Package org.eclipse.jgit.transport

Source Code of org.eclipse.jgit.transport.IndexPack

/*
* Copyright (C) 2008-2010, Google Inc.
* Copyright (C) 2007-2008, Robin Rosenberg <robin.rosenberg@dewire.com>
* Copyright (C) 2008, Shawn O. Pearce <spearce@spearce.org>
* and other copyright owners as documented in the project's IP log.
*
* This program and the accompanying materials are made available
* under the terms of the Eclipse Distribution License v1.0 which
* accompanies this distribution, is reproduced below, and is
* available at http://www.eclipse.org/org/documents/edl-v10.php
*
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or
* without modification, are permitted provided that the following
* conditions are met:
*
* - Redistributions of source code must retain the above copyright
*   notice, this list of conditions and the following disclaimer.
*
* - Redistributions in binary form must reproduce the above
*   copyright notice, this list of conditions and the following
*   disclaimer in the documentation and/or other materials provided
*   with the distribution.
*
* - Neither the name of the Eclipse Foundation, Inc. nor the
*   names of its contributors may be used to endorse or promote
*   products derived from this software without specific prior
*   written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND
* CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES,
* INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
* OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
* SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
* NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
* LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
* CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
* STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
* ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF
* ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/

package org.eclipse.jgit.transport;

import java.io.EOFException;
import java.io.File;
import java.io.FileOutputStream;
import java.io.IOException;
import java.io.InputStream;
import java.io.RandomAccessFile;
import java.security.MessageDigest;
import java.text.MessageFormat;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.List;
import java.util.zip.CRC32;
import java.util.zip.DataFormatException;
import java.util.zip.Deflater;
import java.util.zip.Inflater;

import org.eclipse.jgit.JGitText;
import org.eclipse.jgit.errors.CorruptObjectException;
import org.eclipse.jgit.errors.MissingObjectException;
import org.eclipse.jgit.lib.AnyObjectId;
import org.eclipse.jgit.lib.Constants;
import org.eclipse.jgit.lib.CoreConfig;
import org.eclipse.jgit.lib.InflaterCache;
import org.eclipse.jgit.lib.MutableObjectId;
import org.eclipse.jgit.lib.ObjectChecker;
import org.eclipse.jgit.lib.ObjectDatabase;
import org.eclipse.jgit.lib.ObjectId;
import org.eclipse.jgit.lib.ObjectIdSubclassMap;
import org.eclipse.jgit.lib.ObjectLoader;
import org.eclipse.jgit.lib.ProgressMonitor;
import org.eclipse.jgit.lib.Repository;
import org.eclipse.jgit.lib.ObjectReader;
import org.eclipse.jgit.storage.file.PackIndexWriter;
import org.eclipse.jgit.storage.file.PackLock;
import org.eclipse.jgit.storage.pack.BinaryDelta;
import org.eclipse.jgit.util.NB;

/** Indexes Git pack files for local use. */
public class IndexPack {
  /** Progress message when reading raw data from the pack. */
  public static final String PROGRESS_DOWNLOAD = JGitText.get().receivingObjects;

  /** Progress message when computing names of delta compressed objects. */
  public static final String PROGRESS_RESOLVE_DELTA = JGitText.get().resolvingDeltas;

  /**
   * Size of the internal stream buffer.
   * <p>
   * If callers are going to be supplying IndexPack a BufferedInputStream they
   * should use this buffer size as the size of the buffer for that
   * BufferedInputStream, and any other its may be wrapping. This way the
   * buffers will cascade efficiently and only the IndexPack buffer will be
   * receiving the bulk of the data stream.
   */
  public static final int BUFFER_SIZE = 8192;

  /**
   * Create an index pack instance to load a new pack into a repository.
   * <p>
   * The received pack data and generated index will be saved to temporary
   * files within the repository's <code>objects</code> directory. To use the
   * data contained within them call {@link #renameAndOpenPack()} once the
   * indexing is complete.
   *
   * @param db
   *            the repository that will receive the new pack.
   * @param is
   *            stream to read the pack data from. If the stream is buffered
   *            use {@link #BUFFER_SIZE} as the buffer size for the stream.
   * @return a new index pack instance.
   * @throws IOException
   *             a temporary file could not be created.
   */
  public static IndexPack create(final Repository db, final InputStream is)
      throws IOException {
    final String suffix = ".pack";
    final File objdir = db.getObjectsDirectory();
    final File tmp = File.createTempFile("incoming_", suffix, objdir);
    final String n = tmp.getName();
    final File base;

    base = new File(objdir, n.substring(0, n.length() - suffix.length()));
    final IndexPack ip = new IndexPack(db, is, base);
    ip.setIndexVersion(db.getConfig().get(CoreConfig.KEY)
        .getPackIndexVersion());
    return ip;
  }

  private static enum Source {
    /** Data is read from the incoming stream. */
    INPUT,

