Entry and Entry[])
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
254
diff
changeset
|
24 import org.tmatesoft.hg.repo.HgDataFile; |
5a6ab50b4cbf
Improve memory footprint of tag collection (about 14 Mb saved for cpython repo without HashMap.Entry and Entry[])
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
254
diff
changeset
|
25 import org.tmatesoft.hg.repo.HgLookup; |
5a6ab50b4cbf
Improve memory footprint of tag collection (about 14 Mb saved for cpython repo without HashMap.Entry and Entry[])
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
254
diff
changeset
|
26 import org.tmatesoft.hg.repo.HgManifest; |
326
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
27 import org.tmatesoft.hg.repo.HgManifest.Flags; |
255
5a6ab50b4cbf
Improve memory footprint of tag collection (about 14 Mb saved for cpython repo without HashMap.Entry and Entry[])
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
254
diff
changeset
|
28 import org.tmatesoft.hg.repo.HgRepository; |
5a6ab50b4cbf
Improve memory footprint of tag collection (about 14 Mb saved for cpython repo without HashMap.Entry and Entry[])
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
254
diff
changeset
|
29 import org.tmatesoft.hg.repo.HgTags; |
254
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
30 import org.tmatesoft.hg.repo.HgTags.TagInfo; |
433
be697c3e951e
Revlog.RevisionMap helper class got promoted as TLC, renamed to HgRevisionMap
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
431
diff
changeset
|
31 import org.tmatesoft.hg.repo.HgRevisionMap; |
255
5a6ab50b4cbf
Improve memory footprint of tag collection (about 14 Mb saved for cpython repo without HashMap.Entry and Entry[])
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
254
diff
changeset
|
32 import org.tmatesoft.hg.util.CancelledException; |
5a6ab50b4cbf
Improve memory footprint of tag collection (about 14 Mb saved for cpython repo without HashMap.Entry and Entry[])
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
254
diff
changeset
|
33 import org.tmatesoft.hg.util.Path; |
254
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
34 |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
35 /** |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
36 * @author Marc Strapetz |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
37 */ |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
38 public class MapTagsToFileRevisions { |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
39 |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
40 // Static ================================================================= |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
41 |
255
5a6ab50b4cbf
Improve memory footprint of tag collection (about 14 Mb saved for cpython repo without HashMap.Entry and Entry[])
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
254
diff
changeset
|
42 public static void main(String[] args) throws Exception { |
5a6ab50b4cbf
Improve memory footprint of tag collection (about 14 Mb saved for cpython repo without HashMap.Entry and Entry[])
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
254
diff
changeset
|
43 MapTagsToFileRevisions m = new MapTagsToFileRevisions(); |
5a6ab50b4cbf
Improve memory footprint of tag collection (about 14 Mb saved for cpython repo without HashMap.Entry and Entry[])
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
254
diff
changeset
|
44 System.out.printf("Free mem: %,d\n", Runtime.getRuntime().freeMemory()); |
436
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
45 // m.measurePatchAffectsArbitraryRevisionRead(); |
329
694ebabb5cb3
Refactor revlog patch mechanism, towards patch merging
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
326
diff
changeset
|
46 // m.collectTagsPerFile(); |
307
2f2ab5c27f41
Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
304
diff
changeset
|
47 // m.manifestWalk(); |
2f2ab5c27f41
Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
304
diff
changeset
|
48 // m.changelogWalk(); |
324
283b294d1079
Explore alternatives to access file-changelog combined history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
311
diff
changeset
|
49 // m.revisionMap(); |
436
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
50 m.buildFile2ChangelogRevisionMap(".hgtags", "README", "configure.in", "Misc/NEWS"); |
255
5a6ab50b4cbf
Improve memory footprint of tag collection (about 14 Mb saved for cpython repo without HashMap.Entry and Entry[])
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
254
diff
changeset
|
51 m = null; |
5a6ab50b4cbf
Improve memory footprint of tag collection (about 14 Mb saved for cpython repo without HashMap.Entry and Entry[])
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
254
diff
changeset
|
52 System.gc(); |
5a6ab50b4cbf
Improve memory footprint of tag collection (about 14 Mb saved for cpython repo without HashMap.Entry and Entry[])
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
254
diff
changeset
|
53 System.out.printf("Free mem: %,d\n", Runtime.getRuntime().freeMemory()); |
5a6ab50b4cbf
Improve memory footprint of tag collection (about 14 Mb saved for cpython repo without HashMap.Entry and Entry[])
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
254
diff
changeset
|
54 } |
324
283b294d1079
Explore alternatives to access file-changelog combined history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
311
diff
changeset
|
55 |
329
694ebabb5cb3
Refactor revlog patch mechanism, towards patch merging
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
326
diff
changeset
|
56 |
694ebabb5cb3
Refactor revlog patch mechanism, towards patch merging
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
326
diff
changeset
|
57 // revision == 2406 - 5 ms per run (baseRevision == 2406) |
694ebabb5cb3
Refactor revlog patch mechanism, towards patch merging
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
326
diff
changeset
|
58 // revision == 2405 - 69 ms per run (baseRevision == 1403) |
438
0d128e09d70f
Yet another alternative to collect tags for a file
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
436
diff
changeset
|
59 public void measurePatchAffectsArbitraryRevisionRead() throws Exception { |
329
694ebabb5cb3
Refactor revlog patch mechanism, towards patch merging
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
326
diff
changeset
|
60 final HgRepository repository = new HgLookup().detect(new File("/temp/hg/cpython")); |
694ebabb5cb3
Refactor revlog patch mechanism, towards patch merging
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
326
diff
changeset
|
61 final DoNothingManifestInspector insp = new DoNothingManifestInspector(); |
694ebabb5cb3
Refactor revlog patch mechanism, towards patch merging
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
326
diff
changeset
|
62 final int revision = 2405; |
694ebabb5cb3
Refactor revlog patch mechanism, towards patch merging
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
326
diff
changeset
|
63 // warm-up. |
694ebabb5cb3
Refactor revlog patch mechanism, towards patch merging
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
326
diff
changeset
|
64 repository.getManifest().walk(revision, revision, insp); |
694ebabb5cb3
Refactor revlog patch mechanism, towards patch merging
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
326
diff
changeset
|
65 final int runs = 10; |
694ebabb5cb3
Refactor revlog patch mechanism, towards patch merging
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
326
diff
changeset
|
66 final long start = System.nanoTime(); |
694ebabb5cb3
Refactor revlog patch mechanism, towards patch merging
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
326
diff
changeset
|
67 for (int i = 0; i < runs; i++) { |
694ebabb5cb3
Refactor revlog patch mechanism, towards patch merging
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
326
diff
changeset
|
68 repository.getManifest().walk(revision, revision, insp); |
694ebabb5cb3
Refactor revlog patch mechanism, towards patch merging
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
326
diff
changeset
|
69 } |
694ebabb5cb3
Refactor revlog patch mechanism, towards patch merging
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
326
diff
changeset
|
70 final long end = System.nanoTime(); |
694ebabb5cb3
Refactor revlog patch mechanism, towards patch merging
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
326
diff
changeset
|
71 System.out.printf("%d ms per run\n", (end - start)/ (runs*1000000)); |
694ebabb5cb3
Refactor revlog patch mechanism, towards patch merging
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
326
diff
changeset
|
72 } |
694ebabb5cb3
Refactor revlog patch mechanism, towards patch merging
