annotate test/org/tmatesoft/hg/test/MapTagsToFileRevisions.java @ 307:2f2ab5c27f41

Collect sort reverse indexes along with array sorting
author Artem Tikhomirov <tikhomirov.artem@gmail.com>
date Sat, 24 Sep 2011 04:06:27 +0200
parents 85b8efde5586
children b9592e21176a
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;
254
a620f0663a37 Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff changeset
10
285
6dbbc53fc46d Use Path instead of plain String for manifest file names
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 268
diff changeset
11 import org.tmatesoft.hg.core.HgBadStateException;
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;
307
2f2ab5c27f41 Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 304
diff changeset
17 import org.tmatesoft.hg.internal.ArrayHelper;
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
18 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
19 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
20 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
21 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
22 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
23 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
24 import org.tmatesoft.hg.repo.HgChangelog.RawChangeset;
285
6dbbc53fc46d Use Path instead of plain String for manifest file names
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 268
diff changeset
25 import org.tmatesoft.hg.repo.HgManifest.Flags;
254
a620f0663a37 Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff changeset
26 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
27 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
28 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
29
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 * @author Marc Strapetz
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 public class MapTagsToFileRevisions {
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 // Static =================================================================
a620f0663a37 Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff changeset
36
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
37 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
38 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
39 System.out.printf("Free mem: %,d\n", Runtime.getRuntime().freeMemory());
304
85b8efde5586 Use memory-friendly set implementation to canonicalize filenames and nodeids
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 285
diff changeset
40 // m.collectTagsPerFile();
307
2f2ab5c27f41 Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 304
diff changeset
41 // m.manifestWalk();
2f2ab5c27f41 Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 304
diff changeset
42 // m.changelogWalk();
2f2ab5c27f41 Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 304
diff changeset
43 m.revisionMap();
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
44 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
45 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
46 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
47 }
263
31f67be94e71 RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 259
diff changeset
48
307
2f2ab5c27f41 Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 304
diff changeset
49 /*
2f2ab5c27f41 Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 304
diff changeset
50 * Each 5000 revisions from cpython, total 15 revisions
2f2ab5c27f41 Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 304
diff changeset
51 * Direct clog.getLocalRevision: ~260 ms
2f2ab5c27f41 Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 304
diff changeset
52 * RevisionMap.localRevision: ~265 ms (almost 100% in #init())
2f2ab5c27f41 Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 304
diff changeset
53 * 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
54 * 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
55 * 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
56 */
2f2ab5c27f41 Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 304
diff changeset
57 private void revisionMap() throws Exception {
2f2ab5c27f41 Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 304
diff changeset
58 ArrayHelper ah = new ArrayHelper();
2f2ab5c27f41 Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 304
diff changeset
59 final List<String> initial = Arrays.asList("d", "w", "k", "b", "c", "i", "a", "r", "e", "h");
2f2ab5c27f41 Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 304
diff changeset
60 String[] a = (String[]) initial.toArray();
2f2ab5c27f41 Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 304
diff changeset
61 ah.sort(a);
2f2ab5c27f41 Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 304
diff changeset
62 System.out.println(Arrays.toString(initial.toArray()));
2f2ab5c27f41 Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 304
diff changeset
63 System.out.println(Arrays.toString(a));
2f2ab5c27f41 Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 304
diff changeset
64 System.out.println(Arrays.toString(ah.getReverse()));
2f2ab5c27f41 Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 304
diff changeset
65 Object[] rebuilt = new Object[a.length];
2f2ab5c27f41 Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 304
diff changeset
66 for (int i = 0; i < a.length; i++) {
2f2ab5c27f41 Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 304
diff changeset
67 int indexInOriginal = ah.getReverse()[i];
2f2ab5c27f41 Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 304
diff changeset
68 rebuilt[indexInOriginal-1] = a[i];
2f2ab5c27f41 Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 304
diff changeset
69 }
2f2ab5c27f41 Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 304
diff changeset
70 System.out.println(Arrays.toString(rebuilt));
2f2ab5c27f41 Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 304
diff changeset
71 //
2f2ab5c27f41 Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 304
diff changeset
72 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
73 final HgChangelog clog = repository.getChangelog();
2f2ab5c27f41 Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 304
diff changeset
74 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
75 final int step = 5000;
2f2ab5c27f41 Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 304
diff changeset
76 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
77 revisions.add(clog.getRevision(i));
2f2ab5c27f41 Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 304
diff changeset
78 }
2f2ab5c27f41 Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 304
diff changeset
79 final long s1 = System.nanoTime();
2f2ab5c27f41 Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 304
diff changeset
80 for (Nodeid n : revisions) {
2f2ab5c27f41 Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 304
diff changeset
81 int r = clog.getLocalRevision(n);
2f2ab5c27f41 Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 304
diff changeset
82 if (r % step != 0) {
2f2ab5c27f41 Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 304
diff changeset
83 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
84 }
2f2ab5c27f41 Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 304
diff changeset
85 }
2f2ab5c27f41 Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 304
diff changeset
86 System.out.printf("Direct lookup of %d revisions took %,d ns\n", revisions.size(), System.nanoTime() - s1);
2f2ab5c27f41 Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 304
diff changeset
87 HgChangelog.RevisionMap rmap = clog.new RevisionMap();
2f2ab5c27f41 Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 304
diff changeset
88 final long s2 = System.nanoTime();
2f2ab5c27f41 Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 304
diff changeset
89 rmap.init();
2f2ab5c27f41 Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 304
diff changeset