    /**
     * Data is read from the spooled pack file.
     * <p>
     * During streaming, some (or all) data might be saved into the spooled
     * pack file so it can be randomly accessed later.
     */
    FILE;
  }

  private final Repository repo;

  /**
   * Object database used for loading existing objects
   */
  private final ObjectDatabase objectDatabase;

  private Inflater inflater;

  private final MessageDigest objectDigest;

  private final MutableObjectId tempObjectId;

  private InputStream in;

  private byte[] buf;

  private long bBase;

  private int bOffset;

  private int bAvail;

  private ObjectChecker objCheck;

  private boolean fixThin;

  private boolean keepEmpty;

  private boolean needBaseObjectIds;

  private int outputVersion;

  private final File dstPack;

  private final File dstIdx;

  private long objectCount;

  private PackedObjectInfo[] entries;

  /**
   * Every object contained within the incoming pack.
   * <p>
   * This is a subset of {@link #entries}, as thin packs can add additional
   * objects to {@code entries} by copying already existing objects from the
   * repository onto the end of the thin pack to make it self-contained.
   */
  private ObjectIdSubclassMap<ObjectId> newObjectIds;

  private int deltaCount;

  private int entryCount;

  private final CRC32 crc = new CRC32();

  private ObjectIdSubclassMap<DeltaChain> baseById;

  /**
   * Objects referenced by their name from deltas, that aren't in this pack.
   * <p>
   * This is the set of objects that were copied onto the end of this pack to
   * make it complete. These objects were not transmitted by the remote peer,
   * but instead were assumed to already exist in the local repository.
   */
  private ObjectIdSubclassMap<ObjectId> baseObjectIds;

  private LongMap<UnresolvedDelta> baseByPos;

  private byte[] skipBuffer;

  private MessageDigest packDigest;

  private RandomAccessFile packOut;

  private byte[] packcsum;

  /** If {@link #fixThin} this is the last byte of the original checksum. */
  private long originalEOF;

  private ObjectReader readCurs;

  /**
   * Create a new pack indexer utility.
   *
   * @param db
   * @param src
   *            stream to read the pack data from. If the stream is buffered
   *            use {@link #BUFFER_SIZE} as the buffer size for the stream.
   * @param dstBase
   * @throws IOException
   *             the output packfile could not be created.
   */
  public IndexPack(final Repository db, final InputStream src,
      final File dstBase) throws IOException {
    repo = db;
    objectDatabase = db.getObjectDatabase().newCachedDatabase();
    in = src;
    inflater = InflaterCache.get();
    readCurs = objectDatabase.newReader();
    buf = new byte[BUFFER_SIZE];
    skipBuffer = new byte[512];
    objectDigest = Constants.newMessageDigest();
    tempObjectId = new MutableObjectId();
    packDigest = Constants.newMessageDigest();

    if (dstBase != null) {
      final File dir = dstBase.getParentFile();
      final String nam = dstBase.getName();
      dstPack = new File(dir, nam + ".pack");
      dstIdx = new File(dir, nam + ".idx");
      packOut = new RandomAccessFile(dstPack, "rw");
      packOut.setLength(0);
    } else {
      dstPack = null;
      dstIdx = null;
    }
  }

  /**
   * Set the pack index file format version this instance will create.
   *
   * @param version
   *            the version to write. The special version 0 designates the
   *            oldest (most compatible) format available for the objects.
   * @see PackIndexWriter
   */
  public void setIndexVersion(final int version) {
    outputVersion = version;
  }

  /**
   * Configure this index pack instance to make a thin pack complete.
   * <p>
   * Thin packs are sometimes used during network transfers to allow a delta
   * to be sent without a base object. Such packs are not permitted on disk.
   * They can be fixed by copying the base object onto the end of the pack.
   *
   * @param fix
   *            true to enable fixing a thin pack.
   */
  public void setFixThin(final boolean fix) {
    fixThin = fix;
  }

  /**
   * Configure this index pack instance to keep an empty pack.
   * <p>
   * By default an empty pack (a pack with no objects) is not kept, as doing
   * so is completely pointless. With no objects in the pack there is no data
   * stored by it, so the pack is unnecessary.
   *
   * @param empty true to enable keeping an empty pack.
   */
  public void setKeepEmpty(final boolean empty) {
    keepEmpty = empty;
  }