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
326
diff
changeset
|
73 |
324
283b294d1079
Explore alternatives to access file-changelog combined history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
311
diff
changeset
|
74 /* |
283b294d1079
Explore alternatives to access file-changelog combined history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
311
diff
changeset
|
75 * .hgtags, 261 revisions |
283b294d1079
Explore alternatives to access file-changelog combined history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
311
diff
changeset
|
76 * Approach 1: total 83, init: 0, iteration: 82 |
283b294d1079
Explore alternatives to access file-changelog combined history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
311
diff
changeset
|
77 * Approach 2: total 225, init: 206, iteration: 19 |
283b294d1079
Explore alternatives to access file-changelog combined history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
311
diff
changeset
|
78 * README, 465 revisions |
283b294d1079
Explore alternatives to access file-changelog combined history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
311
diff
changeset
|
79 * Approach 1: total 162, init: 0, iteration: 161 |
283b294d1079
Explore alternatives to access file-changelog combined history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
311
diff
changeset
|
80 * Approach 2: total 231, init: 198, iteration: 32 |
283b294d1079
Explore alternatives to access file-changelog combined history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
311
diff
changeset
|
81 * configure.in, 1109 revisions |
283b294d1079
Explore alternatives to access file-changelog combined history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
311
diff
changeset
|
82 * Approach 1: total 409, init: 1, iteration: 407 |
283b294d1079
Explore alternatives to access file-changelog combined history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
311
diff
changeset
|
83 * Approach 2: total 277, init: 203, iteration: 74 |
283b294d1079
Explore alternatives to access file-changelog combined history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
311
diff
changeset
|
84 */ |
436
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
85 /* New data, 0.9.0v (another CPU!) |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
86 *.hgtags, 306 revisions |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
87 * Approach 0: total 136 |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
88 * Approach 1: total 53, init: 1, iteration: 52 |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
89 * Approach 2: total 95, init: 78, iteration: 17 |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
90 * Approach 3: total 17 |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
91 * |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
92 * README, 499 revisions |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
93 * Approach 0: total 149 |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
94 * Approach 1: total 43, init: 0, iteration: 43 |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
95 * Approach 2: total 102, init: 86, iteration: 16 |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
96 * Approach 3: total 18 |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
97 * |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
98 * configure.in, 1170 revisions |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
99 * Approach 0: total 321 |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
100 * Approach 1: total 116, init: 0, iteration: 115 |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
101 * Approach 2: total 140, init: 79, iteration: 60 |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
102 * Approach 3: total 30 |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
103 * |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
104 * Misc/NEWS, 10863 revisions |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
105 * Approach 0: total 4946 |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
106 * Approach 1: total 309, init: 6, iteration: 302 |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
107 * Approach 2: total 213, init: 63, iteration: 150 |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
108 * Approach 3: total 140 |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
109 */ |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
110 private void buildFile2ChangelogRevisionMap(String... fileNames) throws Exception { |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
111 final HgRepository repository = new HgLookup().detect(new File("/home/artem/hg/cpython")); |
324
283b294d1079
Explore alternatives to access file-changelog combined history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
311
diff
changeset
|
112 final HgChangelog clog = repository.getChangelog(); |
283b294d1079
Explore alternatives to access file-changelog combined history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
311
diff
changeset
|
113 // warm-up |
433
be697c3e951e
Revlog.RevisionMap helper class got promoted as TLC, renamed to HgRevisionMap
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
431
diff
changeset
|
114 HgRevisionMap<HgChangelog> clogMap = new HgRevisionMap<HgChangelog>(clog).init(); |
324
283b294d1079
Explore alternatives to access file-changelog combined history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
311
diff
changeset
|
115 |
436
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
116 for (String fname : fileNames) { |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
117 HgDataFile fileNode = repository.getFileNode(fname); |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
118 // warm-up |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
119 HgRevisionMap<HgDataFile> fileMap = new HgRevisionMap<HgDataFile>(fileNode).init(); |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
120 // |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
121 final int latestRevision = fileNode.getLastRevision(); |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
122 // |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
123 final long start_0 = System.nanoTime(); |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
124 final Map<Nodeid, Nodeid> changesetToNodeid_0 = new HashMap<Nodeid, Nodeid>(); |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
125 for (int fileRevisionIndex = 0; fileRevisionIndex <= latestRevision; fileRevisionIndex++) { |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
126 Nodeid fileRevision = fileNode.getRevision(fileRevisionIndex); |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
127 Nodeid changesetRevision = fileNode.getChangesetRevision(fileRevision); |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
128 changesetToNodeid_0.put(changesetRevision, fileRevision); |
324
283b294d1079
Explore alternatives to access file-changelog combined history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
311
diff
changeset
|
129 } |
436
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
130 final long end_0 = System.nanoTime(); |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
131 // |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
132 final long start_1 = System.nanoTime(); |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
133 fileMap = new HgRevisionMap<HgDataFile>(fileNode).init(); |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
134 final long start_1a = System.nanoTime(); |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
135 final Map<Nodeid, Nodeid> changesetToNodeid_1 = new HashMap<Nodeid, Nodeid>(); |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
136 for (int revision = 0; revision <= latestRevision; revision++) { |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
137 final Nodeid nodeId = fileMap.revision(revision); |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
138 int localCset = fileNode.getChangesetRevisionIndex(revision); |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
139 final Nodeid changesetId = clog.getRevision(localCset); |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
140 // final Nodeid changesetId = fileNode.getChangesetRevision(nodeId); |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
141 changesetToNodeid_1.put(changesetId, nodeId); |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
142 } |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
143 final long end_1 = System.nanoTime(); |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
144 // |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
145 final long start_2 = System.nanoTime(); |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
