Mercurial > hg4j
annotate test/org/tmatesoft/hg/test/MapTagsToFileRevisions.java @ 268:c5980f287cc4
Use StringProxy when parsing manifest to minimize number of useless conversions and array instances
author | Artem Tikhomirov <tikhomirov.artem@gmail.com> |
---|---|
date | Tue, 23 Aug 2011 22:30:56 +0200 |
parents | 31f67be94e71 |
children | 6dbbc53fc46d |
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 |
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
|
3 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
|
4 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
|
5 import java.util.Arrays; |
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
|
6 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
|
7 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
|
8 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
|
9 import java.util.Map; |
263
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
10 import java.util.regex.Pattern; |
254
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
11 |
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
|
12 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
|
13 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
|
14 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
|
15 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
|
16 import org.tmatesoft.hg.core.Nodeid; |
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.repo.HgChangelog; |
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.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
|
19 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
|
20 import org.tmatesoft.hg.repo.HgManifest; |
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
|
21 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
|
22 import org.tmatesoft.hg.repo.HgTags; |
263
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
23 import org.tmatesoft.hg.repo.HgChangelog.RawChangeset; |
254
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
24 import org.tmatesoft.hg.repo.HgTags.TagInfo; |
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
|
25 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
|
26 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
|
27 |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
28 /** |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
29 * @author Marc Strapetz |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
30 */ |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
31 public class MapTagsToFileRevisions { |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
32 |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
33 // Static ================================================================= |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
34 |
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
|
35 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
|
36 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
|
37 System.out.printf("Free mem: %,d\n", Runtime.getRuntime().freeMemory()); |
263
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
38 // Pattern p = Pattern.compile("^doc/[^/]*?\\.[0-9]\\.(x|ht)ml"); |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
39 // System.out.println(p.matcher("doc/asd.2.xml").matches()); |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
40 // System.out.println(p.matcher("doc/zxc.6.html").matches()); |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
41 m.collectTagsPerFile(); |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
42 // m.manifestWalk(); |
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
|
43 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
|
44 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
|
45 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
|
46 } |
263
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
47 |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
48 private void changelogWalk() throws Exception { |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
49 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
|
50 final HgRepository repository = new HgLookup().detect(new File("/temp/hg/cpython")); |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
51 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
|
52 public int xx = 0; |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
53 |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
54 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
|
55 if (xx+revisionNumber < 0) { |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
56 System.out.println(xx); |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
57 System.out.println(revisionNumber); |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
58 } |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
59 xx += revisionNumber; |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
60 } |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
61 }); |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
62 // 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
|
63 // 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
|
64 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
|
65 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
|
66 } |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
67 |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
68 private 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
|
69 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
|
70 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
|
71 final HgRepository repository = new HgLookup().detect(new File("/temp/hg/cpython")); |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
72 repository.getManifest().walk(0, 10000, new HgManifest.Inspector() { |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
73 |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
74 public boolean begin(int mainfestRevision, Nodeid nid, int changelogRevision) { |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
75 return true; |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
76 } |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
77 |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
78 public boolean next(Nodeid nid, String fname, String flags) { |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
79 return true; |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
80 } |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
81 |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
82 public boolean end(int manifestRevision) { |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
83 return true; |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
84 } |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
85 }); |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
86 // 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
|
87 // 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
|
88 // 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
|
89 // 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
|
90 // 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
|
91 // |
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
|
92 // 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
|
93 // 0..10k is 34 seconds now |
263
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
94 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
|
95 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
|
96 } |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
97 |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
