You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

Store Git on any DHT jgit.storage.dht is a storage provider implementation for JGit that permits storing the Git repository in a distributed hashtable, NoSQL system, or other database. The actual underlying storage system is undefined, and can be plugged in by implementing 7 small interfaces: * Database * RepositoryIndexTable * RepositoryTable * RefTable * ChunkTable * ObjectIndexTable * WriteBuffer The storage provider interface tries to assume very little about the underlying storage system, and requires only three key features: * key -> value lookup (a hashtable is suitable) * atomic updates on single rows * asynchronous operations (Java's ExecutorService is easy to use) Most NoSQL database products offer all 3 of these features in their clients, and so does any decent network based cache system like the open source memcache product. Relying only on key equality for data retrevial makes it simple for the storage engine to distribute across multiple machines. Traditional SQL systems could also be used with a JDBC based spi implementation. Before submitting this change I have implemented six storage systems for the spi layer: * Apache HBase[1] * Apache Cassandra[2] * Google Bigtable[3] * an in-memory implementation for unit testing * a JDBC implementation for SQL * a generic cache provider that can ride on top of memcache All six systems came in with an spi layer around 1000 lines of code to implement the above 7 interfaces. This is a huge reduction in size compared to prior attempts to implement a new JGit storage layer. As this package shows, a complete JGit storage implementation is more than 17,000 lines of fairly complex code. A simple cache is provided in storage.dht.spi.cache. Implementers can use CacheDatabase to wrap any other type of Database and perform fast reads against a network based cache service, such as the open source memcached[4]. An implementation of CacheService must be provided to glue this spi onto the network cache. [1] https://github.com/spearce/jgit_hbase [2] https://github.com/spearce/jgit_cassandra [3] http://labs.google.com/papers/bigtable.html [4] http://memcached.org/ Change-Id: I0aa4072781f5ccc019ca421c036adff2c40c4295 Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
13 年之前
Store Git on any DHT jgit.storage.dht is a storage provider implementation for JGit that permits storing the Git repository in a distributed hashtable, NoSQL system, or other database. The actual underlying storage system is undefined, and can be plugged in by implementing 7 small interfaces: * Database * RepositoryIndexTable * RepositoryTable * RefTable * ChunkTable * ObjectIndexTable * WriteBuffer The storage provider interface tries to assume very little about the underlying storage system, and requires only three key features: * key -> value lookup (a hashtable is suitable) * atomic updates on single rows * asynchronous operations (Java's ExecutorService is easy to use) Most NoSQL database products offer all 3 of these features in their clients, and so does any decent network based cache system like the open source memcache product. Relying only on key equality for data retrevial makes it simple for the storage engine to distribute across multiple machines. Traditional SQL systems could also be used with a JDBC based spi implementation. Before submitting this change I have implemented six storage systems for the spi layer: * Apache HBase[1] * Apache Cassandra[2] * Google Bigtable[3] * an in-memory implementation for unit testing * a JDBC implementation for SQL * a generic cache provider that can ride on top of memcache All six systems came in with an spi layer around 1000 lines of code to implement the above 7 interfaces. This is a huge reduction in size compared to prior attempts to implement a new JGit storage layer. As this package shows, a complete JGit storage implementation is more than 17,000 lines of fairly complex code. A simple cache is provided in storage.dht.spi.cache. Implementers can use CacheDatabase to wrap any other type of Database and perform fast reads against a network based cache service, such as the open source memcached[4]. An implementation of CacheService must be provided to glue this spi onto the network cache. [1] https://github.com/spearce/jgit_hbase [2] https://github.com/spearce/jgit_cassandra [3] http://labs.google.com/papers/bigtable.html [4] http://memcached.org/ Change-Id: I0aa4072781f5ccc019ca421c036adff2c40c4295 Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
13 年之前