146 clogMap = new HgRevisionMap<HgChangelog>(clog).init(); |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
147 fileMap = new HgRevisionMap<HgDataFile>(fileNode).init(); |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
148 final Map<Nodeid, Nodeid> changesetToNodeid_2 = new HashMap<Nodeid, Nodeid>(); |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
149 final long start_2a = System.nanoTime(); |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
150 for (int revision = 0; revision <= latestRevision; revision++) { |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
151 Nodeid nidFile = fileMap.revision(revision); |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
152 int localCset = fileNode.getChangesetRevisionIndex(revision); |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
153 Nodeid nidCset = clogMap.revision(localCset); |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
154 changesetToNodeid_2.put(nidCset, nidFile); |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
155 } |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
156 final long end_2 = System.nanoTime(); |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
157 Assert.assertEquals(changesetToNodeid_1, changesetToNodeid_2); |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
158 // |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
159 final long start_3 = System.nanoTime(); |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
160 final Map<Nodeid, Nodeid> changesetToNodeid_3 = new HashMap<Nodeid, Nodeid>(); |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
161 fileNode.indexWalk(0, TIP, new HgDataFile.RevisionInspector() { |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
162 |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
163 public void next(int fileRevisionIndex, Nodeid revision, int linkedRevisionIndex) { |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
164 changesetToNodeid_3.put(clog.getRevision(linkedRevisionIndex), revision); |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
165 } |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
166 }); |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
167 final long end_3 = System.nanoTime(); |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
168 Assert.assertEquals(changesetToNodeid_1, changesetToNodeid_3); |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
169 System.out.printf("%s, %d revisions\n", fname, 1+latestRevision); |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
170 System.out.printf("Approach 0: total %d\n", (end_0 - start_0)/1000000); |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
171 System.out.printf("Approach 1: total %d, init: %d, iteration: %d\n", (end_1 - start_1)/1000000, (start_1a - start_1)/1000000, (end_1 - start_1a)/1000000); |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
172 System.out.printf("Approach 2: total %d, init: %d, iteration: %d\n", (end_2 - start_2)/1000000, (start_2a - start_2)/1000000, (end_2 - start_2a)/1000000); |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
173 System.out.printf("Approach 3: total %d\n", (end_3 - start_3)/1000000); |
9265f7f4b49b
Updated sample to map changeset and file revisions
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
433
diff
changeset
|
174 } |
324
283b294d1079
Explore alternatives to access file-changelog combined history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
311
diff
changeset
|
175 } |
263
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
176 |
307
2f2ab5c27f41
Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
304
diff
changeset
|
177 /* |
2f2ab5c27f41
Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
304
diff
changeset
|
178 * Each 5000 revisions from cpython, total 15 revisions |
367
2fadf8695f8a
Use 'revision index' instead of the vague 'local revision number' concept in the API
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
366
diff
changeset
|
179 * Direct clog.getRevisionIndex: ~260 ms |
2fadf8695f8a
Use 'revision index' instead of the vague 'local revision number' concept in the API
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
366
diff
changeset
|
180 * RevisionMap.revisionIndex: ~265 ms (almost 100% in #init()) |
307
2f2ab5c27f41
Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
304
diff
changeset
|
181 * each 1000'th revision, total 71 revision: 1 230 vs 270 |
2f2ab5c27f41
Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
304
diff
changeset
|
182 * each 2000'th revision, total 36 revision: 620 vs 270 |
2f2ab5c27f41
Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
304
diff
changeset
|
183 * each 3000'th revision, total 24 revision: 410 vs 275 |
2f2ab5c27f41
Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
304
diff
changeset
|
184 */ |
438
0d128e09d70f
Yet another alternative to collect tags for a file
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
436
diff
changeset
|
185 public void revisionMap() throws Exception { |
307
2f2ab5c27f41
Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
304
diff
changeset
|
186 final HgRepository repository = new HgLookup().detect(new File("/temp/hg/cpython")); |
2f2ab5c27f41
Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
304
diff
changeset
|
187 final HgChangelog clog = repository.getChangelog(); |
2f2ab5c27f41
Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
304
diff
changeset
|
188 ArrayList<Nodeid> revisions = new ArrayList<Nodeid>(); |
2f2ab5c27f41
Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
304
diff
changeset
|
189 final int step = 5000; |
2f2ab5c27f41
Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
304
diff
changeset
|
190 for (int i = 0, top = clog.getLastRevision(); i < top; i += step) { |
2f2ab5c27f41
Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
304
diff
changeset
|
191 revisions.add(clog.getRevision(i)); |
2f2ab5c27f41
Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
304
diff
changeset
|
192 } |
2f2ab5c27f41
Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
304
diff
changeset
|
193 final long s1 = System.nanoTime(); |
2f2ab5c27f41
Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
304
diff
changeset
|
194 for (Nodeid n : revisions) { |
367
2fadf8695f8a
Use 'revision index' instead of the vague 'local revision number' concept in the API
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
366
diff
changeset
|
195 int r = clog.getRevisionIndex(n); |
307
2f2ab5c27f41
Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
304
diff
changeset
|
196 if (r % step != 0) { |
2f2ab5c27f41
Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
304
diff
changeset
|
197 throw new IllegalStateException(Integer.toString(r)); |
2f2ab5c27f41
Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
304
diff
changeset
|
198 } |
2f2ab5c27f41
Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
304
diff
changeset
|
199 } |
2f2ab5c27f41
Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
304
diff
changeset
|
200 System.out.printf("Direct lookup of %d revisions took %,d ns\n", revisions.size(), System.nanoTime() - s1); |
433
be697c3e951e
Revlog.RevisionMap helper class got promoted as TLC, renamed to HgRevisionMap
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
431
diff
changeset
|
201 HgRevisionMap<HgChangelog> rmap = new HgRevisionMap<HgChangelog>(clog); |
307
2f2ab5c27f41
Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
304
diff
changeset
|
202 final long s2 = System.nanoTime(); |
2f2ab5c27f41
Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
304
diff
changeset
|
203 rmap.init(); |
2f2ab5c27f41
Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
304
diff
changeset
|
204 final long s3 = System.nanoTime(); |
2f2ab5c27f41
Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
304
diff
changeset
|
205 for (Nodeid n : revisions) { |
367
2fadf8695f8a
Use 'revision index' instead of the vague 'local revision number' concept in the API
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
366
diff
changeset
|
206 int r = rmap.revisionIndex(n); |
307
2f2ab5c27f41
Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
304
diff
changeset
|
207 if (r % step != 0) { |
2f2ab5c27f41
Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
304
diff
changeset
|
208 throw new IllegalStateException(Integer.toString(r)); |
2f2ab5c27f41
Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
304
diff
changeset
|
209 } |
2f2ab5c27f41
Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
304
diff
changeset
|
210 } |
2f2ab5c27f41
Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
304
diff
changeset
|
211 System.out.printf("RevisionMap time: %d ms, of that init() %,d ns\n", (System.nanoTime() - s2) / 1000000, s3 - s2); |
