All Downloads are FREE. Search and download functionalities are using the official Maven repository.

com.google.gerrit.server.update.RepoView Maven / Gradle / Ivy

// Copyright (C) 2017 The Android Open Source Project
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.

package com.google.gerrit.server.update;

import static com.google.common.base.Preconditions.checkArgument;
import static com.google.common.base.Preconditions.checkNotNull;

import com.google.common.collect.Maps;
import com.google.gerrit.reviewdb.client.Project;
import com.google.gerrit.server.git.GitRepositoryManager;
import java.io.IOException;
import java.util.HashMap;
import java.util.Map;
import java.util.Optional;
import org.eclipse.jgit.lib.Config;
import org.eclipse.jgit.lib.ObjectId;
import org.eclipse.jgit.lib.ObjectInserter;
import org.eclipse.jgit.lib.Ref;
import org.eclipse.jgit.lib.Repository;
import org.eclipse.jgit.revwalk.RevWalk;

/**
 * Restricted view of a {@link Repository} for use by {@link BatchUpdateOp} implementations.
 *
 * 

This class serves two purposes in the context of {@link BatchUpdate}. First, the subset of * normal Repository functionality is purely read-only, which prevents implementors from modifying * the repository outside of {@link BatchUpdateOp#updateRepo}. Write operations can only be * performed by calling methods on {@link RepoContext}. * *

Second, the read methods take into account any pending operations on the repository that * implementations have staged using the write methods on {@link RepoContext}. Callers do not have * to worry about whether operations have been performed yet, and the implementation details may * differ between ReviewDb and NoteDb, but callers just don't need to care. */ public class RepoView { private final Repository repo; private final RevWalk rw; private final ObjectInserter inserter; private final ObjectInserter inserterWrapper; private final ChainedReceiveCommands commands; private final boolean closeRepo; RepoView(GitRepositoryManager repoManager, Project.NameKey project) throws IOException { repo = repoManager.openRepository(project); inserter = repo.newObjectInserter(); inserterWrapper = new NonFlushingInserter(inserter); rw = new RevWalk(inserter.newReader()); commands = new ChainedReceiveCommands(repo); closeRepo = true; } RepoView(Repository repo, RevWalk rw, ObjectInserter inserter) { checkArgument( rw.getObjectReader().getCreatedFromInserter() == inserter, "expected RevWalk %s to be created by ObjectInserter %s", rw, inserter); this.repo = checkNotNull(repo); this.rw = checkNotNull(rw); this.inserter = checkNotNull(inserter); inserterWrapper = new NonFlushingInserter(inserter); commands = new ChainedReceiveCommands(repo); closeRepo = false; } /** * Get this repo's configuration. * *

This is the storage-level config you would get with {@link Repository#getConfig()}, not, for * example, the Gerrit-level project config. * * @return a defensive copy of the config; modifications have no effect on the underlying config. */ public Config getConfig() { return new Config(repo.getConfig()); } /** * Get an open revwalk on the repo. * *

Guaranteed to be able to read back any objects inserted in the repository via {@link * RepoContext#getInserter()}, even if objects have not been flushed to the underlying repo. In * particular this includes any object returned by {@link #getRef(String)}, even taking into * account not-yet-executed commands. * * @return revwalk. */ public RevWalk getRevWalk() { return rw; } /** * Read a single ref from the repo. * *

Takes into account any ref update commands added during the course of the update using * {@link RepoContext#addRefUpdate}, even if they have not yet been executed on the underlying * repo. * *

The results of individual ref lookups are cached: calling this method multiple times with * the same ref name will return the same result (unless a command was added in the meantime). The * repo is not reread. * * @param name exact ref name. * @return the value of the ref, if present. * @throws IOException if an error occurred. */ public Optional getRef(String name) throws IOException { return getCommands().get(name); } /** * Look up refs by prefix. * *

Takes into account any ref update commands added during the course of the update using * {@link RepoContext#addRefUpdate}, even if they have not yet been executed on the underlying * repo. * *

For any ref that has previously been accessed with {@link #getRef(String)}, the value in the * result map will be that same cached value. Any refs that have not been previously * accessed are re-scanned from the repo on each call. * * @param prefix ref prefix; must end in '/' or else be empty. * @return a map of ref suffixes to SHA-1s. The refs are all under {@code prefix} and have the * prefix stripped; this matches the behavior of {@link * org.eclipse.jgit.lib.RefDatabase#getRefs(String)}. * @throws IOException if an error occurred. */ public Map getRefs(String prefix) throws IOException { Map result = new HashMap<>( Maps.transformValues(repo.getRefDatabase().getRefs(prefix), Ref::getObjectId)); // First, overwrite any cached reads from the underlying RepoRefCache. If any of these differ, // it's because a ref was updated after the RepoRefCache read it. It feels a little odd to // prefer the *old* value in this case, but it would be weirder to be inconsistent with getRef. // // Mostly this doesn't matter. If the caller was intending to write to the ref, they lost a // race, and they will get a lock failure. If they just want to read, well, the JGit interface // doesn't currently guarantee that any snapshot of multiple refs is consistent, so they were // probably out of luck anyway. commands .getRepoRefCache() .getCachedRefs() .forEach((k, v) -> updateRefIfPrefixMatches(result, prefix, k, v)); // Second, overwrite with any pending commands. commands .getCommands() .values() .forEach( c -> updateRefIfPrefixMatches(result, prefix, c.getRefName(), toOptional(c.getNewId()))); return result; } private static Optional toOptional(ObjectId id) { return id.equals(ObjectId.zeroId()) ? Optional.empty() : Optional.of(id); } private static void updateRefIfPrefixMatches( Map map, String prefix, String fullRefName, Optional maybeId) { if (!fullRefName.startsWith(prefix)) { return; } String suffix = fullRefName.substring(prefix.length()); if (maybeId.isPresent()) { map.put(suffix, maybeId.get()); } else { map.remove(suffix); } } // Not AutoCloseable so callers can't improperly close it. Plus it's never managed with a try // block anyway. void close() { if (closeRepo) { inserter.close(); rw.close(); repo.close(); } } Repository getRepository() { return repo; } ObjectInserter getInserter() { return inserter; } ObjectInserter getInserterWrapper() { return inserterWrapper; } ChainedReceiveCommands getCommands() { return commands; } private static class NonFlushingInserter extends ObjectInserter.Filter { private final ObjectInserter delegate; private NonFlushingInserter(ObjectInserter delegate) { this.delegate = delegate; } @Override protected ObjectInserter delegate() { return delegate; } @Override public void flush() { // Do nothing. } @Override public void close() { // Do nothing; the delegate is closed separately. } } }





© 2015 - 2024 Weber Informatics LLC | Privacy Policy