90 final long s3 = System.nanoTime();
2f2ab5c27f41 Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 304
diff changeset
91 for (Nodeid n : revisions) {
2f2ab5c27f41 Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 304
diff changeset
92 int r = rmap.localRevision(n);
2f2ab5c27f41 Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 304
diff changeset
93 if (r % step != 0) {
2f2ab5c27f41 Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 304
diff changeset
94 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
95 }
2f2ab5c27f41 Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 304
diff changeset
96 }
2f2ab5c27f41 Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 304
diff changeset
97 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
98 }
2f2ab5c27f41 Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 304
diff changeset
99
263
31f67be94e71 RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 259
diff changeset
100 private void changelogWalk() throws Exception {
307
2f2ab5c27f41 Collect sort reverse indexes along with array sorting
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 304
diff changeset
101 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
102 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
103 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
104 public int xx = 0;
31f67be94e71 RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 259
diff changeset
105
31f67be94e71 RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 259
diff changeset
106 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
107 if (xx+revisionNumber < 0) {
31f67be94e71 RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 259
diff changeset
108 System.out.println(xx);
31f67be94e71 RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 259
diff changeset
109 System.out.println(revisionNumber);
31f67be94e71 RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 259
diff changeset
110 }
31f67be94e71 RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 259
diff changeset
111 xx += revisionNumber;
31f67be94e71 RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 259
diff changeset
112 }
31f67be94e71 RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 259
diff changeset
113 });
31f67be94e71 RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 259
diff changeset
114 // 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
115 // 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
116 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
117 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
118 }
31f67be94e71 RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 259
diff changeset
119
31f67be94e71 RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 259
diff changeset
120 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
121 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
122 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
123 final HgRepository repository = new HgLookup().detect(new File("/temp/hg/cpython"));
285
6dbbc53fc46d Use Path instead of plain String for manifest file names
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 268
diff changeset
124 repository.getManifest().walk(0, 10000, new HgManifest.Inspector2() {
263
31f67be94e71 RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 259
diff changeset
125 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
126 return true;
31f67be94e71 RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 259
diff changeset
127 }
31f67be94e71 RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 259
diff changeset
128 public boolean next(Nodeid nid, String fname, String flags) {
285
6dbbc53fc46d Use Path instead of plain String for manifest file names
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 268
diff changeset
129 throw new HgBadStateException(HgManifest.Inspector2.class.getName());
6dbbc53fc46d Use Path instead of plain String for manifest file names
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 268
diff changeset
130 }
6dbbc53fc46d Use Path instead of plain String for manifest file names
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 268
diff changeset
131 public boolean next(Nodeid nid, Path fname, Flags flags) {
263
31f67be94e71 RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 259
diff changeset
132 return true;
31f67be94e71 RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 259
diff changeset
133 }
31f67be94e71 RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 259
diff changeset
134 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
135 return true;
31f67be94e71 RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 259
diff changeset
136 }
31f67be94e71 RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 259
diff changeset
137 });
31f67be94e71 RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 259
diff changeset
138 // 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
139 // 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
140 // 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
141 // 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
142 // 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
143 //
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
144 // 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
145 // 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
146 // 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
147 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
148 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
149 }
31f67be94e71 RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 259
diff changeset
150
31f67be94e71 RevlogStream - reduce number of object instances, reuse when possible
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 259
diff changeset
151 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
152 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
153 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
154 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
155 //
a620f0663a37 Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff changeset
156 // 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
157 //
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
158 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
159 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
160 // 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
161 // 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
162 final Map<Path, Nodeid[]> file2rev2tag = new HashMap<Path, Nodeid[]>();
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
163 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
164 // 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
165 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
166 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
167 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
168 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
169 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
170 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
171 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
172 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
173 }
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
174 }
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
175 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
176 // 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
177 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
178 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
179 tagLocalRevs = copy;
254
a620f0663a37 Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff changeset
180 }
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
181 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
182 //
285
6dbbc53fc46d Use Path instead of plain String for manifest file names
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 268
diff changeset