2f2ab5c27f41
Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
304
diff
changeset
|
212 } |
2f2ab5c27f41
Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
304
diff
changeset
|
213 |
438
0d128e09d70f
Yet another alternative to collect tags for a file
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
436
diff
changeset
|
214 public void changelogWalk() throws Exception { |
307
2f2ab5c27f41
Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
304
diff
changeset
|
215 final HgRepository repository = new HgLookup().detect(new File("/temp/hg/cpython")); |
263
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
216 final long start = System.currentTimeMillis(); |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
217 repository.getChangelog().all(new HgChangelog.Inspector() { |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
218 public int xx = 0; |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
219 |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
220 public void next(int revisionNumber, Nodeid nodeid, RawChangeset cset) { |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
221 if (xx+revisionNumber < 0) { |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
222 System.out.println(xx); |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
223 System.out.println(revisionNumber); |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
224 } |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
225 xx += revisionNumber; |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
226 } |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
227 }); |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
228 // cpython: 17 seconds, mem 132,9 -> 129,0 -> 131,7 |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
229 // cpyhton: 13 seconds. Of that, cumulative Patch.apply takes 8.8 seconds, RevlogStream.Inspector.next - 1.8 |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
230 System.out.printf("Total time: %d ms\n", System.currentTimeMillis() - start); |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
231 System.out.printf("Free mem: %,d\n", Runtime.getRuntime().freeMemory()); |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
232 } |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
233 |
438
0d128e09d70f
Yet another alternative to collect tags for a file
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
436
diff
changeset
|
234 public void manifestWalk() throws Exception { |
268
c5980f287cc4
Use StringProxy when parsing manifest to minimize number of useless conversions and array instances
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
263
diff
changeset
|
235 System.out.println(System.getProperty("java.version")); |
263
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
236 final long start = System.currentTimeMillis(); |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
237 final HgRepository repository = new HgLookup().detect(new File("/temp/hg/cpython")); |
329
694ebabb5cb3
Refactor revlog patch mechanism, towards patch merging
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
326
diff
changeset
|
238 repository.getManifest().walk(0, 10000, new DoNothingManifestInspector()); |
263
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
239 // cpython: 1,1 sec for 0..1000, 43 sec for 0..10000, 115 sec for 0..20000 (Pool with HashMap) |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
240 // 2,4 sec for 1000..2000 |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
241 // cpython -r 1000: 484 files, -r 2000: 1015 files. Iteration 1000..2000; fnamePool.size:1019 nodeidPool.size:2989 |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
242 // nodeidPool for two subsequent revisions only: 840. 37 sec for 0..10000. 99 sec for 0..20k |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
243 // 0..10000 fnamePool: hits:15989152, misses:3020 |
268
c5980f287cc4
Use StringProxy when parsing manifest to minimize number of useless conversions and array instances
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
263
diff
changeset
|
244 // |
c5980f287cc4
Use StringProxy when parsing manifest to minimize number of useless conversions and array instances
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
263
diff
changeset
|
245 // With Pool<StringProxy> for fname and flags, Nodeid's ascii2bin through local array, overall byte[] iteration, |
c5980f287cc4
Use StringProxy when parsing manifest to minimize number of useless conversions and array instances
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
263
diff
changeset
|
246 // 0..10k is 34 seconds now |
304
85b8efde5586
Use memory-friendly set implementation to canonicalize filenames and nodeids
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
285
diff
changeset
|
247 // Another run, 23 seconds now, seems nothing has been changed. Switched to Pool2 with DirectHashSet: 22,5 seconds |
263
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
248 System.out.printf("Total time: %d ms\n", System.currentTimeMillis() - start); |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
249 System.out.printf("Free mem: %,d\n", Runtime.getRuntime().freeMemory()); |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
250 } |
326
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
251 |
433
be697c3e951e
Revlog.RevisionMap helper class got promoted as TLC, renamed to HgRevisionMap
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
431
diff
changeset
|
252 private int[] collectLocalTagRevisions(HgRevisionMap<HgChangelog> clogrmap, TagInfo[] allTags, IntMap<List<TagInfo>> tagLocalRev2TagInfo) { |
326
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
253 int[] tagLocalRevs = new int[allTags.length]; |
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
254 int x = 0; |
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
255 for (int i = 0; i < allTags.length; i++) { |
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
256 final Nodeid tagRevision = allTags[i].revision(); |
368
8107b95f4280
Update Javadoc with 'revision index'
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
367
diff
changeset
|
257 final int tagRevisionIndex = clogrmap.revisionIndex(tagRevision); |
8107b95f4280
Update Javadoc with 'revision index'
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
367
diff
changeset
|
258 if (tagRevisionIndex != HgRepository.BAD_REVISION) { |
8107b95f4280
Update Javadoc with 'revision index'
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
367
diff
changeset
|
259 tagLocalRevs[x++] = tagRevisionIndex; |
8107b95f4280
Update Javadoc with 'revision index'
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
367
diff
changeset
|
260 List<TagInfo> tagsAssociatedWithRevision = tagLocalRev2TagInfo.get(tagRevisionIndex); |
326
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
261 if (tagsAssociatedWithRevision == null) { |
368
8107b95f4280
Update Javadoc with 'revision index'
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
367
diff
changeset
|
262 tagLocalRev2TagInfo.put(tagRevisionIndex, tagsAssociatedWithRevision = new LinkedList<TagInfo>()); |
326
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
263 } |
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
264 tagsAssociatedWithRevision.add(allTags[i]); |
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
265 } |
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
266 } |
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
267 if (x != allTags.length) { |
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
268 // some tags were removed (recorded Nodeid.NULL tagname) |
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
269 int[] copy = new int[x]; |
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
270 System.arraycopy(tagLocalRevs, 0, copy, 0, x); |
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
271 tagLocalRevs = copy; |
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
272 } |
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
273 return tagLocalRevs; |
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
274 } |
263
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
275 |
438
0d128e09d70f
Yet another alternative to collect tags for a file
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
436
diff
changeset
|
276 public void collectTagsPerFile() throws HgException, CancelledException { |
254
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
277 final long start = System.currentTimeMillis(); |
438
0d128e09d70f
Yet another alternative to collect tags for a file
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
436
diff
changeset
|
278 final HgRepository repository = new HgLookup().detect(new File("/home/artem/hg/cpython")); |