98 private 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
|
99 final long start = System.currentTimeMillis(); |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
100 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
|
101 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
|
102 // |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
103 // 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
|
104 // |
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
|
105 final TagInfo[] allTags = new TagInfo[tags.getTags().size()]; |
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
|
106 tags.getTags().values().toArray(allTags); |
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
|
107 // 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
|
108 // is index of corresponding TagInfo in allTags; |
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
|
109 final Map<String, Nodeid[]> file2rev2tag = new HashMap<String, Nodeid[]>(); |
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
|
110 System.out.printf("Collecting manifests for %d tags\n", allTags.length); |
254
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
111 // effective translation of changeset revisions to their local indexes |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
112 final HgChangelog.RevisionMap clogrmap = repository.getChangelog().new RevisionMap().init(); |
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
|
113 int[] tagLocalRevs = new int[allTags.length]; |
259
ea0c0de86d0e
Avoid IAE when repository has removed tags. Do not duplicate pools that are already in ManifestParser
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
256
diff
changeset
|
114 int x = 0; |
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
|
115 for (int i = 0; i < allTags.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
|
116 final Nodeid tagRevision = allTags[i].revision(); |
259
ea0c0de86d0e
Avoid IAE when repository has removed tags. Do not duplicate pools that are already in ManifestParser
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
256
diff
changeset
|
117 final int tagLocalRev = clogrmap.localRevision(tagRevision); |
ea0c0de86d0e
Avoid IAE when repository has removed tags. Do not duplicate pools that are already in ManifestParser
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
256
diff
changeset
|
118 if (tagLocalRev != HgRepository.BAD_REVISION) { |
ea0c0de86d0e
Avoid IAE when repository has removed tags. Do not duplicate pools that are already in ManifestParser
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
256
diff
changeset
|
119 tagLocalRevs[x++] = tagLocalRev; |
ea0c0de86d0e
Avoid IAE when repository has removed tags. Do not duplicate pools that are already in ManifestParser
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
256
diff
changeset
|
120 } |
ea0c0de86d0e
Avoid IAE when repository has removed tags. Do not duplicate pools that are already in ManifestParser
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
256
diff
changeset
|
121 } |
ea0c0de86d0e
Avoid IAE when repository has removed tags. Do not duplicate pools that are already in ManifestParser
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
256
diff
changeset
|
122 if (x != allTags.length) { |
ea0c0de86d0e
Avoid IAE when repository has removed tags. Do not duplicate pools that are already in ManifestParser
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
256
diff
changeset
|
123 // some tags were removed (recorded Nodeid.NULL tagname) |
ea0c0de86d0e
Avoid IAE when repository has removed tags. Do not duplicate pools that are already in ManifestParser
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
256
diff
changeset
|
124 int[] copy = new int[x]; |
ea0c0de86d0e
Avoid IAE when repository has removed tags. Do not duplicate pools that are already in ManifestParser
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
256
diff
changeset
|
125 System.arraycopy(tagLocalRevs, 0, copy, 0, x); |
ea0c0de86d0e
Avoid IAE when repository has removed tags. Do not duplicate pools that are already in ManifestParser
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
256
diff
changeset
|
126 tagLocalRevs = copy; |
254
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
127 } |
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
|
128 System.out.printf("Prepared tag revisions to analyze: %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
|
129 // |
263
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
130 final int[] counts = new int[2]; |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
131 HgManifest.Inspector emptyInsp = new HgManifest.Inspector() { |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
132 public boolean begin(int mainfestRevision, Nodeid nid, int changelogRevision) { |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
133 counts[0]++; |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
134 return true; |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
135 } |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
136 public boolean next(Nodeid nid, String fname, String flags) { |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
137 counts[1]++; |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
138 return true; |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
139 } |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
140 public boolean end(int manifestRevision) { |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
141 return true; |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
142 } |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
143 }; |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
144 // final long start0 = System.currentTimeMillis(); |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
145 // repository.getManifest().walk(emptyInsp, tagLocalRevs[0]); // warm-up |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
146 // final long start1 = System.currentTimeMillis(); |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
147 // repository.getManifest().walk(emptyInsp, tagLocalRevs[0]); // warm-up |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
148 // counts[0] = counts[1] = 0; |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
149 // final long start2 = System.currentTimeMillis(); |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
150 // repository.getManifest().walk(emptyInsp, tagLocalRevs); |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
151 // // No-op iterate over selected revisions: 11719 ms (revs:189, files: 579652), warm-up: 218 ms. Cache built: 25281 ms |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
152 // // No-op iterate over selected revisions: 11719 ms (revs:189, files: 579652), warm-up1: 234 ms, warm-up2: 16 ms. Cache built: 25375 ms |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
153 // System.out.printf("No-op iterate over selected revisions: %d ms (revs:%d, files: %d), warm-up1: %d ms, warm-up2: %d ms \n", System.currentTimeMillis() - start2, counts[0], counts[1], start1-start0, start2-start1); |
31f67be94e71
RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
259
diff
changeset
|
154 // |
254
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
155 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
|
156 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
|
157 |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
158 public boolean begin(int mainfestRevision, Nodeid nid, int changelogRevision) { |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
159 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