183 repository.getManifest().walk(new HgManifest.Inspector2() {
256
b61ed0f2c4da Yet another slight perf/mem improvement for the tags per file sample
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 255
diff changeset
184 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
185
a620f0663a37 Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff changeset
186 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
187 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
188 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
189 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
190 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
191 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
192 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
193 // 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
194 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
195 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
196 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
197 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
198 }
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
199 }
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
200 }
256
b61ed0f2c4da Yet another slight perf/mem improvement for the tags per file sample
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 255
diff changeset
201 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
202 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
203 }
a620f0663a37 Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff changeset
204 return true;
a620f0663a37 Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff changeset
205 }
a620f0663a37 Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff changeset
206
a620f0663a37 Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff changeset
207 public boolean next(Nodeid nid, String fname, String flags) {
285
6dbbc53fc46d Use Path instead of plain String for manifest file names
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 268
diff changeset
208 throw new HgBadStateException(HgManifest.Inspector2.class.getName());
6dbbc53fc46d Use Path instead of plain String for manifest file names
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 268
diff changeset
209 }
6dbbc53fc46d Use Path instead of plain String for manifest file names
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 268
diff changeset
210
6dbbc53fc46d Use Path instead of plain String for manifest file names
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 268
diff changeset
211 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
212 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
213 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
214 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
215 }
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
216 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
217 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
218 break;
b61ed0f2c4da Yet another slight perf/mem improvement for the tags per file sample
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents: 255
diff changeset
219 }
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
220 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
221 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
222 }
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
223 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
224 }
a620f0663a37 Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff changeset
225 return true;
a620f0663a37 Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff changeset
226 }
a620f0663a37 Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff changeset
227
a620f0663a37 Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff changeset
228 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
229 return true;
a620f0663a37 Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff changeset
230 }
a620f0663a37 Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff changeset
231
a620f0663a37 Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff changeset
232 }, 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
233 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
234 //
a620f0663a37 Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff changeset
235 // 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
236 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
237 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
238 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
239 // 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
240 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
241 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
242 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
243 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
244 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
245 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
246 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
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 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
250 }
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
251 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
252 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
253 }
a620f0663a37 Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff changeset
254
a620f0663a37 Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff changeset
255 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
256 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
257 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
258 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
259 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
260 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
261 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
262 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
263 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
264 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
265 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
266 // 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
267 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
268 if (fileRevision == null) {
a620f0663a37 Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff changeset
269 continue;
a620f0663a37 Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff changeset
270 }
a620f0663a37 Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff changeset
271
a620f0663a37 Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff changeset
272 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
273 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
274 if (revisionTags == null) {
a620f0663a37 Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff changeset
275 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
276 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
277 }
a620f0663a37 Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff changeset
278 revisionTags.add(tagName);
a620f0663a37 Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff changeset
279 }
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 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
282 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
283 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
284 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
285 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
286 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
287 }
a620f0663a37 Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff changeset
288 }
a620f0663a37 Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff changeset
289 });
a620f0663a37 Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff changeset
290 }
a620f0663a37 Collect tags for a file - improve performance of 'sparse' manifest reads
Artem Tikhomirov <tikhomirov.artem@gmail.com>
parents:
diff changeset
291 }