Examples of org.sonatype.nexus.proxy.walker.ParentOMatic

org.sonatype.nexus.proxy.walker.ParentOMatic
A helper class to "optimize" when some sort of "gathering for later processing" (most notable collection of directories to start Walks from) of repository paths happens, that are to be processed in some subsequent step by "walking" them (recurse them, a la "visitor" pattern). This utility class simply maintains a "tree" with marked nodes, while the meaning of "marked" is left to class user. One assumption does exists against "marked": some recursive processing is planned against it. It currently applies these simple rules to it.
- rule A: if parent node of the currently added node is already "marked", do not mark the currently added node (reason: it will be anyway processes when marked parent processing starts to recurse)
- rule B: if all the children nodes of the currently added node's parent as "marked", mark the parent and unmark it's children (reason: it's usually cheaper to fire off one walk level above, from parent, instead firing, for example 100, independent walks from children one by one)
Note: all the input paths are expected to be "normalized ones": being absolute, using generic "/" character as path separator (since these are NOT File paths, but just hierarchical paths of strings). For example: {@link RepositoryItemUid#getPath()} returns paths like these. See {@link RepositoryItemUid} for explanation.
This class also "optimizes" the tree size to lessen memory use. This "optimization" can be turned off, see constructors.
This class makes use of {@link Node} to implement the tree hierarchy. @author cstamas @since 2.0


  // ==


  protected Node<Payload> buildRoot(final List<String> entries, final int maxDepth) {
    // no rule B!
    final ParentOMatic parentOMatic = new ParentOMatic(true, true, false);
    for (String entry : entries) {
      parentOMatic.addAndMarkPath(entry);
    }
    if (maxDepth != Integer.MAX_VALUE) {
      // cut the tree to maxDepth
      parentOMatic.cutNodesDeeperThan(maxDepth);
    }
    return parentOMatic.getRoot();
  }

View Full Code Here

    final PrefixCollectorProcessor prefixCollectorProcessor = new PrefixCollectorProcessor();
    context.getProcessors().add(prefixCollectorProcessor);


    try {
      walker.walk(context);
      final ParentOMatic parentOMatic = prefixCollectorProcessor.getParentOMatic();
      if (parentOMatic.getRoot().isLeaf()) {
        // tree is basically empty, so make the list too
        discoveryResult.recordSuccess(ID, "Repository crawled successfully (is empty)",
            new ArrayListPrefixSource(Collections.<String>emptyList()));
      }
      else {

View Full Code Here

      extends AbstractWalkerProcessor
  {
    private final ParentOMatic parentOMatic;


    public PrefixCollectorProcessor() {
      this.parentOMatic = new ParentOMatic();
    }

View Full Code Here

          cacheAltered = altered1 || altered2;
        }
      }
    }
    else {
      final ParentOMatic parentOMatic = new ParentOMatic(false);
      final DefaultWalkerContext context = new DefaultWalkerContext(this, request);
      final Collection<String> nfcPaths = getNotFoundCache().listKeysInCache();


      for (String nfcPath : nfcPaths) {
        final DefaultStorageNotFoundItem nfcItem =
            new DefaultStorageNotFoundItem(this, new ResourceStoreRequest(nfcPath));


        if (filter.shouldProcess(context, nfcItem)) {
          parentOMatic.addAndMarkPath(nfcPath);
        }
      }


      for (String path : parentOMatic.getMarkedPaths()) {
        boolean removed = getNotFoundCache().remove(path);
        cacheAltered = cacheAltered || removed;
      }
    }

View Full Code Here

  @Override
  protected List<String> diveIn(final ScrapeContext context, final Page page)
      throws IOException
  {
    // we use the great and all-mighty ParentOMatic
    final ParentOMatic parentOMatic = new ParentOMatic();
    diveIn(context, page, 0, parentOMatic, parentOMatic.getRoot());
    // Special case: scraped with 0 entry, we consider this as an error
    // Remote repo empty? Why are you proxying it? Or worse, some scrape
    // exotic index page and we end up with 0 entries by mistake?
    if (parentOMatic.getRoot().isLeaf()) {
      context.stop("Remote recognized as " + getTargetedServer()
          + ", but scraped 0 entries. This is considered a failure.");
      return null;
    }
    final List<String> entries = parentOMatic.getAllLeafPaths();
    return entries;
  }

View Full Code Here

          "Collecting deletable snapshots on repository " + repository.getId() + " from storage directory "
              + repository.getLocalUrl()
      );
    }


    final ParentOMatic parentOMatic = new ParentOMatic();


    // create a walker to collect deletables and let it loose on collections only
    final SnapshotRemoverWalkerProcessor snapshotRemoveProcessor =
        new SnapshotRemoverWalkerProcessor(repository, request, parentOMatic);


    final DefaultWalkerContext ctxMain =
        new DefaultWalkerContext(repository,
            new ResourceStoreRequest("/"),
            new DottedStoreWalkerFilter(),
            TraversalType.BREADTH_FIRST,
            false);
    ctxMain.getContext().put(DeleteOperation.DELETE_OPERATION_CTX_KEY, getDeleteOperation(request));
    ctxMain.getProcessors().add(snapshotRemoveProcessor);
    walker.walk(ctxMain);


    if (ctxMain.getStopCause() != null) {
      result.setSuccessful(false);
    }


    // and collect results
    result.setDeletedSnapshots(snapshotRemoveProcessor.getDeletedSnapshots());
    result.setDeletedFiles(snapshotRemoveProcessor.getDeletedFiles());


    if (log.isDebugEnabled()) {
      log.debug(
          "Collected and deleted " + snapshotRemoveProcessor.getDeletedSnapshots()
              + " snapshots with alltogether " + snapshotRemoveProcessor.getDeletedFiles()
              + " files on repository " + repository.getId()
      );
    }


    // if we are processing a hosted-snapshot repository, we need to rebuild maven metadata
    // without this if below, the walk would happen against proxy repositories too, but doing nothing!
    if (repository.getRepositoryKind().isFacetAvailable(HostedRepository.class)) {
      // expire NFC since we might create new maven metadata files
      repository.expireNotFoundCaches(new ResourceStoreRequest(RepositoryItemUid.PATH_ROOT));


      RecreateMavenMetadataWalkerProcessor metadataRebuildProcessor =
          new RecreateMavenMetadataWalkerProcessor(log, getDeleteOperation(request));


      for (String path : parentOMatic.getMarkedPaths()) {
        TaskUtil.checkInterruption();


        DefaultWalkerContext ctxMd =
            new DefaultWalkerContext(repository, new ResourceStoreRequest(path),
                new DottedStoreWalkerFilter());

View Full Code Here

TOP

Related Classes of org.sonatype.nexus.proxy.walker.ParentOMatic

org.sonatype.nexus.maven.tasks.DefaultSnapshotRemover

org.sonatype.nexus.proxy.maven.routing.internal.LocalContentDiscovererImpl

org.sonatype.nexus.proxy.maven.routing.internal.LocalContentDiscovererImpl$PrefixCollectorProcessor

org.sonatype.nexus.proxy.maven.routing.internal.PathMatcher

org.sonatype.nexus.proxy.maven.routing.internal.scrape.AbstractGeneratedIndexPageScraper

org.sonatype.nexus.proxy.repository.AbstractRepository

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.