tikhomirov@13: /* tikhomirov@390: * Copyright (c) 2010-2012 TMate Software Ltd tikhomirov@74: * tikhomirov@74: * This program is free software; you can redistribute it and/or modify tikhomirov@74: * it under the terms of the GNU General Public License as published by tikhomirov@74: * the Free Software Foundation; version 2 of the License. tikhomirov@74: * tikhomirov@74: * This program is distributed in the hope that it will be useful, tikhomirov@74: * but WITHOUT ANY WARRANTY; without even the implied warranty of tikhomirov@74: * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the tikhomirov@74: * GNU General Public License for more details. tikhomirov@74: * tikhomirov@74: * For information on how to redistribute this software under tikhomirov@74: * the terms of a license other than GNU General Public License tikhomirov@102: * contact TMate Software at support@hg4j.com tikhomirov@2: */ tikhomirov@74: package org.tmatesoft.hg.repo; tikhomirov@74: tikhomirov@390: import static org.tmatesoft.hg.core.Nodeid.NULL; tikhomirov@390: import static org.tmatesoft.hg.repo.HgRepository.BAD_REVISION; tikhomirov@218: import static org.tmatesoft.hg.repo.HgRepository.TIP; tikhomirov@218: tikhomirov@232: import java.io.ByteArrayOutputStream; tikhomirov@157: import java.io.IOException; tikhomirov@218: import java.util.ArrayList; tikhomirov@218: import java.util.Arrays; tikhomirov@326: import java.util.HashMap; tikhomirov@326: import java.util.Map; tikhomirov@157: tikhomirov@390: import org.tmatesoft.hg.core.HgBadStateException; tikhomirov@366: import org.tmatesoft.hg.core.HgException; tikhomirov@354: import org.tmatesoft.hg.core.HgInvalidControlFileException; tikhomirov@403: import org.tmatesoft.hg.core.HgInvalidRevisionException; tikhomirov@74: import org.tmatesoft.hg.core.Nodeid; tikhomirov@157: import org.tmatesoft.hg.internal.DataAccess; tikhomirov@266: import org.tmatesoft.hg.internal.DigestHelper; tikhomirov@320: import org.tmatesoft.hg.internal.EncodingHelper; tikhomirov@232: import org.tmatesoft.hg.internal.Experimental; tikhomirov@312: import org.tmatesoft.hg.internal.IterateControlMediator; tikhomirov@218: import org.tmatesoft.hg.internal.Lifecycle; tikhomirov@304: import org.tmatesoft.hg.internal.Pool2; tikhomirov@77: import org.tmatesoft.hg.internal.RevlogStream; tikhomirov@312: import org.tmatesoft.hg.util.CancelSupport; tikhomirov@232: import org.tmatesoft.hg.util.Path; tikhomirov@312: import org.tmatesoft.hg.util.ProgressSupport; tikhomirov@74: tikhomirov@2: tikhomirov@2: /** tikhomirov@2: * tikhomirov@74: * @author Artem Tikhomirov tikhomirov@74: * @author TMate Software Ltd. tikhomirov@2: */ tikhomirov@2: public class HgManifest extends Revlog { tikhomirov@218: private RevisionMapper revisionMap; tikhomirov@412: private EncodingHelper encodingHelper; tikhomirov@285: tikhomirov@285: public enum Flags { tikhomirov@413: Exec, Link; // FIXME REVISIT consider REGULAR instead of null tikhomirov@285: tikhomirov@285: static Flags parse(String flags) { tikhomirov@285: if ("x".equalsIgnoreCase(flags)) { tikhomirov@285: return Exec; tikhomirov@285: } tikhomirov@285: if ("l".equalsIgnoreCase(flags)) { tikhomirov@285: return Link; tikhomirov@285: } tikhomirov@285: if (flags == null) { tikhomirov@285: return null; tikhomirov@285: } tikhomirov@285: throw new IllegalStateException(flags); tikhomirov@285: } tikhomirov@285: tikhomirov@285: static Flags parse(byte[] data, int start, int length) { tikhomirov@285: if (length == 0) { tikhomirov@285: return null; tikhomirov@285: } tikhomirov@285: if (length == 1) { tikhomirov@285: if (data[start] == 'x') { tikhomirov@285: return Exec; tikhomirov@285: } tikhomirov@285: if (data[start] == 'l') { tikhomirov@285: return Link; tikhomirov@285: } tikhomirov@285: // FALL THROUGH tikhomirov@285: } tikhomirov@285: throw new IllegalStateException(new String(data, start, length)); tikhomirov@285: } tikhomirov@285: tikhomirov@285: String nativeString() { tikhomirov@285: if (this == Exec) { tikhomirov@285: return "x"; tikhomirov@285: } tikhomirov@285: if (this == Link) { tikhomirov@285: return "l"; tikhomirov@285: } tikhomirov@285: throw new IllegalStateException(toString()); tikhomirov@285: } tikhomirov@285: } tikhomirov@2: tikhomirov@412: /*package-local*/ HgManifest(HgRepository hgRepo, RevlogStream content, EncodingHelper eh) { tikhomirov@21: super(hgRepo, content); tikhomirov@412: encodingHelper = eh; tikhomirov@13: } tikhomirov@13: tikhomirov@218: /** tikhomirov@390: * Walks manifest revisions that correspond to specified range of changesets. The order in which manifest versions get reported tikhomirov@390: * to the inspector corresponds to physical order of manifest revisions, not that of changesets (with few exceptions as noted below). tikhomirov@390: * That is, for cset-manifest revision pairs: tikhomirov@390: *
tikhomirov@390: 	 *   3  8
tikhomirov@390: 	 *   4  7
tikhomirov@390: 	 *   5  9
tikhomirov@390: 	 * 
tikhomirov@390: * call walk(3,5, insp) would yield (4,7), (3,8) and (5,9) to the inspector; tikhomirov@390: * different order of arguments, walk(5, 3, insp), makes no difference. tikhomirov@390: * tikhomirov@390: *

Physical layout of mercurial files (revlog) doesn't impose any restriction on whether manifest and changeset revisions shall go tikhomirov@390: * incrementally, nor it mandates presence of manifest version for a changeset. Thus, there might be changesets that record {@link Nodeid#NULL} tikhomirov@390: * as corresponding manifest revision. This situation is deemed exceptional now and what would inspector get depends on whether tikhomirov@390: * start or end arguments point to such changeset, or such changeset happen to be somewhere inside the range tikhomirov@390: * [start..end]. Implementation does it best to report empty manifests (Inspector.begin(BAD_REVISION, NULL, csetRevIndex); tikhomirov@390: * followed immediately by Inspector.end(BAD_REVISION) when start and/or end point to changeset with no associated tikhomirov@390: * manifest revision. However, if changeset-manifest revision pairs look like: tikhomirov@390: *

tikhomirov@390: 	 *   3  8
tikhomirov@390: 	 *   4  -1 (cset records null revision for manifest)
tikhomirov@390: 	 *   5  9
tikhomirov@390: 	 * 
tikhomirov@390: * call walk(3,5, insp) would yield only (3,8) and (5,9) to the inspector, without additional empty tikhomirov@390: * Inspector.begin(); Inspector.end() call pair. tikhomirov@218: * tikhomirov@218: * @param start changelog (not manifest!) revision to begin with tikhomirov@218: * @param end changelog (not manifest!) revision to stop, inclusive. tikhomirov@390: * @param inspector manifest revision visitor, can't be null tikhomirov@403: * @throws HgInvalidRevisionException if start or end specify non-existent revision index tikhomirov@403: * @throws IllegalArgumentException if start or end is not a revision index tikhomirov@390: * @throws HgInvalidControlFileException if access to revlog index/data entry failed tikhomirov@218: */ tikhomirov@403: public void walk(int start, int end, final Inspector inspector) throws HgInvalidRevisionException, HgInvalidControlFileException { tikhomirov@196: if (inspector == null) { tikhomirov@196: throw new IllegalArgumentException(); tikhomirov@196: } tikhomirov@390: final int csetFirst = start <= end ? start : end, csetLast = start > end ? start : end; tikhomirov@390: int manifestFirst, manifestLast, i = 0; tikhomirov@390: do { tikhomirov@390: manifestFirst = fromChangelog(csetFirst+i); tikhomirov@403: if (manifestFirst == BAD_REVISION) { tikhomirov@390: inspector.begin(BAD_REVISION, NULL, csetFirst+i); tikhomirov@390: inspector.end(BAD_REVISION); tikhomirov@390: } tikhomirov@390: i++; tikhomirov@403: } while (manifestFirst == BAD_REVISION && csetFirst+i <= csetLast); tikhomirov@403: if (manifestFirst == BAD_REVISION) { tikhomirov@390: getRepo().getContext().getLog().info(getClass(), "None of changesets [%d..