  /**
   * Configure this index pack instance to keep track of new objects.
   * <p>
   * By default an index pack doesn't save the new objects that were created
   * when it was instantiated. Setting this flag to {@code true} allows the
   * caller to use {@link #getNewObjectIds()} to retrieve that list.
   *
   * @param b {@code true} to enable keeping track of new objects.
   */
  public void setNeedNewObjectIds(boolean b) {
    if (b)
      newObjectIds = new ObjectIdSubclassMap<ObjectId>();
    else
      newObjectIds = null;
  }

  private boolean needNewObjectIds() {
    return newObjectIds != null;
  }

  /**
   * Configure this index pack instance to keep track of the objects assumed
   * for delta bases.
   * <p>
   * By default an index pack doesn't save the objects that were used as delta
   * bases. Setting this flag to {@code true} will allow the caller to
   * use {@link #getBaseObjectIds()} to retrieve that list.
   *
   * @param b {@code true} to enable keeping track of delta bases.
   */
  public void setNeedBaseObjectIds(boolean b) {
    this.needBaseObjectIds = b;
  }

  /** @return the new objects that were sent by the user */
  public ObjectIdSubclassMap<ObjectId> getNewObjectIds() {
    if (newObjectIds != null)
      return newObjectIds;
    return new ObjectIdSubclassMap<ObjectId>();
  }

  /** @return set of objects the incoming pack assumed for delta purposes */
  public ObjectIdSubclassMap<ObjectId> getBaseObjectIds() {
    if (baseObjectIds != null)
      return baseObjectIds;
    return new ObjectIdSubclassMap<ObjectId>();
  }

  /**
   * Configure the checker used to validate received objects.
   * <p>
   * Usually object checking isn't necessary, as Git implementations only
   * create valid objects in pack files. However, additional checking may be
   * useful if processing data from an untrusted source.
   *
   * @param oc
   *            the checker instance; null to disable object checking.
   */
  public void setObjectChecker(final ObjectChecker oc) {
    objCheck = oc;
  }

  /**
   * Configure the checker used to validate received objects.
   * <p>
   * Usually object checking isn't necessary, as Git implementations only
   * create valid objects in pack files. However, additional checking may be
   * useful if processing data from an untrusted source.
   * <p>
   * This is shorthand for:
   *
   * <pre>
   * setObjectChecker(on ? new ObjectChecker() : null);
   * </pre>
   *
   * @param on
   *            true to enable the default checker; false to disable it.
   */
  public void setObjectChecking(final boolean on) {
    setObjectChecker(on ? new ObjectChecker() : null);
  }

  /**
   * Consume data from the input stream until the packfile is indexed.
   *
   * @param progress
   *            progress feedback
   *
   * @throws IOException
   */
  public void index(final ProgressMonitor progress) throws IOException {
    progress.start(2 /* tasks */);
    try {
      try {
        readPackHeader();

        entries = new PackedObjectInfo[(int) objectCount];
        baseById = new ObjectIdSubclassMap<DeltaChain>();
        baseByPos = new LongMap<UnresolvedDelta>();

        progress.beginTask(PROGRESS_DOWNLOAD, (int) objectCount);
        for (int done = 0; done < objectCount; done++) {
          indexOneObject();
          progress.update(1);
          if (progress.isCancelled())
            throw new IOException(JGitText.get().downloadCancelled);
        }
        readPackFooter();
        endInput();
        progress.endTask();
        if (deltaCount > 0) {
          if (packOut == null)
            throw new IOException(JGitText.get().needPackOut);
          resolveDeltas(progress);
          if (entryCount < objectCount) {
            if (!fixThin) {
              throw new IOException(MessageFormat.format(
                  JGitText.get().packHasUnresolvedDeltas, (objectCount - entryCount)));
            }
            fixThinPack(progress);
          }
        }
        if (packOut != null && (keepEmpty || entryCount > 0))
          packOut.getChannel().force(true);

        packDigest = null;
        baseById = null;
        baseByPos = null;

        if (dstIdx != null && (keepEmpty || entryCount > 0))
          writeIdx();