Store Git on any DHT jgit.storage.dht is a storage provider implementation for JGit that permits storing the Git repository in a distributed hashtable, NoSQL system, or other database. The actual underlying storage system is undefined, and can be plugged in by implementing 7 small interfaces: * Database * RepositoryIndexTable * RepositoryTable * RefTable * ChunkTable * ObjectIndexTable * WriteBuffer The storage provider interface tries to assume very little about the underlying storage system, and requires only three key features: * key -> value lookup (a hashtable is suitable) * atomic updates on single rows * asynchronous operations (Java's ExecutorService is easy to use) Most NoSQL database products offer all 3 of these features in their clients, and so does any decent network based cache system like the open source memcache product. Relying only on key equality for data retrevial makes it simple for the storage engine to distribute across multiple machines. Traditional SQL systems could also be used with a JDBC based spi implementation. Before submitting this change I have implemented six storage systems for the spi layer: * Apache HBase[1] * Apache Cassandra[2] * Google Bigtable[3] * an in-memory implementation for unit testing * a JDBC implementation for SQL * a generic cache provider that can ride on top of memcache All six systems came in with an spi layer around 1000 lines of code to implement the above 7 interfaces. This is a huge reduction in size compared to prior attempts to implement a new JGit storage layer. As this package shows, a complete JGit storage implementation is more than 17,000 lines of fairly complex code. A simple cache is provided in storage.dht.spi.cache. Implementers can use CacheDatabase to wrap any other type of Database and perform fast reads against a network based cache service, such as the open source memcached[4]. An implementation of CacheService must be provided to glue this spi onto the network cache. [1] https://github.com/spearce/jgit_hbase [2] https://github.com/spearce/jgit_cassandra [3] http://labs.google.com/papers/bigtable.html [4] http://memcached.org/ Change-Id: I0aa4072781f5ccc019ca421c036adff2c40c4295 Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
13 年之前
Store Git on any DHT jgit.storage.dht is a storage provider implementation for JGit that permits storing the Git repository in a distributed hashtable, NoSQL system, or other database. The actual underlying storage system is undefined, and can be plugged in by implementing 7 small interfaces: * Database * RepositoryIndexTable * RepositoryTable * RefTable * ChunkTable * ObjectIndexTable * WriteBuffer The storage provider interface tries to assume very little about the underlying storage system, and requires only three key features: * key -> value lookup (a hashtable is suitable) * atomic updates on single rows * asynchronous operations (Java's ExecutorService is easy to use) Most NoSQL database products offer all 3 of these features in their clients, and so does any decent network based cache system like the open source memcache product. Relying only on key equality for data retrevial makes it simple for the storage engine to distribute across multiple machines. Traditional SQL systems could also be used with a JDBC based spi implementation. Before submitting this change I have implemented six storage systems for the spi layer: * Apache HBase[1] * Apache Cassandra[2] * Google Bigtable[3] * an in-memory implementation for unit testing * a JDBC implementation for SQL * a generic cache provider that can ride on top of memcache All six systems came in with an spi layer around 1000 lines of code to implement the above 7 interfaces. This is a huge reduction in size compared to prior attempts to implement a new JGit storage layer. As this package shows, a complete JGit storage implementation is more than 17,000 lines of fairly complex code. A simple cache is provided in storage.dht.spi.cache. Implementers can use CacheDatabase to wrap any other type of Database and perform fast reads against a network based cache service, such as the open source memcached[4]. An implementation of CacheService must be provided to glue this spi onto the network cache. [1] https://github.com/spearce/jgit_hbase [2] https://github.com/spearce/jgit_cassandra [3] http://labs.google.com/papers/bigtable.html [4] http://memcached.org/ Change-Id: I0aa4072781f5ccc019ca421c036adff2c40c4295 Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
13 年之前