|
160 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
|
161 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
|
162 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
|
163 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
|
164 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
|
165 // 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
|
166 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
|
167 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
|
168 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
|
169 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
|
170 } |
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
|
171 } |
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
|
172 } |
256
b61ed0f2c4da
Yet another slight perf/mem improvement for the tags per file sample
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
255
diff
changeset
|
173 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
|
174 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
|
175 } |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
176 return true; |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
177 } |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
178 |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
179 public boolean next(Nodeid nid, String fname, String 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
|
180 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
|
181 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
|
182 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
|
183 } |
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
|
184 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
|
185 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
|
186 break; |
b61ed0f2c4da
Yet another slight perf/mem improvement for the tags per file sample
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
255
diff
changeset
|
187 } |
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
|
188 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
|
189 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
|
190 } |
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
|
191 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
|
192 } |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
193 return true; |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
194 } |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
195 |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
196 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
|
197 return true; |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
198 } |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
199 |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
200 }, 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
|
201 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
|
202 // |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
203 // 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
|
204 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
|
205 HgDataFile fileNode = repository.getFileNode(targetPath); |
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
|
206 final Nodeid[] allTagsOfTheFile = file2rev2tag.get(targetPath.toString()); |
254
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
207 // TODO if fileNode.isCopy, repeat for each getCopySourceName() |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
208 for (int localFileRev = 0; localFileRev < fileNode.getRevisionCount(); localFileRev++) { |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
209 Nodeid fileRev = fileNode.getRevision(localFileRev); |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
210 int changesetLocalRev = fileNode.getChangesetLocalRevision(localFileRev); |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
211 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
|
212 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
|
213 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
|
214 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
|
215 } |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
216 } |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
217 System.out.printf("%3d%7d%s\n", localFileRev, changesetLocalRev, associatedTags); |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
218 } |
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
|
219 System.out.printf("Total time: %d ms\n", System.currentTimeMillis() - start); |
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
|
220 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
|
221 } |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
222 |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
223 public static void main2(String[] args) throws HgException, CancelledException { |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
224 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
|
225 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
|
226 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
|
227 final Map<String, HgTags.TagInfo> tagToInfo = tags.getTags(); |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
228 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
|
229 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
|
230 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
|
231 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
|
232 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
|
233 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
|
234 // 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.) |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
235 final Nodeid fileRevision = manifest.getFileRevision(repository.getChangelog().getLocalRevision(nodeId), targetPath); |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
236 if (fileRevision == null) { |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
237 continue; |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
238 } |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
239 |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
240 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
|
241 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
|
242 if (revisionTags == null) { |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
243 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
|
244 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
|
245 } |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
246 revisionTags.add(tagName); |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
247 } |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
248 |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
249 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
|
250 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
|
251 logCommand.execute(new HgChangesetHandler() { |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
252 public void next(HgChangeset changeset) { |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
253 if (changeset.getAffectedFiles().contains(targetPath)) { |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
254 System.out.println(changeset.getRevision() + " " + changeSetRevisionToTags.get(changeset.getNodeid())); |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
255 } |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
256 } |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
257 }); |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
258 } |
a620f0663a37
Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff
changeset
|
259 } |