      } finally {
        try {
          if (readCurs != null)
            readCurs.release();
        } finally {
          readCurs = null;
        }

        try {
          InflaterCache.release(inflater);
        } finally {
          inflater = null;
          objectDatabase.close();
        }

        progress.endTask();
        if (packOut != null)
          packOut.close();
      }

      if (keepEmpty || entryCount > 0) {
        if (dstPack != null)
          dstPack.setReadOnly();
        if (dstIdx != null)
          dstIdx.setReadOnly();
      }
    } catch (IOException err) {
      if (dstPack != null)
        dstPack.delete();
      if (dstIdx != null)
        dstIdx.delete();
      throw err;
    }
  }

  private void resolveDeltas(final ProgressMonitor progress)
      throws IOException {
    progress.beginTask(PROGRESS_RESOLVE_DELTA, deltaCount);
    final int last = entryCount;
    for (int i = 0; i < last; i++) {
      final int before = entryCount;
      resolveDeltas(entries[i]);
      progress.update(entryCount - before);
      if (progress.isCancelled())
        throw new IOException(JGitText.get().downloadCancelledDuringIndexing);
    }
    progress.endTask();
  }

  private void resolveDeltas(final PackedObjectInfo oe) throws IOException {
    final int oldCRC = oe.getCRC();
    if (baseById.get(oe) != null || baseByPos.containsKey(oe.getOffset()))
      resolveDeltas(oe.getOffset(), oldCRC, Constants.OBJ_BAD, null, oe);
  }

  private void resolveDeltas(final long pos, final int oldCRC, int type,
      byte[] data, PackedObjectInfo oe) throws IOException {
    crc.reset();
    position(pos);
    int c = readFrom(Source.FILE);
    final int typeCode = (c >> 4) & 7;
    long sz = c & 15;
    int shift = 4;
    while ((c & 0x80) != 0) {
      c = readFrom(Source.FILE);
      sz += (c & 0x7f) << shift;
      shift += 7;
    }

    switch (typeCode) {
    case Constants.OBJ_COMMIT:
    case Constants.OBJ_TREE:
    case Constants.OBJ_BLOB:
    case Constants.OBJ_TAG:
      type = typeCode;
      data = inflateAndReturn(Source.FILE, sz);
      break;
    case Constants.OBJ_OFS_DELTA: {
      c = readFrom(Source.FILE) & 0xff;
      while ((c & 128) != 0)
        c = readFrom(Source.FILE) & 0xff;
      data = BinaryDelta.apply(data, inflateAndReturn(Source.FILE, sz));
      break;
    }
    case Constants.OBJ_REF_DELTA: {
      crc.update(buf, fill(Source.FILE, 20), 20);
      use(20);
      data = BinaryDelta.apply(data, inflateAndReturn(Source.FILE, sz));
      break;
    }
    default:
      throw new IOException(MessageFormat.format(JGitText.get().unknownObjectType, typeCode));
    }

    final int crc32 = (int) crc.getValue();
    if (oldCRC != crc32)
      throw new IOException(MessageFormat.format(JGitText.get().corruptionDetectedReReadingAt, pos));
    if (oe == null) {
      objectDigest.update(Constants.encodedTypeString(type));
      objectDigest.update((byte) ' ');
      objectDigest.update(Constants.encodeASCII(data.length));
      objectDigest.update((byte) 0);
      objectDigest.update(data);
      tempObjectId.fromRaw(objectDigest.digest(), 0);

      verifySafeObject(tempObjectId, type, data);
      oe = new PackedObjectInfo(pos, crc32, tempObjectId);
      addObjectAndTrack(oe);
    }

    resolveChildDeltas(pos, type, data, oe);
  }

  private UnresolvedDelta removeBaseById(final AnyObjectId id){
    final DeltaChain d = baseById.get(id);
    return d != null ? d.remove() : null;
  }

  private static UnresolvedDelta reverse(UnresolvedDelta c) {
    UnresolvedDelta tail = null;
    while (c != null) {
      final UnresolvedDelta n = c.next;
      c.next = tail;
      tail = c;
      c = n;
    }
    return tail;
  }

  private void resolveChildDeltas(final long pos, int type, byte[] data,
      PackedObjectInfo oe) throws IOException {
    UnresolvedDelta a = reverse(removeBaseById(oe));
    UnresolvedDelta b = reverse(baseByPos.remove(pos));
    while (a != null && b != null) {
      if (a.position < b.position) {
        resolveDeltas(a.position, a.crc, type, data, null);
        a = a.next;
      } else {
        resolveDeltas(b.position, b.crc, type, data, null);
        b = b.next;
      }
    }
    resolveChildDeltaChain(type, data, a);
    resolveChildDeltaChain(type, data, b);
  }

  private void resolveChildDeltaChain(final int type, final byte[] data,
      UnresolvedDelta a) throws IOException {
    while (a != null) {
      resolveDeltas(a.position, a.crc, type, data, null);
      a = a.next;
    }
  }

  private void fixThinPack(final ProgressMonitor progress) throws IOException {
    growEntries();

    if (needBaseObjectIds)
      baseObjectIds = new ObjectIdSubclassMap<ObjectId>();