Store Git on any DHT jgit.storage.dht is a storage provider implementation for JGit that permits storing the Git repository in a distributed hashtable, NoSQL system, or other database. The actual underlying storage system is undefined, and can be plugged in by implementing 7 small interfaces: * Database * RepositoryIndexTable * RepositoryTable * RefTable * ChunkTable * ObjectIndexTable * WriteBuffer The storage provider interface tries to assume very little about the underlying storage system, and requires only three key features: * key -> value lookup (a hashtable is suitable) * atomic updates on single rows * asynchronous operations (Java's ExecutorService is easy to use) Most NoSQL database products offer all 3 of these features in their clients, and so does any decent network based cache system like the open source memcache product. Relying only on key equality for data retrevial makes it simple for the storage engine to distribute across multiple machines. Traditional SQL systems could also be used with a JDBC based spi implementation. Before submitting this change I have implemented six storage systems for the spi layer: * Apache HBase[1] * Apache Cassandra[2] * Google Bigtable[3] * an in-memory implementation for unit testing * a JDBC implementation for SQL * a generic cache provider that can ride on top of memcache All six systems came in with an spi layer around 1000 lines of code to implement the above 7 interfaces. This is a huge reduction in size compared to prior attempts to implement a new JGit storage layer. As this package shows, a complete JGit storage implementation is more than 17,000 lines of fairly complex code. A simple cache is provided in storage.dht.spi.cache. Implementers can use CacheDatabase to wrap any other type of Database and perform fast reads against a network based cache service, such as the open source memcached[4]. An implementation of CacheService must be provided to glue this spi onto the network cache. [1] https://github.com/spearce/jgit_hbase [2] https://github.com/spearce/jgit_cassandra [3] http://labs.google.com/papers/bigtable.html [4] http://memcached.org/ Change-Id: I0aa4072781f5ccc019ca421c036adff2c40c4295 Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
13 年之前
Store Git on any DHT jgit.storage.dht is a storage provider implementation for JGit that permits storing the Git repository in a distributed hashtable, NoSQL system, or other database. The actual underlying storage system is undefined, and can be plugged in by implementing 7 small interfaces: * Database * RepositoryIndexTable * RepositoryTable * RefTable * ChunkTable * ObjectIndexTable * WriteBuffer The storage provider interface tries to assume very little about the underlying storage system, and requires only three key features: * key -> value lookup (a hashtable is suitable) * atomic updates on single rows * asynchronous operations (Java's ExecutorService is easy to use) Most NoSQL database products offer all 3 of these features in their clients, and so does any decent network based cache system like the open source memcache product. Relying only on key equality for data retrevial makes it simple for the storage engine to distribute across multiple machines. Traditional SQL systems could also be used with a JDBC based spi implementation. Before submitting this change I have implemented six storage systems for the spi layer: * Apache HBase[1] * Apache Cassandra[2] * Google Bigtable[3] * an in-memory implementation for unit testing * a JDBC implementation for SQL * a generic cache provider that can ride on top of memcache All six systems came in with an spi layer around 1000 lines of code to implement the above 7 interfaces. This is a huge reduction in size compared to prior attempts to implement a new JGit storage layer. As this package shows, a complete JGit storage implementation is more than 17,000 lines of fairly complex code. A simple cache is provided in storage.dht.spi.cache. Implementers can use CacheDatabase to wrap any other type of Database and perform fast reads against a network based cache service, such as the open source memcached[4]. An implementation of CacheService must be provided to glue this spi onto the network cache. [1] https://github.com/spearce/jgit_hbase [2] https://github.com/spearce/jgit_cassandra [3] http://labs.google.com/papers/bigtable.html [4] http://memcached.org/ Change-Id: I0aa4072781f5ccc019ca421c036adff2c40c4295 Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
13 年之前