%d] have associated manifest revision", csetFirst, csetLast); tikhomirov@390: // we ran through all revisions in [start..end] and none of them had manifest. tikhomirov@390: // we reported that to inspector and proceeding is done now. tikhomirov@390: return; tikhomirov@390: } tikhomirov@390: i = 0; tikhomirov@390: do { tikhomirov@390: manifestLast = fromChangelog(csetLast-i); tikhomirov@403: if (manifestLast == BAD_REVISION) { tikhomirov@390: inspector.begin(BAD_REVISION, NULL, csetLast-i); tikhomirov@390: inspector.end(BAD_REVISION); tikhomirov@390: } tikhomirov@390: i++; tikhomirov@403: } while (manifestLast == BAD_REVISION && csetLast-i >= csetFirst); tikhomirov@403: if (manifestLast == BAD_REVISION) { tikhomirov@390: // hmm, manifestFirst != -1 here, hence there's i from [csetFirst..csetLast] for which manifest entry exists, tikhomirov@390: // and thus it's impossible to run into manifestLast == -1. Nevertheless, never hurts to check. tikhomirov@390: throw new HgBadStateException(String.format("Manifest %d-%d(!) for cset range [%d..%d] ", manifestFirst, manifestLast, csetFirst, csetLast)); tikhomirov@390: } tikhomirov@390: if (manifestLast < manifestFirst) { tikhomirov@300: // there are tool-constructed repositories that got order of changeset revisions completely different from that of manifest tikhomirov@390: int x = manifestLast; tikhomirov@390: manifestLast = manifestFirst; tikhomirov@390: manifestFirst = x; tikhomirov@300: } tikhomirov@412: content.iterate(manifestFirst, manifestLast, true, new ManifestParser(inspector, encodingHelper)); tikhomirov@19: } tikhomirov@218: tikhomirov@254: /** tikhomirov@390: * "Sparse" iteration of the manifest, more effective than accessing revisions one by one. tikhomirov@390: *

Inspector is invoked for each changeset revision supplied, even when there's no manifest tikhomirov@390: * revision associated with a changeset (@see {@link #walk(int, int, Inspector)} for more details when it happens). Order inspector tikhomirov@390: * gets invoked doesn't resemble order of changeset revisions supplied, manifest revisions are reported in the order they appear tikhomirov@390: * in manifest revlog (with exception of changesets with missing manifest that may be reported in any order). tikhomirov@254: * tikhomirov@390: * @param inspector manifest revision visitor, can't be null tikhomirov@390: * @param revisionIndexes local indexes of changesets to visit, non-null tikhomirov@403: * @throws HgInvalidRevisionException if argument specifies non-existent revision index tikhomirov@403: * @throws HgInvalidControlFileException if access to revlog index/data entry failed tikhomirov@254: */ tikhomirov@403: public void walk(final Inspector inspector, int... revisionIndexes) throws HgInvalidRevisionException, HgInvalidControlFileException { tikhomirov@367: if (inspector == null || revisionIndexes == null) { tikhomirov@254: throw new IllegalArgumentException(); tikhomirov@254: } tikhomirov@390: int[] manifestRevs = toManifestRevisionIndexes(revisionIndexes, inspector); tikhomirov@412: content.iterate(manifestRevs, true, new ManifestParser(inspector, encodingHelper)); tikhomirov@254: } tikhomirov@254: tikhomirov@390: // tikhomirov@390: /** tikhomirov@403: * Tells manifest revision number that corresponds to the given changeset. May return {@link HgRepository#BAD_REVISION} tikhomirov@403: * if changeset has no associated manifest (cset records NULL nodeid for manifest). tikhomirov@403: * @return manifest revision index, non-negative, or {@link HgRepository#BAD_REVISION}. tikhomirov@403: * @throws HgInvalidRevisionException if method argument specifies non-existent revision index tikhomirov@403: * @throws IllegalArgumentException if argument is not a revision index tikhomirov@403: * @throws HgInvalidControlFileException if access to revlog index/data entry failed tikhomirov@390: */ tikhomirov@403: /*package-local*/ int fromChangelog(int changesetRevisionIndex) throws HgInvalidRevisionException, HgInvalidControlFileException { tikhomirov@368: if (HgInternals.wrongRevisionIndex(changesetRevisionIndex)) { tikhomirov@368: throw new IllegalArgumentException(String.valueOf(changesetRevisionIndex)); tikhomirov@218: } tikhomirov@368: if (changesetRevisionIndex == HgRepository.WORKING_COPY || changesetRevisionIndex == HgRepository.BAD_REVISION) { tikhomirov@257: throw new IllegalArgumentException("Can't use constants like WORKING_COPY or BAD_REVISION"); tikhomirov@257: } tikhomirov@257: // revisionNumber == TIP is processed by RevisionMapper tikhomirov@218: if (revisionMap == null) { tikhomirov@218: revisionMap = new RevisionMapper(getRepo()); tikhomirov@218: content.iterate(0, TIP, false, revisionMap); tikhomirov@218: } tikhomirov@368: return revisionMap.at(changesetRevisionIndex); tikhomirov@218: } tikhomirov@218: tikhomirov@232: /** tikhomirov@232: * Extracts file revision as it was known at the time of given changeset. tikhomirov@232: * tikhomirov@367: * @param changelogRevisionIndex local changeset index tikhomirov@232: * @param file path to file in question tikhomirov@232: * @return file revision or null if manifest at specified revision doesn't list such file tikhomirov@403: * @throws HgInvalidRevisionException if method argument specifies non-existent revision index tikhomirov@403: * @throws HgInvalidControlFileException if access to revlog index/data entry failed tikhomirov@232: */ tikhomirov@326: @Experimental(reason="Perhaps, HgDataFile shall own this method, or get a delegate?") tikhomirov@403: public Nodeid getFileRevision(int changelogRevisionIndex, final Path file) throws HgInvalidRevisionException, HgInvalidControlFileException { tikhomirov@367: return getFileRevisions(file, changelogRevisionIndex).get(changelogRevisionIndex); tikhomirov@326: } tikhomirov@326: tikhomirov@326: // XXX package-local, IntMap, and HgDataFile getFileRevisionAt(int... localChangelogRevisions) tikhomirov@326: @Experimental(reason="@see #getFileRevision") tikhomirov@403: public Map getFileRevisions(final Path file, int... changelogRevisionIndexes) throws HgInvalidRevisionException, HgInvalidControlFileException { tikhomirov@403: // TODO need tests tikhomirov@390: int[] manifestRevisionIndexes = toManifestRevisionIndexes(changelogRevisionIndexes, null); tikhomirov@367: final HashMap rv = new HashMap(changelogRevisionIndexes.length); tikhomirov@367: content.iterate(manifestRevisionIndexes, true, new RevlogStream.Inspector() { tikhomirov@232: tikhomirov@366: public void next(int revisionNumber, int actualLen, int baseRevision, int linkRevision, int parent1Revision, int parent2Revision, byte[] nodeid, DataAccess data) throws HgException { tikhomirov@232: ByteArrayOutputStream bos = new ByteArrayOutputStream(); tikhomirov@232: try { tikhomirov@232: byte b; tikhomirov@232: while (!data.isEmpty() && (b = data.readByte()) != '\n') { tikhomirov@232: if (b != 0) { tikhomirov@232: bos.write(b); tikhomirov@232: } else { tikhomirov@232: String fname = new String(bos.toByteArray()); tikhomirov@232: bos.reset(); tikhomirov@232: if (file.toString().equals(fname)) { tikhomirov@232: byte[] nid = new byte[40]; tikhomirov@232: data.readBytes(nid, 0, 40); tikhomirov@326: rv.put(linkRevision, Nodeid.fromAscii(nid, 0, 40)); tikhomirov@232: break; tikhomirov@326: } else { tikhomirov@326: data.skip(40); tikhomirov@232: } tikhomirov@232: // else skip to the end of line tikhomirov@232: while (!data.isEmpty() && (b = data.readByte()) != '\n') tikhomirov@232: ; tikhomirov@232: } tikhomirov@232: } tikhomirov@232: } catch (IOException ex) { tikhomirov@366: throw new HgException(ex); tikhomirov@232: } tikhomirov@232: } tikhomirov@232: }); tikhomirov@326: return rv; tikhomirov@232: } tikhomirov@326: tikhomirov@326: tikhomirov@390: /** tikhomirov@390: * @param changelogRevisionIndexes non-null tikhomirov@390: * @param inspector may be null if reporting of missing manifests is not needed tikhomirov@403: * @throws HgInvalidRevisionException if arguments specify non-existent revision index tikhomirov@403: * @throws IllegalArgumentException if any index argument is not a revision index tikhomirov@403: * @throws HgInvalidControlFileException if access to revlog index/data entry failed tikhomirov@390: */ tikhomirov@403: private int[] toManifestRevisionIndexes(int[] changelogRevisionIndexes, Inspector