    packDigest.reset();
    originalEOF = packOut.length() - 20;
    final Deflater def = new Deflater(Deflater.DEFAULT_COMPRESSION, false);
    final List<DeltaChain> missing = new ArrayList<DeltaChain>(64);
    long end = originalEOF;
    for (final DeltaChain baseId : baseById) {
      if (baseId.head == null)
        continue;
      if (needBaseObjectIds)
        baseObjectIds.add(baseId);
      final ObjectLoader ldr;
      try {
        ldr = readCurs.open(baseId);
      } catch (MissingObjectException notFound) {
        missing.add(baseId);
        continue;
      }
      final byte[] data = ldr.getCachedBytes();
      final int typeCode = ldr.getType();
      final PackedObjectInfo oe;

      crc.reset();
      packOut.seek(end);
      writeWhole(def, typeCode, data);
      oe = new PackedObjectInfo(end, (int) crc.getValue(), baseId);
      entries[entryCount++] = oe;
      end = packOut.getFilePointer();

      resolveChildDeltas(oe.getOffset(), typeCode, data, oe);
      if (progress.isCancelled())
        throw new IOException(JGitText.get().downloadCancelledDuringIndexing);
    }
    def.end();

    for (final DeltaChain base : missing) {
      if (base.head != null)
        throw new MissingObjectException(base, "delta base");
    }

    if (end - originalEOF < 20) {
      // Ugly corner case; if what we appended on to complete deltas
      // doesn't completely cover the SHA-1 we have to truncate off
      // we need to shorten the file, otherwise we will include part
      // of the old footer as object content.
      packOut.setLength(end);
    }

    fixHeaderFooter(packcsum, packDigest.digest());
  }

  private void writeWhole(final Deflater def, final int typeCode,
      final byte[] data) throws IOException {
    int sz = data.length;
    int hdrlen = 0;
    buf[hdrlen++] = (byte) ((typeCode << 4) | sz & 15);
    sz >>>= 4;
    while (sz > 0) {
      buf[hdrlen - 1] |= 0x80;
      buf[hdrlen++] = (byte) (sz & 0x7f);
      sz >>>= 7;
    }
    packDigest.update(buf, 0, hdrlen);
    crc.update(buf, 0, hdrlen);
    packOut.write(buf, 0, hdrlen);
    def.reset();
    def.setInput(data);
    def.finish();
    while (!def.finished()) {
      final int datlen = def.deflate(buf);
      packDigest.update(buf, 0, datlen);
      crc.update(buf, 0, datlen);
      packOut.write(buf, 0, datlen);
    }
  }

  private void fixHeaderFooter(final byte[] origcsum, final byte[] tailcsum)
      throws IOException {
    final MessageDigest origDigest = Constants.newMessageDigest();
    final MessageDigest tailDigest = Constants.newMessageDigest();
    long origRemaining = originalEOF;

    packOut.seek(0);
    bAvail = 0;
    bOffset = 0;
    fill(Source.FILE, 12);

    {
      final int origCnt = (int) Math.min(bAvail, origRemaining);
      origDigest.update(buf, 0, origCnt);
      origRemaining -= origCnt;
      if (origRemaining == 0)
        tailDigest.update(buf, origCnt, bAvail - origCnt);
    }

    NB.encodeInt32(buf, 8, entryCount);
    packOut.seek(0);
    packOut.write(buf, 0, 12);
    packOut.seek(bAvail);

    packDigest.reset();
    packDigest.update(buf, 0, bAvail);
    for (;;) {
      final int n = packOut.read(buf);
      if (n < 0)
        break;
      if (origRemaining != 0) {
        final int origCnt = (int) Math.min(n, origRemaining);
        origDigest.update(buf, 0, origCnt);
        origRemaining -= origCnt;
        if (origRemaining == 0)
          tailDigest.update(buf, origCnt, n - origCnt);
      } else
        tailDigest.update(buf, 0, n);

      packDigest.update(buf, 0, n);
    }

    if (!Arrays.equals(origDigest.digest(), origcsum)
        || !Arrays.equals(tailDigest.digest(), tailcsum))
      throw new IOException(JGitText.get().packCorruptedWhileWritingToFilesystem);

    packcsum = packDigest.digest();
    packOut.write(packcsum);
  }

  private void growEntries() {
    final PackedObjectInfo[] ne;

    ne = new PackedObjectInfo[(int) objectCount + baseById.size()];
    System.arraycopy(entries, 0, ne, 0, entryCount);
    entries = ne;
  }

  private void writeIdx() throws IOException {
    Arrays.sort(entries, 0, entryCount);
    List<PackedObjectInfo> list = Arrays.asList(entries);
    if (entryCount < entries.length)
      list = list.subList(0, entryCount);

