Mercurial > hg4j
view src/org/tmatesoft/hg/repo/Revlog.java @ 256:b61ed0f2c4da
Yet another slight perf/mem improvement for the tags per file sample
author | Artem Tikhomirov <tikhomirov.artem@gmail.com> |
---|---|
date | Tue, 16 Aug 2011 15:45:47 +0200 |
parents | 0e01f9182e16 |
children | 9fb50c04f03c |
line wrap: on
line source
/* * Copyright (c) 2010-2011 TMate Software Ltd * * This program is free software; you can redistribute it and/or modify * it under the terms of the GNU General Public License as published by * the Free Software Foundation; version 2 of the License. * * This program is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * * For information on how to redistribute this software under * the terms of a license other than GNU General Public License * contact TMate Software at support@hg4j.com */ package org.tmatesoft.hg.repo; import static org.tmatesoft.hg.core.Nodeid.NULL; import static org.tmatesoft.hg.repo.HgRepository.BAD_REVISION; import static org.tmatesoft.hg.repo.HgRepository.TIP; import java.io.IOException; import java.nio.ByteBuffer; import java.util.Arrays; import java.util.Collection; import java.util.HashSet; import java.util.LinkedList; import java.util.List; import org.tmatesoft.hg.core.HgBadStateException; import org.tmatesoft.hg.core.HgException; import org.tmatesoft.hg.core.Nodeid; import org.tmatesoft.hg.internal.DataAccess; import org.tmatesoft.hg.internal.RevlogStream; import org.tmatesoft.hg.util.ByteChannel; import org.tmatesoft.hg.util.CancelSupport; import org.tmatesoft.hg.util.CancelledException; import org.tmatesoft.hg.util.ProgressSupport; /** * Base class for all Mercurial entities that are serialized in a so called revlog format (changelog, manifest, data files). * * Implementation note: * Hides actual actual revlog stream implementation and its access methods (i.e. RevlogStream.Inspector), iow shall not expose anything internal * in public methods. * * @author Artem Tikhomirov * @author TMate Software Ltd. */ abstract class Revlog { private final HgRepository repo; protected final RevlogStream content; protected Revlog(HgRepository hgRepo, RevlogStream contentStream) { if (hgRepo == null) { throw new IllegalArgumentException(); } if (contentStream == null) { throw new IllegalArgumentException(); } repo = hgRepo; content = contentStream; } // invalid Revlog protected Revlog(HgRepository hgRepo) { repo = hgRepo; content = null; } public final HgRepository getRepo() { return repo; } public final int getRevisionCount() { return content.revisionCount(); } public final int getLastRevision() { return content.revisionCount() - 1; } public final Nodeid getRevision(int revision) { // XXX cache nodeids? return Nodeid.fromBinary(content.nodeid(revision), 0); } /** * Get local revision number (index) of the specified revision. * * For occasional queries, this method works with decent performance, despite its O(n/2) approach. * Alternatively, if you need to perform multiple queries (e.g. at least 15-20), {@link RevisionMap} may come handy. * * @param nid revision to look up * @return revision index, or {@link HgRepository#BAD_REVISION} if specified revision not found. */ public final int getLocalRevision(Nodeid nid) { int revision = content.findLocalRevisionNumber(nid); if (revision == BAD_REVISION) { throw new IllegalArgumentException(String.format("%s doesn't represent a revision of %s", nid.toString(), this /*XXX HgDataFile.getPath might be more suitable here*/)); } return revision; } // Till now, i follow approach that NULL nodeid is never part of revlog public final boolean isKnown(Nodeid nodeid) { final int rn = content.findLocalRevisionNumber(nodeid); if (BAD_REVISION == rn) { return false; } if (rn < 0 || rn >= content.revisionCount()) { // Sanity check throw new IllegalStateException(); } return true; } /** * Access to revision data as is (decompressed, but otherwise unprocessed, i.e. not parsed for e.g. changeset or manifest entries) * @param nodeid */ protected void rawContent(Nodeid nodeid, ByteChannel sink) throws HgException, IOException, CancelledException { rawContent(getLocalRevision(nodeid), sink); } /** * @param revision - repo-local index of this file change (not a changelog revision number!) */ protected void rawContent(int revision, ByteChannel sink) throws HgException, IOException, CancelledException { if (sink == null) { throw new IllegalArgumentException(); } ContentPipe insp = new ContentPipe(sink, 0); insp.checkCancelled(); content.iterate(revision, revision, true, insp); insp.checkFailed(); } /** * XXX perhaps, return value Nodeid[2] and boolean needNodeids is better (and higher level) API for this query? * * @param revision - revision to query parents, or {@link HgRepository#TIP} * @param parentRevisions - int[2] to get local revision numbers of parents (e.g. {6, -1}) * @param parent1 - byte[20] or null, if parent's nodeid is not needed * @param parent2 - byte[20] or null, if second parent's nodeid is not needed * @return */ public void parents(int revision, int[] parentRevisions, byte[] parent1, byte[] parent2) { if (revision != TIP && !(revision >= 0 && revision < content.revisionCount())) { throw new IllegalArgumentException(String.valueOf(revision)); } if (parentRevisions == null || parentRevisions.length < 2) { throw new IllegalArgumentException(String.valueOf(parentRevisions)); } if (parent1 != null && parent1.length < 20) { throw new IllegalArgumentException(parent1.toString()); } if (parent2 != null && parent2.length < 20) { throw new IllegalArgumentException(parent2.toString()); } class ParentCollector implements RevlogStream.Inspector { public int p1 = -1; public int p2 = -1; public byte[] nodeid; public void next(int revisionNumber, int actualLen, int baseRevision, int linkRevision, int parent1Revision, int parent2Revision, byte[] nodeid, DataAccess da) { p1 = parent1Revision; p2 = parent2Revision; this.nodeid = new byte[20]; // nodeid arg now comes in 32 byte from (as in file format description), however upper 12 bytes are zeros. System.arraycopy(nodeid, nodeid.length > 20 ? nodeid.length - 20 : 0, this.nodeid, 0, 20); } }; ParentCollector pc = new ParentCollector(); content.iterate(revision, revision, false, pc); parentRevisions[0] = pc.p1; parentRevisions[1] = pc.p2; if (parent1 != null) { if (parentRevisions[0] == -1) { Arrays.fill(parent1, 0, 20, (byte) 0); } else { content.iterate(parentRevisions[0], parentRevisions[0], false, pc); System.arraycopy(pc.nodeid, 0, parent1, 0, 20); } } if (parent2 != null) { if (parentRevisions[1] == -1) { Arrays.fill(parent2, 0, 20, (byte) 0); } else { content.iterate(parentRevisions[1], parentRevisions[1], false, pc); System.arraycopy(pc.nodeid, 0, parent2, 0, 20); } } } /* * XXX think over if it's better to do either: * pw = getChangelog().new ParentWalker(); pw.init() and pass pw instance around as needed * or * add Revlog#getParentWalker(), static class, make cons() and #init package-local, and keep SoftReference to allow walker reuse. * * and yes, walker is not a proper name */ public final class ParentWalker { private Nodeid[] sequential; // natural repository order, childrenOf rely on ordering private Nodeid[] sorted; // for binary search private int[] sorted2natural; private Nodeid[] firstParent; private Nodeid[] secondParent; // Nodeid instances shall be shared between all arrays public ParentWalker() { } public HgRepository getRepo() { return Revlog.this.getRepo(); } public void init() { final RevlogStream stream = Revlog.this.content; final int revisionCount = stream.revisionCount(); firstParent = new Nodeid[revisionCount]; // although branches/merges are less frequent, and most of secondParent would be -1/null, some sort of // SparseOrderedList might be handy, provided its inner structures do not overweight simplicity of an array secondParent = new Nodeid[revisionCount]; // sequential = new Nodeid[revisionCount]; sorted = new Nodeid[revisionCount]; RevlogStream.Inspector insp = new RevlogStream.Inspector() { int ix = 0; public void next(int revisionNumber, int actualLen, int baseRevision, int linkRevision, int parent1Revision, int parent2Revision, byte[] nodeid, DataAccess da) { if (ix != revisionNumber) { // XXX temp code, just to make sure I understand what's going on here // FIXME check against cpython or another tool-mangled repository throw new IllegalStateException(); } if (parent1Revision >= revisionNumber || parent2Revision >= revisionNumber) { throw new IllegalStateException(); // sanity, revisions are sequential } final Nodeid nid = new Nodeid(nodeid, true); sequential[ix] = sorted[ix] = nid; if (parent1Revision != -1) { assert parent1Revision < ix; firstParent[ix] = sequential[parent1Revision]; } if (parent2Revision != -1) { // revlog of DataAccess.java has p2 set when p1 is -1 assert parent2Revision < ix; secondParent[ix] = sequential[parent2Revision]; } ix++; } }; stream.iterate(0, TIP, false, insp); Arrays.sort(sorted); sorted2natural = new int[revisionCount]; for (int i = 0; i < revisionCount; i++) { Nodeid n = sequential[i]; int x = Arrays.binarySearch(sorted, n); assertSortedIndex(x); sorted2natural[x] = i; } } private void assertSortedIndex(int x) { if (x < 0) { throw new HgBadStateException(); } } // FIXME need to decide whether Nodeid(00 * 20) is always known or not // right now Nodeid.NULL is not recognized as known if passed to this method, // caller is supposed to make explicit check public boolean knownNode(Nodeid nid) { return Arrays.binarySearch(sorted, nid) >= 0; } /** * null if none. only known nodes (as per #knownNode) are accepted as arguments */ public Nodeid firstParent(Nodeid nid) { int x = Arrays.binarySearch(sorted, nid); assertSortedIndex(x); int i = sorted2natural[x]; return firstParent[i]; } // never null, Nodeid.NULL if none known public Nodeid safeFirstParent(Nodeid nid) { Nodeid rv = firstParent(nid); return rv == null ? Nodeid.NULL : rv; } public Nodeid secondParent(Nodeid nid) { int x = Arrays.binarySearch(sorted, nid); assertSortedIndex(x); int i = sorted2natural[x]; return secondParent[i]; } public Nodeid safeSecondParent(Nodeid nid) { Nodeid rv = secondParent(nid); return rv == null ? Nodeid.NULL : rv; } public boolean appendParentsOf(Nodeid nid, Collection<Nodeid> c) { int x = Arrays.binarySearch(sorted, nid); assertSortedIndex(x); int i = sorted2natural[x]; Nodeid p1 = firstParent[i]; boolean modified = false; if (p1 != null) { modified = c.add(p1); } Nodeid p2 = secondParent[i]; if (p2 != null) { modified = c.add(p2) || modified; } return modified; } // XXX alternative (and perhaps more reliable) approach would be to make a copy of allNodes and remove // nodes, their parents and so on. // @return ordered collection of all children rooted at supplied nodes. Nodes shall not be descendants of each other! // Nodeids shall belong to this revlog public List<Nodeid> childrenOf(List<Nodeid> roots) { HashSet<Nodeid> parents = new HashSet<Nodeid>(); LinkedList<Nodeid> result = new LinkedList<Nodeid>(); int earliestRevision = Integer.MAX_VALUE; assert sequential.length == firstParent.length && firstParent.length == secondParent.length; // first, find earliest index of roots in question, as there's no sense // to check children among nodes prior to branch's root node for (Nodeid r : roots) { int x = Arrays.binarySearch(sorted, r); assertSortedIndex(x); int i = sorted2natural[x]; if (i < earliestRevision) { earliestRevision = i; } parents.add(sequential[i]); // add canonical instance in hope equals() is bit faster when can do a == } for (int i = earliestRevision + 1; i < sequential.length; i++) { if (parents.contains(firstParent[i]) || parents.contains(secondParent[i])) { parents.add(sequential[i]); // to find next child result.add(sequential[i]); } } return result; } /** * @param nid possibly parent node, shall be {@link #knownNode(Nodeid) known} in this revlog. * @return <code>true</code> if there's any node in this revlog that has specified node as one of its parents. */ public boolean hasChildren(Nodeid nid) { int x = Arrays.binarySearch(sorted, nid); assertSortedIndex(x); int i = sorted2natural[x]; assert firstParent.length == secondParent.length; // just in case later I implement sparse array for secondParent assert firstParent.length == sequential.length; // to use == instead of equals, take the same Nodeid instance we used to fill all the arrays. final Nodeid canonicalNode = sequential[i]; i++; // no need to check node itself. child nodes may appear in sequential only after revision in question for (; i < sequential.length; i++) { // FIXME likely, not very effective. // May want to optimize it with another (Tree|Hash)Set, created on demand on first use, // however, need to be careful with memory usage if (firstParent[i] == canonicalNode || secondParent[i] == canonicalNode) { return true; } } return false; } } /** * Effective int to Nodeid and vice versa translation. It's advised to use this class instead of * multiple {@link Revlog#getLocalRevision(Nodeid)} calls. * * getLocalRevision(Nodeid) with