Store Git on any DHT jgit.storage.dht is a storage provider implementation for JGit that permits storing the Git repository in a distributed hashtable, NoSQL system, or other database. The actual underlying storage system is undefined, and can be plugged in by implementing 7 small interfaces: * Database * RepositoryIndexTable * RepositoryTable * RefTable * ChunkTable * ObjectIndexTable * WriteBuffer The storage provider interface tries to assume very little about the underlying storage system, and requires only three key features: * key -> value lookup (a hashtable is suitable) * atomic updates on single rows * asynchronous operations (Java's ExecutorService is easy to use) Most NoSQL database products offer all 3 of these features in their clients, and so does any decent network based cache system like the open source memcache product. Relying only on key equality for data retrevial makes it simple for the storage engine to distribute across multiple machines. Traditional SQL systems could also be used with a JDBC based spi implementation. Before submitting this change I have implemented six storage systems for the spi layer: * Apache HBase[1] * Apache Cassandra[2] * Google Bigtable[3] * an in-memory implementation for unit testing * a JDBC implementation for SQL * a generic cache provider that can ride on top of memcache All six systems came in with an spi layer around 1000 lines of code to implement the above 7 interfaces. This is a huge reduction in size compared to prior attempts to implement a new JGit storage layer. As this package shows, a complete JGit storage implementation is more than 17,000 lines of fairly complex code. A simple cache is provided in storage.dht.spi.cache. Implementers can use CacheDatabase to wrap any other type of Database and perform fast reads against a network based cache service, such as the open source memcached[4]. An implementation of CacheService must be provided to glue this spi onto the network cache. [1] https://github.com/spearce/jgit_hbase [2] https://github.com/spearce/jgit_cassandra [3] http://labs.google.com/papers/bigtable.html [4] http://memcached.org/ Change-Id: I0aa4072781f5ccc019ca421c036adff2c40c4295 Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
13 年之前
Store Git on any DHT jgit.storage.dht is a storage provider implementation for JGit that permits storing the Git repository in a distributed hashtable, NoSQL system, or other database. The actual underlying storage system is undefined, and can be plugged in by implementing 7 small interfaces: * Database * RepositoryIndexTable * RepositoryTable * RefTable * ChunkTable * ObjectIndexTable * WriteBuffer The storage provider interface tries to assume very little about the underlying storage system, and requires only three key features: * key -> value lookup (a hashtable is suitable) * atomic updates on single rows * asynchronous operations (Java's ExecutorService is easy to use) Most NoSQL database products offer all 3 of these features in their clients, and so does any decent network based cache system like the open source memcache product. Relying only on key equality for data retrevial makes it simple for the storage engine to distribute across multiple machines. Traditional SQL systems could also be used with a JDBC based spi implementation. Before submitting this change I have implemented six storage systems for the spi layer: * Apache HBase[1] * Apache Cassandra[2] * Google Bigtable[3] * an in-memory implementation for unit testing * a JDBC implementation for SQL * a generic cache provider that can ride on top of memcache All six systems came in with an spi layer around 1000 lines of code to implement the above 7 interfaces. This is a huge reduction in size compared to prior attempts to implement a new JGit storage layer. As this package shows, a complete JGit storage implementation is more than 17,000 lines of fairly complex code. A simple cache is provided in storage.dht.spi.cache. Implementers can use CacheDatabase to wrap any other type of Database and perform fast reads against a network based cache service, such as the open source memcached[4]. An implementation of CacheService must be provided to glue this spi onto the network cache. [1] https://github.com/spearce/jgit_hbase [2] https://github.com/spearce/jgit_cassandra [3] http://labs.google.com/papers/bigtable.html [4] http://memcached.org/ Change-Id: I0aa4072781f5ccc019ca421c036adff2c40c4295 Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
13 年之前