    final FileOutputStream os = new FileOutputStream(dstIdx);
    try {
      final PackIndexWriter iw;
      if (outputVersion <= 0)
        iw = PackIndexWriter.createOldestPossible(os, list);
      else
        iw = PackIndexWriter.createVersion(os, outputVersion);
      iw.write(list, packcsum);
      os.getChannel().force(true);
    } finally {
      os.close();
    }
  }

  private void readPackHeader() throws IOException {
    final int hdrln = Constants.PACK_SIGNATURE.length + 4 + 4;
    final int p = fill(Source.INPUT, hdrln);
    for (int k = 0; k < Constants.PACK_SIGNATURE.length; k++)
      if (buf[p + k] != Constants.PACK_SIGNATURE[k])
        throw new IOException(JGitText.get().notAPACKFile);

    final long vers = NB.decodeUInt32(buf, p + 4);
    if (vers != 2 && vers != 3)
      throw new IOException(MessageFormat.format(JGitText.get().unsupportedPackVersion, vers));
    objectCount = NB.decodeUInt32(buf, p + 8);
    use(hdrln);
  }

  private void readPackFooter() throws IOException {
    sync();
    final byte[] cmpcsum = packDigest.digest();
    final int c = fill(Source.INPUT, 20);
    packcsum = new byte[20];
    System.arraycopy(buf, c, packcsum, 0, 20);
    use(20);
    if (packOut != null)
      packOut.write(packcsum);

    if (!Arrays.equals(cmpcsum, packcsum))
      throw new CorruptObjectException(JGitText.get().corruptObjectPackfileChecksumIncorrect);
  }

  // Cleanup all resources associated with our input parsing.
  private void endInput() {
    in = null;
    skipBuffer = null;
  }

  // Read one entire object or delta from the input.
  private void indexOneObject() throws IOException {
    final long pos = position();

    crc.reset();
    int c = readFrom(Source.INPUT);
    final int typeCode = (c >> 4) & 7;
    long sz = c & 15;
    int shift = 4;
    while ((c & 0x80) != 0) {
      c = readFrom(Source.INPUT);
      sz += (c & 0x7f) << shift;
      shift += 7;
    }

    switch (typeCode) {
    case Constants.OBJ_COMMIT:
    case Constants.OBJ_TREE:
    case Constants.OBJ_BLOB:
    case Constants.OBJ_TAG:
      whole(typeCode, pos, sz);
      break;
    case Constants.OBJ_OFS_DELTA: {
      c = readFrom(Source.INPUT);
      long ofs = c & 127;
      while ((c & 128) != 0) {
        ofs += 1;
        c = readFrom(Source.INPUT);
        ofs <<= 7;
        ofs += (c & 127);
      }
      final long base = pos - ofs;
      final UnresolvedDelta n;
      inflateAndSkip(Source.INPUT, sz);
      n = new UnresolvedDelta(pos, (int) crc.getValue());
      n.next = baseByPos.put(base, n);
      deltaCount++;
      break;
    }
    case Constants.OBJ_REF_DELTA: {
      c = fill(Source.INPUT, 20);
      crc.update(buf, c, 20);
      final ObjectId base = ObjectId.fromRaw(buf, c);
      use(20);
      DeltaChain r = baseById.get(base);
      if (r == null) {
        r = new DeltaChain(base);
        baseById.add(r);
      }
      inflateAndSkip(Source.INPUT, sz);
      r.add(new UnresolvedDelta(pos, (int) crc.getValue()));
      deltaCount++;
      break;
    }
    default:
      throw new IOException(MessageFormat.format(JGitText.get().unknownObjectType, typeCode));
    }
  }

  private void whole(final int type, final long pos, final long sz)
      throws IOException {
    final byte[] data = inflateAndReturn(Source.INPUT, sz);
    objectDigest.update(Constants.encodedTypeString(type));
    objectDigest.update((byte) ' ');
    objectDigest.update(Constants.encodeASCII(sz));
    objectDigest.update((byte) 0);
    objectDigest.update(data);
    tempObjectId.fromRaw(objectDigest.digest(), 0);

    verifySafeObject(tempObjectId, type, data);
    final int crc32 = (int) crc.getValue();
    addObjectAndTrack(new PackedObjectInfo(pos, crc32, tempObjectId));
  }

  private void verifySafeObject(final AnyObjectId id, final int type,
      final byte[] data) throws IOException {
    if (objCheck != null) {
      try {
        objCheck.check(type, data);
      } catch (CorruptObjectException e) {
        throw new IOException(MessageFormat.format(JGitText.get().invalidObject
            , Constants.typeString(type) , id.name() , e.getMessage()));
      }
    }

    try {
      final ObjectLoader ldr = readCurs.open(id, type);
      final byte[] existingData = ldr.getCachedBytes();
      if (!Arrays.equals(data, existingData)) {
        throw new IOException(MessageFormat.format(JGitText.get().collisionOn, id.name()));
      }
    } catch (MissingObjectException notLocal) {
      // This is OK, we don't have a copy of the object locally
      // but the API throws when we try to read it as usually its
      // an error to read something that doesn't exist.
    }
  }