254
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
279 final HgTags tags = repository.getTags(); |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
280 // |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
281 // build cache |
255
5a6ab50b4cbf
Improve memory footprint of tag collection (about 14 Mb saved for cpython repo without HashMap.Entry and Entry[])
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
254
diff
changeset
|
282 // |
423
9c9c442b5f2e
Major refactoring of exception handling. Low-level API uses RuntimeExceptions, while checked are left for higher level
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
368
diff
changeset
|
283 final TagInfo[] allTags = new TagInfo[tags.getAllTags().size()]; |
9c9c442b5f2e
Major refactoring of exception handling. Low-level API uses RuntimeExceptions, while checked are left for higher level
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
368
diff
changeset
|
284 tags.getAllTags().values().toArray(allTags); |
326
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
285 // effective translation of changeset revisions to their local indexes |
433
be697c3e951e
Revlog.RevisionMap helper class got promoted as TLC, renamed to HgRevisionMap
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
431
diff
changeset
|
286 final HgRevisionMap<HgChangelog> clogrmap = new HgRevisionMap<HgChangelog>(repository.getChangelog()).init(); |
326
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
287 // map to look up tag by changeset local number |
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
288 final IntMap<List<TagInfo>> tagLocalRev2TagInfo = new IntMap<List<TagInfo>>(allTags.length); |
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
289 System.out.printf("Collecting manifests for %d tags\n", allTags.length); |
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
290 final int[] tagLocalRevs = collectLocalTagRevisions(clogrmap, allTags, tagLocalRev2TagInfo); |
329
694ebabb5cb3
Refactor revlog patch mechanism, towards patch merging
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
326
diff
changeset
|
291 System.out.printf("Prepared %d tag revisions to analyze: %d ms\n", tagLocalRevs.length, System.currentTimeMillis() - start); |
326
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
292 |
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
293 final Path targetPath = Path.create("README"); |
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
294 // |
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
295 collectTagsPerFile_Approach_1(clogrmap, tagLocalRevs, allTags, targetPath); |
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
296 System.out.printf("Total time: %d ms\n", System.currentTimeMillis() - start); |
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
297 |
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
298 System.out.println("\nApproach 2"); |
438
0d128e09d70f
Yet another alternative to collect tags for a file
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
436
diff
changeset
|
299 collectTagsPerFile_Approach_2(repository, tagLocalRevs, tagLocalRev2TagInfo, targetPath); |
326
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
300 } |
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
301 |
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
302 // Approach 1. Build map with all files, their revisions and corresponding tags |
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
303 // |
438
0d128e09d70f
Yet another alternative to collect tags for a file
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
436
diff
changeset
|
304 private void collectTagsPerFile_Approach_1(final HgRevisionMap<HgChangelog> clogrmap, final int[] tagLocalRevs, final TagInfo[] allTags, Path targetPath) throws HgException { |
326
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
305 HgRepository repository = clogrmap.getRepo(); |
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
306 final long start = System.currentTimeMillis(); |
255
5a6ab50b4cbf
Improve memory footprint of tag collection (about 14 Mb saved for cpython repo without HashMap.Entry and Entry[])
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
254
diff
changeset
|
307 // file2rev2tag value is array of revisions, always of allTags.length. Revision index in the array |
5a6ab50b4cbf
Improve memory footprint of tag collection (about 14 Mb saved for cpython repo without HashMap.Entry and Entry[])
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
254
diff
changeset
|
308 // is index of corresponding TagInfo in allTags; |
285
6dbbc53fc46d
Use Path instead of plain String for manifest file names
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
268
diff
changeset
|
309 final Map<Path, Nodeid[]> file2rev2tag = new HashMap<Path, Nodeid[]>(); |
424
6437d261048a
Deprecated code removed
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
423
diff
changeset
|
310 repository.getManifest().walk(new HgManifest.Inspector() { |
256
b61ed0f2c4da
Yet another slight perf/mem improvement for the tags per file sample
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
255
diff
changeset
|
311 private int[] tagIndexAtRev = new int[4]; // it's unlikely there would be a lot of tags associated with a given cset |
254
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
312 |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
313 public boolean begin(int mainfestRevision, Nodeid nid, int changelogRevision) { |
326
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
314 // may do better here using tagLocalRev2TagInfo, but need to change a lot, too lazy now |
254
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
315 Nodeid cset = clogrmap.revision(changelogRevision); |
256
b61ed0f2c4da
Yet another slight perf/mem improvement for the tags per file sample
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
255
diff
changeset
|
316 Arrays.fill(tagIndexAtRev, -1); |
b61ed0f2c4da
Yet another slight perf/mem improvement for the tags per file sample
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
255
diff
changeset
|
317 for (int i = 0, x = 0; i < allTags.length; i++) { |
255
5a6ab50b4cbf
Improve memory footprint of tag collection (about 14 Mb saved for cpython repo without HashMap.Entry and Entry[])
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
254
diff
changeset
|
318 if (cset.equals(allTags[i].revision())) { |
256
b61ed0f2c4da
Yet another slight perf/mem improvement for the tags per file sample
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
255
diff
changeset
|
319 tagIndexAtRev[x++] = i; |
b61ed0f2c4da
Yet another slight perf/mem improvement for the tags per file sample
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
255
diff
changeset
|
320 if (x == tagIndexAtRev.length) { |
b61ed0f2c4da
Yet another slight perf/mem improvement for the tags per file sample
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
255
diff
changeset
|
321 // expand twice as much |
b61ed0f2c4da
Yet another slight perf/mem improvement for the tags per file sample
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
255
diff
changeset
|
322 int[] expanded = new int[x << 1]; |
b61ed0f2c4da
Yet another slight perf/mem improvement for the tags per file sample
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
255
diff
changeset
|
323 System.arraycopy(tagIndexAtRev, 0, expanded, 0, x); |
b61ed0f2c4da
Yet another slight perf/mem improvement for the tags per file sample
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
255
diff
changeset
|
324 expanded[x] = -1; // just in case there'd be no more tags associated with this cset |
b61ed0f2c4da
Yet another slight perf/mem improvement for the tags per file sample
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
255
diff
changeset
|
325 tagIndexAtRev = expanded; |
b61ed0f2c4da
Yet another slight perf/mem improvement for the tags per file sample
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
255
diff
changeset
|
326 } |
255
5a6ab50b4cbf
Improve memory footprint of tag collection (about 14 Mb saved for cpython repo without HashMap.Entry and Entry[])
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
254
diff
changeset
|
327 } |
5a6ab50b4cbf
Improve memory footprint of tag collection (about 14 Mb saved for cpython repo without HashMap.Entry and Entry[])
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
254
diff
changeset
|
328 } |
256
b61ed0f2c4da
Yet another slight perf/mem improvement for the tags per file sample
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
255
diff
changeset
|
329 if (tagIndexAtRev[0] == -1) { |
254
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