straightforward lookup approach performs O(n/2) * #localRevision() is log(n), plus initialization is O(n) */ public final class RevisionMap { /* * in fact, initialization is much slower as it instantiates Nodeids, while #getLocalRevision * compares directly against byte buffer. Measuring cpython with 70k+ gives 3 times difference (47 vs 171) * for complete changelog iteration. */ /* * XXX 3 * (x * 4) bytes. Can I do better? */ private Nodeid[] sequential; // natural repository order, childrenOf rely on ordering private Nodeid[] sorted; // for binary search private int[] sorted2natural; public RevisionMap() { } public HgRepository getRepo() { return Revlog.this.getRepo(); } /** * @return <code>this</code> for convenience. */ public RevisionMap init(/*XXX Pool<Nodeid> to reuse nodeids, if possible. */) { // XXX HgRepository.register((RepoChangeListener) this); // listen to changes in repo, re-init if needed? final int revisionCount = Revlog.this.content.revisionCount(); sequential = new Nodeid[revisionCount]; sorted = new Nodeid[revisionCount]; sorted2natural = new int[revisionCount]; RevlogStream.Inspector insp = new RevlogStream.Inspector() { public void next(int revisionNumber, int actualLen, int baseRevision, int linkRevision, int parent1Revision, int parent2Revision, byte[] nodeid, DataAccess data) { final Nodeid nid = new Nodeid(nodeid, true); sequential[revisionNumber] = sorted[revisionNumber] = nid; } }; Revlog.this.content.iterate(0, TIP, false, insp); Arrays.sort(sorted); for (int i = 0; i < revisionCount; i++) { Nodeid n = sequential[i]; int x = Arrays.binarySearch(sorted, n); sorted2natural[x] = i; } return this; } public Nodeid revision(int localRevision) { return sequential[localRevision]; } public int localRevision(Nodeid revision) { if (revision == null || NULL.equals(revision)) { return BAD_REVISION; } int x = Arrays.binarySearch(sorted, revision); if (x < 0) { return BAD_REVISION; } return sorted2natural[x]; } } protected static class ContentPipe implements RevlogStream.Inspector, CancelSupport { private final ByteChannel sink; private final CancelSupport cancelSupport; private Exception failure; private final int offset; /** * @param _sink - cannot be <code>null</code> * @param seekOffset - when positive, orders to pipe bytes to the sink starting from specified offset, not from the first byte available in DataAccess */ public ContentPipe(ByteChannel _sink, int seekOffset) { assert _sink != null; sink = _sink; cancelSupport = CancelSupport.Factory.get(_sink); offset = seekOffset; } protected void prepare(int revisionNumber, DataAccess da) throws HgException, IOException { if (offset > 0) { // save few useless reset/rewind operations da.seek(offset); } } public void next(int revisionNumber, int actualLen, int baseRevision, int linkRevision, int parent1Revision, int parent2Revision, byte[] nodeid, DataAccess da) { try { prepare(revisionNumber, da); // XXX perhaps, prepare shall return DA (sliced, if needed) final ProgressSupport progressSupport = ProgressSupport.Factory.get(sink); ByteBuffer buf = ByteBuffer.allocate(512); progressSupport.start(da.length()); while (!da.isEmpty()) { cancelSupport.checkCancelled(); da.readBytes(buf); buf.flip(); // XXX I may not rely on returned number of bytes but track change in buf position instead. int consumed = sink.write(buf); // FIXME in fact, bad sink implementation (that consumes no bytes) would result in endless loop. Need to account for this buf.compact(); progressSupport.worked(consumed); } progressSupport.done(); // XXX shall specify whether #done() is invoked always or only if completed successfully. } catch (IOException ex) { recordFailure(ex); } catch (CancelledException ex) { recordFailure(ex); } catch (HgException ex) { recordFailure(ex); } } public void checkCancelled() throws CancelledException { cancelSupport.checkCancelled(); } protected void recordFailure(Exception ex) { assert failure == null; failure = ex; } public void checkFailed() throws HgException, IOException, CancelledException { if (failure == null) { return; } if (failure instanceof IOException) { throw (IOException) failure; } if (failure instanceof CancelledException) { throw (CancelledException) failure; } if (failure instanceof HgException) { throw (HgException) failure; } throw new HgBadStateException(failure); } } }