  // Current position of {@link #bOffset} within the entire file.
  private long position() {
    return bBase + bOffset;
  }

  private void position(final long pos) throws IOException {
    packOut.seek(pos);
    bBase = pos;
    bOffset = 0;
    bAvail = 0;
  }

  // Consume exactly one byte from the buffer and return it.
  private int readFrom(final Source src) throws IOException {
    if (bAvail == 0)
      fill(src, 1);
    bAvail--;
    final int b = buf[bOffset++] & 0xff;
    crc.update(b);
    return b;
  }

  // Consume cnt bytes from the buffer.
  private void use(final int cnt) {
    bOffset += cnt;
    bAvail -= cnt;
  }

  // Ensure at least need bytes are available in in {@link #buf}.
  private int fill(final Source src, final int need) throws IOException {
    while (bAvail < need) {
      int next = bOffset + bAvail;
      int free = buf.length - next;
      if (free + bAvail < need) {
        switch(src){
        case INPUT:
          sync();
          break;
        case FILE:
          if (bAvail > 0)
            System.arraycopy(buf, bOffset, buf, 0, bAvail);
          bOffset = 0;
          break;
        }
        next = bAvail;
        free = buf.length - next;
      }
      switch(src){
      case INPUT:
        next = in.read(buf, next, free);
        break;
      case FILE:
        next = packOut.read(buf, next, free);
        break;
      }
      if (next <= 0)
        throw new EOFException(JGitText.get().packfileIsTruncated);
      bAvail += next;
    }
    return bOffset;
  }

  // Store consumed bytes in {@link #buf} up to {@link #bOffset}.
  private void sync() throws IOException {
    packDigest.update(buf, 0, bOffset);
    if (packOut != null)
      packOut.write(buf, 0, bOffset);
    if (bAvail > 0)
      System.arraycopy(buf, bOffset, buf, 0, bAvail);
    bBase += bOffset;
    bOffset = 0;
  }

  private void inflateAndSkip(final Source src, final long inflatedSize)
      throws IOException {
    inflate(src, inflatedSize, skipBuffer, false /* do not keep result */);
  }

  private byte[] inflateAndReturn(final Source src, final long inflatedSize)
      throws IOException {
    final byte[] dst = new byte[(int) inflatedSize];
    inflate(src, inflatedSize, dst, true /* keep result in dst */);
    return dst;
  }

  private void inflate(final Source src, final long inflatedSize,
      final byte[] dst, final boolean keep) throws IOException {
    final Inflater inf = inflater;
    try {
      int off = 0;
      long cnt = 0;
      int p = fill(src, 24);
      inf.setInput(buf, p, bAvail);

      for (;;) {
        int r = inf.inflate(dst, off, dst.length - off);
        if (r == 0) {
          if (inf.finished())
            break;
          if (inf.needsInput()) {
            if (p >= 0) {
              crc.update(buf, p, bAvail);
              use(bAvail);
            }
            p = fill(src, 24);
            inf.setInput(buf, p, bAvail);
          } else {
            throw new CorruptObjectException(MessageFormat.format(
                JGitText.get().packfileCorruptionDetected,
                JGitText.get().unknownZlibError));
          }
        }
        cnt += r;
        if (keep)
          off += r;
      }

      if (cnt != inflatedSize) {
        throw new CorruptObjectException(MessageFormat.format(JGitText
            .get().packfileCorruptionDetected,
            JGitText.get().wrongDecompressedLength));
      }

      int left = bAvail - inf.getRemaining();
      if (left > 0) {
        crc.update(buf, p, left);
        use(left);
      }
    } catch (DataFormatException dfe) {
      throw new CorruptObjectException(MessageFormat.format(JGitText
          .get().packfileCorruptionDetected, dfe.getMessage()));
    } finally {
      inf.reset();
    }
  }

  private static class DeltaChain extends ObjectId {
    UnresolvedDelta head;