330 System.out.println("Can't happen, provided we iterate over revisions with tags only"); |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
331 } |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
332 return true; |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
333 } |
285
6dbbc53fc46d
Use Path instead of plain String for manifest file names
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
268
diff
changeset
|
334 |
6dbbc53fc46d
Use Path instead of plain String for manifest file names
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
268
diff
changeset
|
335 public boolean next(Nodeid nid, Path fname, HgManifest.Flags flags) { |
255
5a6ab50b4cbf
Improve memory footprint of tag collection (about 14 Mb saved for cpython repo without HashMap.Entry and Entry[])
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
254
diff
changeset
|
336 Nodeid[] m = file2rev2tag.get(fname); |
254
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
337 if (m == null) { |
255
5a6ab50b4cbf
Improve memory footprint of tag collection (about 14 Mb saved for cpython repo without HashMap.Entry and Entry[])
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
254
diff
changeset
|
338 file2rev2tag.put(fname, m = new Nodeid[allTags.length]); |
254
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
339 } |
255
5a6ab50b4cbf
Improve memory footprint of tag collection (about 14 Mb saved for cpython repo without HashMap.Entry and Entry[])
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
254
diff
changeset
|
340 for (int tagIndex : tagIndexAtRev) { |
256
b61ed0f2c4da
Yet another slight perf/mem improvement for the tags per file sample
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
255
diff
changeset
|
341 if (tagIndex == -1) { |
b61ed0f2c4da
Yet another slight perf/mem improvement for the tags per file sample
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
255
diff
changeset
|
342 break; |
b61ed0f2c4da
Yet another slight perf/mem improvement for the tags per file sample
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
255
diff
changeset
|
343 } |
255
5a6ab50b4cbf
Improve memory footprint of tag collection (about 14 Mb saved for cpython repo without HashMap.Entry and Entry[])
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
254
diff
changeset
|
344 if (m[tagIndex] != null) { |
5a6ab50b4cbf
Improve memory footprint of tag collection (about 14 Mb saved for cpython repo without HashMap.Entry and Entry[])
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
254
diff
changeset
|
345 System.out.printf("There's another revision (%s) associated with tag %s already while we try to associate %s\n", m[tagIndex].shortNotation(), allTags[tagIndex].name(), nid.shortNotation()); |
5a6ab50b4cbf
Improve memory footprint of tag collection (about 14 Mb saved for cpython repo without HashMap.Entry and Entry[])
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
254
diff
changeset
|
346 } |
5a6ab50b4cbf
Improve memory footprint of tag collection (about 14 Mb saved for cpython repo without HashMap.Entry and Entry[])
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
254
diff
changeset
|
347 m[tagIndex] = nid; |
254
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
348 } |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
349 return true; |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
350 } |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
351 |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
352 public boolean end(int manifestRevision) { |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
353 return true; |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
354 } |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
355 |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
356 }, tagLocalRevs); |
255
5a6ab50b4cbf
Improve memory footprint of tag collection (about 14 Mb saved for cpython repo without HashMap.Entry and Entry[])
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
254
diff
changeset
|
357 System.out.printf("Cache built: %d ms\n", System.currentTimeMillis() - start); |
254
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
358 // |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
359 // look up specific file. This part is fast. |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
360 HgDataFile fileNode = repository.getFileNode(targetPath); |
285
6dbbc53fc46d
Use Path instead of plain String for manifest file names
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
268
diff
changeset
|
361 final Nodeid[] allTagsOfTheFile = file2rev2tag.get(targetPath); |
254
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
362 // TODO if fileNode.isCopy, repeat for each getCopySourceName() |
368
8107b95f4280
Update Javadoc with 'revision index'
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
367
diff
changeset
|
363 for (int fileRevIndex = 0; fileRevIndex < fileNode.getRevisionCount(); fileRevIndex++) { |
8107b95f4280
Update Javadoc with 'revision index'
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
367
diff
changeset
|
364 Nodeid fileRev = fileNode.getRevision(fileRevIndex); |
8107b95f4280
Update Javadoc with 'revision index'
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
367
diff
changeset
|
365 int changesetRevIndex = fileNode.getChangesetRevisionIndex(fileRevIndex); |
254
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
366 List<String> associatedTags = new LinkedList<String>(); |
255
5a6ab50b4cbf
Improve memory footprint of tag collection (about 14 Mb saved for cpython repo without HashMap.Entry and Entry[])
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
254
diff
changeset
|
367 for (int i = 0; i < allTagsOfTheFile.length; i++) { |
5a6ab50b4cbf
Improve memory footprint of tag collection (about 14 Mb saved for cpython repo without HashMap.Entry and Entry[])
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
254
diff
changeset
|
368 if (fileRev.equals(allTagsOfTheFile[i])) { |
5a6ab50b4cbf
Improve memory footprint of tag collection (about 14 Mb saved for cpython repo without HashMap.Entry and Entry[])
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
254
diff
changeset
|
369 associatedTags.add(allTags[i].name()); |
254
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
370 } |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
371 } |
368
8107b95f4280
Update Javadoc with 'revision index'
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
367
diff
changeset
|
372 System.out.printf("%3d%7d%s\n", fileRevIndex, changesetRevIndex, associatedTags); |
254
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
373 } |
326
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
374 } |
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
375 |
438
0d128e09d70f
Yet another alternative to collect tags for a file
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
436
diff
changeset
|
376 private void collectTagsPerFile_Approach_2(HgRepository repository, final int[] tagLocalRevs, final IntMap<List<TagInfo>> tagRevIndex2TagInfo, Path targetPath) throws HgException { |
326
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
377 // |
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
378 // Approach 2. No all-file map. Collect file revisions recorded at the time of tagging, |
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
379 // then for each file revision check if it is among those above, and if yes, take corresponding tags |
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
380 HgDataFile fileNode = repository.getFileNode(targetPath); |
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
381 final long start2 = System.nanoTime(); |
426
063b0663495a
HgManifest#getFileRevisions refactored into #walkFileRevisions to match pattern throught rest of the library
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
424
diff
changeset
|
382 final Map<Integer, Nodeid> fileRevisionAtTagRevision = new HashMap<Integer, Nodeid>(); |
438
0d128e09d70f
Yet another alternative to collect tags for a file
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
436
diff
changeset
|
383 final Map<Nodeid, List<String>> fileRev2TagNames = new HashMap<Nodeid, List<String>>(); |
426
063b0663495a
HgManifest#getFileRevisions refactored into #walkFileRevisions to match pattern throught rest of the library
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
424
diff
changeset
|
384 HgManifest.Inspector collectFileRevAtCset = new HgManifest.Inspector() { |
063b0663495a
HgManifest#getFileRevisions refactored into #walkFileRevisions to match pattern throught rest of the library
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
424
diff
changeset
|
385 |
063b0663495a