inspector) throws HgInvalidRevisionException, HgInvalidControlFileException { tikhomirov@368: int[] manifestRevs = new int[changelogRevisionIndexes.length]; tikhomirov@326: boolean needsSort = false; tikhomirov@390: int j = 0; tikhomirov@367: for (int i = 0; i < changelogRevisionIndexes.length; i++) { tikhomirov@367: final int manifestRevisionIndex = fromChangelog(changelogRevisionIndexes[i]); tikhomirov@403: if (manifestRevisionIndex == BAD_REVISION) { tikhomirov@390: if (inspector != null) { tikhomirov@390: inspector.begin(BAD_REVISION, NULL, changelogRevisionIndexes[i]); tikhomirov@390: inspector.end(BAD_REVISION); tikhomirov@390: } tikhomirov@390: // othrwise, ignore changeset without manifest tikhomirov@390: } else { tikhomirov@390: manifestRevs[j] = manifestRevisionIndex; tikhomirov@390: if (j > 0 && manifestRevs[j-1] > manifestRevisionIndex) { tikhomirov@390: needsSort = true; tikhomirov@390: } tikhomirov@390: j++; tikhomirov@326: } tikhomirov@326: } tikhomirov@326: if (needsSort) { tikhomirov@390: Arrays.sort(manifestRevs, 0, j); tikhomirov@326: } tikhomirov@390: if (j == manifestRevs.length) { tikhomirov@390: return manifestRevs; tikhomirov@390: } else { tikhomirov@390: int[] rv = new int[j]; tikhomirov@390: //Arrays.copyOfRange tikhomirov@390: System.arraycopy(manifestRevs, 0, rv, 0, j); tikhomirov@390: return rv; tikhomirov@390: } tikhomirov@326: } tikhomirov@326: tikhomirov@19: public interface Inspector { tikhomirov@222: boolean begin(int mainfestRevision, Nodeid nid, int changelogRevision); tikhomirov@285: /** tikhomirov@285: * @deprecated switch to {@link Inspector2#next(Nodeid, Path, Flags)} tikhomirov@285: */ tikhomirov@285: @Deprecated tikhomirov@19: boolean next(Nodeid nid, String fname, String flags); tikhomirov@222: boolean end(int manifestRevision); tikhomirov@2: } tikhomirov@267: tikhomirov@285: @Experimental(reason="Explore Path alternative for filenames and enum for flags") tikhomirov@285: public interface Inspector2 extends Inspector { tikhomirov@285: boolean next(Nodeid nid, Path fname, Flags flags); tikhomirov@267: } tikhomirov@326: tikhomirov@268: /** tikhomirov@268: * When Pool uses Strings directly, tikhomirov@268: * ManifestParser creates new String instance with new char[] value, and does byte->char conversion. tikhomirov@268: * For cpython repo, walk(0..10k), there are over 16 million filenames, of them only 3020 unique. tikhomirov@268: * This means there are 15.9 million useless char[] instances and byte->char conversions tikhomirov@268: * tikhomirov@285: * When String (Path) is wrapped into {@link PathProxy}, there's extra overhead of byte[] representation tikhomirov@285: * of the String, but these are only for unique Strings (Paths) (3020 in the example above). Besides, I save tikhomirov@268: * useless char[] and byte->char conversions. tikhomirov@268: */ tikhomirov@285: private static class PathProxy { tikhomirov@267: private byte[] data; tikhomirov@268: private int start; tikhomirov@268: private final int hash, length; tikhomirov@285: private Path result; tikhomirov@412: private final EncodingHelper encHelper; tikhomirov@267: tikhomirov@412: public PathProxy(byte[] data, int start, int length, EncodingHelper eh) { tikhomirov@267: this.data = data; tikhomirov@267: this.start = start; tikhomirov@267: this.length = length; tikhomirov@412: this.encHelper = eh; tikhomirov@267: tikhomirov@268: // copy from String.hashCode(). In fact, not necessarily match result of String(data).hashCode tikhomirov@268: // just need some nice algorithm here tikhomirov@267: int h = 0; tikhomirov@267: byte[] d = data; tikhomirov@267: for (int i = 0, off = start, len = length; i < len; i++) { tikhomirov@267: h = 31 * h + d[off++]; tikhomirov@267: } tikhomirov@267: hash = h; tikhomirov@267: } tikhomirov@267: tikhomirov@267: @Override tikhomirov@267: public boolean equals(Object obj) { tikhomirov@285: if (false == obj instanceof PathProxy) { tikhomirov@267: return false; tikhomirov@267: } tikhomirov@285: PathProxy o = (PathProxy) obj; tikhomirov@267: if (o.result != null && result != null) { tikhomirov@267: return result.equals(o.result); tikhomirov@267: } tikhomirov@268: if (o.length != length || o.hash != hash) { tikhomirov@267: return false; tikhomirov@267: } tikhomirov@267: for (int i = 0, x = o.start, y = start; i < length; i++) { tikhomirov@267: if (o.data[x++] != data[y++]) { tikhomirov@267: return false; tikhomirov@267: } tikhomirov@267: } tikhomirov@267: return true; tikhomirov@267: } tikhomirov@267: @Override tikhomirov@267: public int hashCode() { tikhomirov@267: return hash; tikhomirov@267: } tikhomirov@267: tikhomirov@285: public Path freeze() { tikhomirov@267: if (result == null) { tikhomirov@412: result = Path.create(encHelper.fromManifest(data, start, length)); tikhomirov@268: // release reference to bigger data array, make a copy of relevant part only tikhomirov@285: // use original bytes, not those from String above to avoid cache misses due to different encodings tikhomirov@268: byte[] d = new byte[length]; tikhomirov@268: System.arraycopy(data, start, d, 0, length); tikhomirov@268: data = d; tikhomirov@268: start = 0; tikhomirov@267: } tikhomirov@267: return result; tikhomirov@267: } tikhomirov@267: } tikhomirov@196: tikhomirov@312: private static class ManifestParser implements RevlogStream.Inspector, Lifecycle { tikhomirov@196: private final Inspector inspector; tikhomirov@285: private final Inspector2 inspector2; tikhomirov@304: private Pool2 nodeidPool, thisRevPool; tikhomirov@304: private final Pool2 fnamePool; tikhomirov@266: private byte[] nodeidLookupBuffer = new byte[20]; // get reassigned each time new Nodeid is added to pool tikhomirov@312: private final ProgressSupport progressHelper; tikhomirov@312: private IterateControlMediator iterateControl; tikhomirov@412: private final EncodingHelper encHelper; tikhomirov@262: tikhomirov@412: public ManifestParser(Inspector delegate, EncodingHelper eh) { tikhomirov@196: assert delegate != null; tikhomirov@196: inspector = delegate; tikhomirov@285: inspector2 = delegate instanceof Inspector2 ? (Inspector2) delegate : null; tikhomirov@412: encHelper = eh; tikhomirov@304: nodeidPool = new Pool2(); tikhomirov@304: fnamePool = new Pool2(); tikhomirov@304: thisRevPool = new Pool2(); tikhomirov@312: progressHelper = ProgressSupport.Factory.get(delegate); tikhomirov@196: } tikhomirov@268: tikhomirov@366: public void next(int revisionNumber, int actualLen, int baseRevision, int linkRevision, int parent1Revision, int parent2Revision, byte[] nodeid, DataAccess da) throws HgException { tikhomirov@196: try { tikhomirov@312: if (!inspector.begin(revisionNumber, new Nodeid(nodeid, true), linkRevision)) { tikhomirov@312: iterateControl.stop(); tikhomirov@312: return; tikhomirov@312: } tikhomirov@372: if (!da.isEmpty()) { tikhomirov@372: // although unlikely, manifest entry may be empty, when all files have been deleted from the repository tikhomirov@372: Path fname = null; tikhomirov@372: Flags flags = null; tikhomirov@372: Nodeid nid = null; tikhomirov@372: int i; tikhomirov@372: byte[] data = da.byteArray(); tikhomirov@372: for (i = 0; i < actualLen; i++) { tikhomirov@372: int x = i; tikhomirov@372: for( ; data[i] != '\n' && i < actualLen; i++) { tikhomirov@372: if (fname == null && data[i] == 0) { tikhomirov@412: PathProxy px = fnamePool.unify(new PathProxy(data, x, i - x, encHelper)); tikhomirov@372: // if (cached = fnamePool.unify(px))== px then cacheMiss, else cacheHit tikhomirov@372: // cpython 0..10k: hits: 15 989 152, misses: 3020 tikhomirov@372: fname = px.freeze(); tikhomirov@372: x = i+1; tikhomirov@372: } tikhomirov@196: } tikhomirov@372: if (i < actualLen) { tikhomirov@372: assert data[i] == '\n'; tikhomirov@372: int nodeidLen = i - x < 40 ? i-x : 40; // if > 40, there are flags tikhomirov@372: DigestHelper.ascii2bin(data, x, nodeidLen, nodeidLookupBuffer); // ignore return value as it's unlikely to have NULL in manifest tikhomirov@372: nid = new Nodeid(nodeidLookupBuffer, false); // this Nodeid is for pool lookup only, mock object tikhomirov@372: Nodeid cached = nodeidPool.unify(nid); tikhomirov@372: if (cached == nid) { tikhomirov@372: // buffer now belongs to the cached nodeid tikhomirov@372: nodeidLookupBuffer = new byte[20]; tikhomirov@372: } else { tikhomirov@372: nid = cached; // use existing version, discard the lookup object tikhomirov@372: } // for cpython 0..10k, cache hits are 15 973 301, vs 18871 misses. tikhomirov@372: thisRevPool.record(nid); // memorize revision for the next iteration. tikhomirov@372: if (nodeidLen + x < i) { tikhomirov@372: // 'x' and 'l' for executable bits and symlinks? tikhomirov@372: // hg --debug manifest shows 644 for each regular file in my repo tikhomirov@372: // for cpython 0..10k, there are 4361062 flag checks, and there's only 1 unique flag tikhomirov@372: flags = Flags.parse(data, x + nodeidLen, i-x-nodeidLen); tikhomirov@372: } else { tikhomirov@372: flags = null; tikhomirov@372: } tikhomirov@372: boolean good2go; tikhomirov@372: if (inspector2 == null) { tikhomirov@372: String flagString = flags == null ? null : flags.nativeString(); tikhomirov@372: good2go = inspector.next(nid, fname.toString(), flagString); tikhomirov@372: } else { tikhomirov@372: good2go = inspector2.next(nid, fname, flags); tikhomirov@372: } tikhomirov@372: if (!good2go) { tikhomirov@372: iterateControl.stop(); tikhomirov@372: return; tikhomirov@372: } tikhomirov@372: } tikhomirov@372: nid = null; tikhomirov@372: fname = null; tikhomirov@372: flags = null; tikhomirov@196: } tikhomirov@196: } tikhomirov@312: if (!inspector.end(revisionNumber)) { tikhomirov@312: iterateControl.stop(); tikhomirov@312: return; tikhomirov@312: } tikhomirov@262: // tikhomirov@262: // keep only actual file revisions, found at this version tikhomirov@262: // (next manifest is likely to refer to most of them, although in specific cases tikhomirov@262: // like commit in another branch a lot may be useless) tikhomirov@262: nodeidPool.clear(); tikhomirov@304: Pool2 t = nodeidPool; tikhomirov@262: nodeidPool = thisRevPool; tikhomirov@264: thisRevPool = t; tikhomirov@313: iterateControl.checkCancelled(); tikhomirov@312: progressHelper.worked(1); tikhomirov@196: } catch (IOException ex) { tikhomirov@366: throw new HgException(ex); tikhomirov@196: } tikhomirov@196: } tikhomirov@312: tikhomirov@312: public void start(int count, Callback callback, Object token) { tikhomirov@312: CancelSupport cs = CancelSupport.Factory.get(inspector, null); tikhomirov@312: iterateControl = new IterateControlMediator(cs, callback); tikhomirov@312: progressHelper.start(count); tikhomirov@312: } tikhomirov@312: tikhomirov@312: public void finish(Object token) { tikhomirov@312: progressHelper.done(); tikhomirov@312: } tikhomirov@196: } tikhomirov@218: tikhomirov@218: private static class RevisionMapper implements RevlogStream.Inspector, Lifecycle { tikhomirov@218: tikhomirov@403: private final int changelogRevisionCount; tikhomirov@218: private int[] changelog2manifest; tikhomirov@218: private final HgRepository repo; tikhomirov@218: tikhomirov@218: public RevisionMapper(HgRepository hgRepo) { tikhomirov@218: repo = hgRepo; tikhomirov@403: changelogRevisionCount = repo.getChangelog().getRevisionCount(); tikhomirov@218: } tikhomirov@218: tikhomirov@403: /** tikhomirov@403: * Get index of manifest revision that corresponds to specified changeset tikhomirov@403: * @param changesetRevisionIndex non-negative index of changelog revision, or {@link HgRepository#TIP} tikhomirov@403: * @return index of manifest revision, or {@link HgRepository#BAD_REVISION} if changeset doesn't reference a valid manifest tikhomirov@403: * @throws HgInvalidRevisionException if method argument specifies non-existent revision index tikhomirov@403: */ tikhomirov@403: public int at(int changesetRevisionIndex) throws HgInvalidRevisionException { tikhomirov@403: if (changesetRevisionIndex == TIP) { tikhomirov@403: changesetRevisionIndex = changelogRevisionCount - 1; tikhomirov@403: } tikhomirov@403: if (changesetRevisionIndex >= changelogRevisionCount) { tikhomirov@403: throw new HgInvalidRevisionException(changesetRevisionIndex); tikhomirov@257: } tikhomirov@218: if (changelog2manifest != null) { tikhomirov@403: return changelog2manifest[changesetRevisionIndex]; tikhomirov@218: } tikhomirov@403: return changesetRevisionIndex; tikhomirov@218: } tikhomirov@218: tikhomirov@324: // XXX likely can be replaced with Revlog.RevisionInspector tikhomirov@218: public void next(int revisionNumber, int actualLen, int baseRevision, int linkRevision, int parent1Revision, int parent2Revision, byte[] nodeid, DataAccess data) { tikhomirov@218: if (changelog2manifest != null) { tikhomirov@220: // next assertion is not an error, rather assumption check, which is too development-related to be explicit exception - tikhomirov@220: // I just wonder if there are manifests that have two entries pointing to single changeset. It seems unrealistic, though - tikhomirov@220: // changeset records one and only one manifest nodeid tikhomirov@403: assert changelog2manifest[linkRevision] == BAD_REVISION : String.format("revision:%d, link:%d, already linked to revision:%d", revisionNumber, linkRevision, changelog2manifest[linkRevision]); tikhomirov@218: changelog2manifest[linkRevision] = revisionNumber; tikhomirov@218: } else { tikhomirov@218: if (revisionNumber != linkRevision) { tikhomirov@403: changelog2manifest = new int[changelogRevisionCount]; tikhomirov@403: Arrays.fill(changelog2manifest, BAD_REVISION); tikhomirov@218: for (int i = 0; i < revisionNumber; changelog2manifest[i] = i, i++) tikhomirov@218: ; tikhomirov@218: changelog2manifest[linkRevision] = revisionNumber; tikhomirov@218: } tikhomirov@218: } tikhomirov@218: } tikhomirov@218: tikhomirov@218: public void start(int count, Callback callback, Object token) { tikhomirov@403: if (count != changelogRevisionCount) { tikhomirov@403: assert count < changelogRevisionCount; // no idea what to do if manifest has more revisions than changelog tikhomirov@218: // the way how manifest may contain more revisions than changelog, as I can imagine, is a result of tikhomirov@218: // some kind of an import tool (e.g. from SVN or CVS), that creates manifest and changelog independently. tikhomirov@218: // Note, it's pure guess, I didn't see such repository yet (although the way manifest revisions tikhomirov@218: // in cpython repo are numbered makes me think aforementioned way) tikhomirov@403: changelog2manifest = new int[changelogRevisionCount]; tikhomirov@403: Arrays.fill(changelog2manifest, BAD_REVISION); tikhomirov@218: } tikhomirov@218: } tikhomirov@218: tikhomirov@218: public void finish(Object token) { tikhomirov@218: if (changelog2manifest == null) { tikhomirov@218: return; tikhomirov@218: } tikhomirov@218: // I assume there'd be not too many revisions we don't know manifest of tikhomirov@218: ArrayList undefinedChangelogRevision = new ArrayList(); tikhomirov@218: for (int i = 0; i < changelog2manifest.length; i++) { tikhomirov@403: if (changelog2manifest[i] == BAD_REVISION) { tikhomirov@218: undefinedChangelogRevision.add(i); tikhomirov@218: } tikhomirov@218: } tikhomirov@218: for (int u : undefinedChangelogRevision) { tikhomirov@354: try { tikhomirov@366: Nodeid manifest = repo.getChangelog().range(u, u).get(0).manifest(); tikhomirov@390: // TODO calculate those missing effectively (e.g. cache and sort nodeids to speed lookup tikhomirov@366: // right away in the #next (may refactor ParentWalker's sequential and sorted into dedicated helper and reuse here) tikhomirov@385: if (manifest.isNull()) { tikhomirov@385: repo.getContext().getLog().warn(getClass(), "Changeset %d has no associated manifest entry", u); tikhomirov@390: // keep -1 in the changelog2manifest map. tikhomirov@385: } else { tikhomirov@385: changelog2manifest[u] = repo.getManifest().getRevisionIndex(manifest); tikhomirov@385: } tikhomirov@354: } catch (HgInvalidControlFileException ex) { tikhomirov@354: // FIXME need to propagate the error up to client tikhomirov@354: repo.getContext().getLog().error(getClass(), ex, null); tikhomirov@354: } tikhomirov@218: } tikhomirov@218: } tikhomirov@218: } tikhomirov@2: }