    DeltaChain(final AnyObjectId id) {
      super(id);
    }

    UnresolvedDelta remove() {
      final UnresolvedDelta r = head;
      if (r != null)
        head = null;
      return r;
    }

    void add(final UnresolvedDelta d) {
      d.next = head;
      head = d;
    }
  }

  private static class UnresolvedDelta {
    final long position;

    final int crc;

    UnresolvedDelta next;

    UnresolvedDelta(final long headerOffset, final int crc32) {
      position = headerOffset;
      crc = crc32;
    }
  }

  /**
   * Rename the pack to it's final name and location and open it.
   * <p>
   * If the call completes successfully the repository this IndexPack instance
   * was created with will have the objects in the pack available for reading
   * and use, without needing to scan for packs.
   *
   * @throws IOException
   *             The pack could not be inserted into the repository's objects
   *             directory. The pack no longer exists on disk, as it was
   *             removed prior to throwing the exception to the caller.
   */
  public void renameAndOpenPack() throws IOException {
    renameAndOpenPack(null);
  }

  /**
   * Rename the pack to it's final name and location and open it.
   * <p>
   * If the call completes successfully the repository this IndexPack instance
   * was created with will have the objects in the pack available for reading
   * and use, without needing to scan for packs.
   *
   * @param lockMessage
   *            message to place in the pack-*.keep file. If null, no lock
   *            will be created, and this method returns null.
   * @return the pack lock object, if lockMessage is not null.
   * @throws IOException
   *             The pack could not be inserted into the repository's objects
   *             directory. The pack no longer exists on disk, as it was
   *             removed prior to throwing the exception to the caller.
   */
  public PackLock renameAndOpenPack(final String lockMessage)
      throws IOException {
    if (!keepEmpty && entryCount == 0) {
      cleanupTemporaryFiles();
      return null;
    }

    final MessageDigest d = Constants.newMessageDigest();
    final byte[] oeBytes = new byte[Constants.OBJECT_ID_LENGTH];
    for (int i = 0; i < entryCount; i++) {
      final PackedObjectInfo oe = entries[i];
      oe.copyRawTo(oeBytes, 0);
      d.update(oeBytes);
    }

    final String name = ObjectId.fromRaw(d.digest()).name();
    final File packDir = new File(repo.getObjectsDirectory(), "pack");
    final File finalPack = new File(packDir, "pack-" + name + ".pack");
    final File finalIdx = new File(packDir, "pack-" + name + ".idx");
    final PackLock keep = new PackLock(finalPack, repo.getFS());

    if (!packDir.exists() && !packDir.mkdir() && !packDir.exists()) {
      // The objects/pack directory isn't present, and we are unable
      // to create it. There is no way to move this pack in.
      //
      cleanupTemporaryFiles();
      throw new IOException(MessageFormat.format(JGitText.get().cannotCreateDirectory, packDir.getAbsolutePath()));
    }

    if (finalPack.exists()) {
      // If the pack is already present we should never replace it.
      //
      cleanupTemporaryFiles();
      return null;
    }

    if (lockMessage != null) {
      // If we have a reason to create a keep file for this pack, do
      // so, or fail fast and don't put the pack in place.
      //
      try {
        if (!keep.lock(lockMessage))
          throw new IOException(MessageFormat.format(JGitText.get().cannotLockPackIn, finalPack));
      } catch (IOException e) {
        cleanupTemporaryFiles();
        throw e;
      }
    }

    if (!dstPack.renameTo(finalPack)) {
      cleanupTemporaryFiles();
      keep.unlock();
      throw new IOException(MessageFormat.format(JGitText.get().cannotMovePackTo, finalPack));
    }

    if (!dstIdx.renameTo(finalIdx)) {
      cleanupTemporaryFiles();
      keep.unlock();
      if (!finalPack.delete())
        finalPack.deleteOnExit();
      throw new IOException(MessageFormat.format(JGitText.get().cannotMoveIndexTo, finalIdx));
    }

    try {
      repo.openPack(finalPack, finalIdx);
    } catch (IOException err) {
      keep.unlock();
      finalPack.delete();
      finalIdx.delete();
      throw err;
    }

    return lockMessage != null ? keep : null;
  }

  private void cleanupTemporaryFiles() {
    if (!dstIdx.delete())
      dstIdx.deleteOnExit();
    if (!dstPack.delete())
      dstPack.deleteOnExit();
  }

  private void addObjectAndTrack(PackedObjectInfo oe) {
    entries[entryCount++] = oe;
    if (needNewObjectIds())
      newObjectIds.add(oe);
  }
}
TOP

Related Classes of org.eclipse.jgit.transport.IndexPack

TOP
Copyright © 2018 www.massapi.com. All rights reserved.
All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.