HgManifest#getFileRevisions refactored into #walkFileRevisions to match pattern throught rest of the library
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
424
diff
changeset
|
386 private int csetRevIndex; |
063b0663495a
HgManifest#getFileRevisions refactored into #walkFileRevisions to match pattern throught rest of the library
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
424
diff
changeset
|
387 |
063b0663495a
HgManifest#getFileRevisions refactored into #walkFileRevisions to match pattern throught rest of the library
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
424
diff
changeset
|
388 public boolean next(Nodeid nid, Path fname, Flags flags) { |
063b0663495a
HgManifest#getFileRevisions refactored into #walkFileRevisions to match pattern throught rest of the library
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
424
diff
changeset
|
389 fileRevisionAtTagRevision.put(csetRevIndex, nid); |
438
0d128e09d70f
Yet another alternative to collect tags for a file
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
436
diff
changeset
|
390 if (tagRevIndex2TagInfo.containsKey(csetRevIndex)) { |
0d128e09d70f
Yet another alternative to collect tags for a file
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
436
diff
changeset
|
391 List<String> tags = fileRev2TagNames.get(nid); |
0d128e09d70f
Yet another alternative to collect tags for a file
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
436
diff
changeset
|
392 if (tags == null) { |
0d128e09d70f
Yet another alternative to collect tags for a file
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
436
diff
changeset
|
393 fileRev2TagNames.put(nid, tags = new ArrayList<String>(3)); |
0d128e09d70f
Yet another alternative to collect tags for a file
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
436
diff
changeset
|
394 } |
0d128e09d70f
Yet another alternative to collect tags for a file
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
436
diff
changeset
|
395 for (TagInfo ti : tagRevIndex2TagInfo.get(csetRevIndex)) { |
0d128e09d70f
Yet another alternative to collect tags for a file
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
436
diff
changeset
|
396 tags.add(ti.name()); |
0d128e09d70f
Yet another alternative to collect tags for a file
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
436
diff
changeset
|
397 } |
0d128e09d70f
Yet another alternative to collect tags for a file
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
436
diff
changeset
|
398 } |
426
063b0663495a
HgManifest#getFileRevisions refactored into #walkFileRevisions to match pattern throught rest of the library
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
424
diff
changeset
|
399 return true; |
063b0663495a
HgManifest#getFileRevisions refactored into #walkFileRevisions to match pattern throught rest of the library
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
424
diff
changeset
|
400 } |
063b0663495a
HgManifest#getFileRevisions refactored into #walkFileRevisions to match pattern throught rest of the library
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
424
diff
changeset
|
401 |
063b0663495a
HgManifest#getFileRevisions refactored into #walkFileRevisions to match pattern throught rest of the library
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
424
diff
changeset
|
402 public boolean end(int manifestRevision) { |
063b0663495a
HgManifest#getFileRevisions refactored into #walkFileRevisions to match pattern throught rest of the library
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
424
diff
changeset
|
403 return true; |
063b0663495a
HgManifest#getFileRevisions refactored into #walkFileRevisions to match pattern throught rest of the library
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
424
diff
changeset
|
404 } |
063b0663495a
HgManifest#getFileRevisions refactored into #walkFileRevisions to match pattern throught rest of the library
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
424
diff
changeset
|
405 |
063b0663495a
HgManifest#getFileRevisions refactored into #walkFileRevisions to match pattern throught rest of the library
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
424
diff
changeset
|
406 public boolean begin(int mainfestRevision, Nodeid nid, int changelogRevision) { |
063b0663495a
HgManifest#getFileRevisions refactored into #walkFileRevisions to match pattern throught rest of the library
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
424
diff
changeset
|
407 csetRevIndex = changelogRevision; |
063b0663495a
HgManifest#getFileRevisions refactored into #walkFileRevisions to match pattern throught rest of the library
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
424
diff
changeset
|
408 return true; |
063b0663495a
HgManifest#getFileRevisions refactored into #walkFileRevisions to match pattern throught rest of the library
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
424
diff
changeset
|
409 } |
063b0663495a
HgManifest#getFileRevisions refactored into #walkFileRevisions to match pattern throught rest of the library
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
424
diff
changeset
|
410 }; |
063b0663495a
HgManifest#getFileRevisions refactored into #walkFileRevisions to match pattern throught rest of the library
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
424
diff
changeset
|
411 repository.getManifest().walkFileRevisions(targetPath, collectFileRevAtCset,tagLocalRevs); |
438
0d128e09d70f
Yet another alternative to collect tags for a file
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
436
diff
changeset
|
412 |
326
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
413 final long start2a = System.nanoTime(); |
438
0d128e09d70f
Yet another alternative to collect tags for a file
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
436
diff
changeset
|
414 fileNode.indexWalk(0, TIP, new HgDataFile.RevisionInspector() { |
326
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
415 |
368
8107b95f4280
Update Javadoc with 'revision index'
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
367
diff
changeset
|
416 public void next(int fileRevisionIndex, Nodeid fileRevision, int changesetRevisionIndex) { |
326
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
417 List<String> associatedTags = new LinkedList<String>(); |
438
0d128e09d70f
Yet another alternative to collect tags for a file
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
436
diff
changeset
|
418 |
0d128e09d70f
Yet another alternative to collect tags for a file
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
436
diff
changeset
|
419 for (int taggedRevision : tagLocalRevs) { |
326
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
420 // current file revision can't appear in tags that point to earlier changelog revisions (they got own file revision) |
438
0d128e09d70f
Yet another alternative to collect tags for a file
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
436
diff
changeset
|
421 if (taggedRevision >= changesetRevisionIndex) { |
326
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
422 // z points to some changeset with tag |
438
0d128e09d70f
Yet another alternative to collect tags for a file
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
436
diff
changeset
|
423 Nodeid wasKnownAs = fileRevisionAtTagRevision.get(taggedRevision); |
326
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
424 if (wasKnownAs.equals(fileRevision)) { |
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
425 // has tag associated with changeset at index z |
438
0d128e09d70f
Yet another alternative to collect tags for a file
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
436
diff
changeset
|
426 List<TagInfo> tagsAtRev = tagRevIndex2TagInfo.get(taggedRevision); |
326
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
427 assert tagsAtRev != null; |
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
428 for (TagInfo ti : tagsAtRev) { |
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
429 associatedTags.add(ti.name()); |
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
430 } |
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
431 } |
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
432 } |
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
433 } |
438
0d128e09d70f
Yet another alternative to collect tags for a file
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
436
diff
changeset
|
434 // |
368
8107b95f4280
Update Javadoc with 'revision index'
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
367
diff
changeset
|
435 System.out.printf("%3d%7d%s\n", fileRevisionIndex, changesetRevisionIndex, associatedTags); |
326
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
436 } |
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
437 }); |
438
0d128e09d70f
Yet another alternative to collect tags for a file
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
436
diff
changeset
|
438 for (int i = 0, lastRev = fileNode.getLastRevision(); i <= lastRev; i++) { |
0d128e09d70f
Yet another alternative to collect tags for a file
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
436
diff
changeset
|
439 Nodeid fileRevision = fileNode.getRevision(i); |
0d128e09d70f
Yet another alternative to collect tags for a file
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
436
diff
changeset
|
440 List<String> associatedTags2 = fileRev2TagNames.get(fileRevision); |
0d128e09d70f
Yet another alternative to collect tags for a file
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
436
diff
changeset
|
441 int changesetRevIndex = fileNode.getChangesetRevisionIndex(i); |
0d128e09d70f
Yet another alternative to collect tags for a file
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
436
diff
changeset
|
442 System.out.printf("%3d%7d%s\n", i, changesetRevIndex, associatedTags2 == null ? Collections.emptyList() : associatedTags2); |
0d128e09d70f
Yet another alternative to collect tags for a file
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
436
diff
changeset
|
443 } |
326
d42a45a2c9d6
Alternative tag collection approach for a file history
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
324
diff
changeset
|
444 System.out.printf("Alternative total time: %d ms, of that init: %d ms\n", (System.nanoTime() - start2)/1000000, (start2a-start2)/1000000); |
255
5a6ab50b4cbf
Improve memory footprint of tag collection (about 14 Mb saved for cpython repo without HashMap.Entry and Entry[])
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
254
diff
changeset
|
445 System.out.printf("Free mem: %,d\n", Runtime.getRuntime().freeMemory()); |
254
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
446 } |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
447 |
424
6437d261048a
Deprecated code removed
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
423
diff
changeset
|
448 static class DoNothingManifestInspector implements HgManifest.Inspector { |
329
694ebabb5cb3
Refactor revlog patch mechanism, towards patch merging
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
326
diff
changeset
|
449 public boolean begin(int mainfestRevision, Nodeid nid, int changelogRevision) { |
694ebabb5cb3
Refactor revlog patch mechanism, towards patch merging
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
326
diff
changeset
|
450 return true; |
694ebabb5cb3
Refactor revlog patch mechanism, towards patch merging
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
326
diff
changeset
|
451 } |
694ebabb5cb3
Refactor revlog patch mechanism, towards patch merging
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
326
diff
changeset
|
452 public boolean next(Nodeid nid, Path fname, Flags flags) { |
694ebabb5cb3
Refactor revlog patch mechanism, towards patch merging
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
326
diff
changeset
|
453 return true; |
694ebabb5cb3
Refactor revlog patch mechanism, towards patch merging
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
326
diff
changeset
|
454 } |
694ebabb5cb3
Refactor revlog patch mechanism, towards patch merging
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
326
diff
changeset
|
455 public boolean end(int manifestRevision) { |
694ebabb5cb3
Refactor revlog patch mechanism, towards patch merging
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
326
diff
changeset
|
456 return true; |
694ebabb5cb3
Refactor revlog patch mechanism, towards patch merging
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
326
diff
changeset
|
457 } |
694ebabb5cb3
Refactor revlog patch mechanism, towards patch merging
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
326
diff
changeset
|
458 } |
694ebabb5cb3
Refactor revlog patch mechanism, towards patch merging
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
326
diff
changeset
|
459 |
423
9c9c442b5f2e
Major refactoring of exception handling. Low-level API uses RuntimeExceptions, while checked are left for higher level
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
368
diff
changeset
|
460 public static void main2(String[] args) throws HgCallbackTargetException, HgException, CancelledException { |
254
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
461 final HgRepository repository = new HgLookup().detect(new File("/temp/hg/cpython")); |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
462 final Path targetPath = Path.create("README"); |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
463 final HgTags tags = repository.getTags(); |
423
9c9c442b5f2e
Major refactoring of exception handling. Low-level API uses RuntimeExceptions, while checked are left for higher level
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
368
diff
changeset
|
464 final Map<String, HgTags.TagInfo> tagToInfo = tags.getAllTags(); |
254
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
465 final HgManifest manifest = repository.getManifest(); |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
466 final Map<Nodeid, List<String>> changeSetRevisionToTags = new HashMap<Nodeid, List<String>>(); |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
467 final HgDataFile fileNode = repository.getFileNode(targetPath); |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
468 for (String tagName : tagToInfo.keySet()) { |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
469 final HgTags.TagInfo info = tagToInfo.get(tagName); |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
470 final Nodeid nodeId = info.revision(); |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
471 // TODO: This is not correct as we can't be sure that file at the corresponding revision is actually our target file (which may have been renamed, etc.) |
367
2fadf8695f8a
Use 'revision index' instead of the vague 'local revision number' concept in the API
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
366
diff
changeset
|
472 final Nodeid fileRevision = manifest.getFileRevision(repository.getChangelog().getRevisionIndex(nodeId), targetPath); |
254
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
473 if (fileRevision == null) { |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
474 continue; |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
475 } |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
476 |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
477 final Nodeid changeSetRevision = fileNode.getChangesetRevision(fileRevision); |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
478 List<String> revisionTags = changeSetRevisionToTags.get(changeSetRevision); |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
479 if (revisionTags == null) { |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
480 revisionTags = new ArrayList<String>(); |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
481 changeSetRevisionToTags.put(changeSetRevision, revisionTags); |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
482 } |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
483 revisionTags.add(tagName); |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
484 } |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
485 |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
486 final HgLogCommand logCommand = new HgLogCommand(repository); |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
487 logCommand.file(targetPath, true); |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
488 logCommand.execute(new HgChangesetHandler() { |
427
31a89587eb04
FIXMEs: consistent names, throws for commands and their handlers. Use of checked exceptions in hi-level api
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
426
diff
changeset
|
489 public void cset(HgChangeset changeset) { |
254
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
490 if (changeset.getAffectedFiles().contains(targetPath)) { |
423
9c9c442b5f2e
Major refactoring of exception handling. Low-level API uses RuntimeExceptions, while checked are left for higher level
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
368
diff
changeset
|
491 System.out.println(changeset.getRevisionIndex() + " " + changeSetRevisionToTags.get(changeset.getNodeid())); |
254
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
492 } |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
493 } |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
494 }); |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
495 } |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
496 } |