Store Git on any DHT jgit.storage.dht is a storage provider implementation for JGit that permits storing the Git repository in a distributed hashtable, NoSQL system, or other database. The actual underlying storage system is undefined, and can be plugged in by implementing 7 small interfaces: * Database * RepositoryIndexTable * RepositoryTable * RefTable * ChunkTable * ObjectIndexTable * WriteBuffer The storage provider interface tries to assume very little about the underlying storage system, and requires only three key features: * key -> value lookup (a hashtable is suitable) * atomic updates on single rows * asynchronous operations (Java's ExecutorService is easy to use) Most NoSQL database products offer all 3 of these features in their clients, and so does any decent network based cache system like the open source memcache product. Relying only on key equality for data retrevial makes it simple for the storage engine to distribute across multiple machines. Traditional SQL systems could also be used with a JDBC based spi implementation. Before submitting this change I have implemented six storage systems for the spi layer: * Apache HBase[1] * Apache Cassandra[2] * Google Bigtable[3] * an in-memory implementation for unit testing * a JDBC implementation for SQL * a generic cache provider that can ride on top of memcache All six systems came in with an spi layer around 1000 lines of code to implement the above 7 interfaces. This is a huge reduction in size compared to prior attempts to implement a new JGit storage layer. As this package shows, a complete JGit storage implementation is more than 17,000 lines of fairly complex code. A simple cache is provided in storage.dht.spi.cache. Implementers can use CacheDatabase to wrap any other type of Database and perform fast reads against a network based cache service, such as the open source memcached[4]. An implementation of CacheService must be provided to glue this spi onto the network cache. [1] https://github.com/spearce/jgit_hbase [2] https://github.com/spearce/jgit_cassandra [3] http://labs.google.com/papers/bigtable.html [4] http://memcached.org/ Change-Id: I0aa4072781f5ccc019ca421c036adff2c40c4295 Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
13 年之前
Store Git on any DHT jgit.storage.dht is a storage provider implementation for JGit that permits storing the Git repository in a distributed hashtable, NoSQL system, or other database. The actual underlying storage system is undefined, and can be plugged in by implementing 7 small interfaces: * Database * RepositoryIndexTable * RepositoryTable * RefTable * ChunkTable * ObjectIndexTable * WriteBuffer The storage provider interface tries to assume very little about the underlying storage system, and requires only three key features: * key -> value lookup (a hashtable is suitable) * atomic updates on single rows * asynchronous operations (Java's ExecutorService is easy to use) Most NoSQL database products offer all 3 of these features in their clients, and so does any decent network based cache system like the open source memcache product. Relying only on key equality for data retrevial makes it simple for the storage engine to distribute across multiple machines. Traditional SQL systems could also be used with a JDBC based spi implementation. Before submitting this change I have implemented six storage systems for the spi layer: * Apache HBase[1] * Apache Cassandra[2] * Google Bigtable[3] * an in-memory implementation for unit testing * a JDBC implementation for SQL * a generic cache provider that can ride on top of memcache All six systems came in with an spi layer around 1000 lines of code to implement the above 7 interfaces. This is a huge reduction in size compared to prior attempts to implement a new JGit storage layer. As this package shows, a complete JGit storage implementation is more than 17,000 lines of fairly complex code. A simple cache is provided in storage.dht.spi.cache. Implementers can use CacheDatabase to wrap any other type of Database and perform fast reads against a network based cache service, such as the open source memcached[4]. An implementation of CacheService must be provided to glue this spi onto the network cache. [1] https://github.com/spearce/jgit_hbase [2] https://github.com/spearce/jgit_cassandra [3] http://labs.google.com/papers/bigtable.html [4] http://memcached.org/ Change-Id: I0aa4072781f5ccc019ca421c036adff2c40c4295 Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
13 年之前
Store Git on any DHT jgit.storage.dht is a storage provider implementation for JGit that permits storing the Git repository in a distributed hashtable, NoSQL system, or other database. The actual underlying storage system is undefined, and can be plugged in by implementing 7 small interfaces: * Database * RepositoryIndexTable * RepositoryTable * RefTable * ChunkTable * ObjectIndexTable * WriteBuffer The storage provider interface tries to assume very little about the underlying storage system, and requires only three key features: * key -> value lookup (a hashtable is suitable) * atomic updates on single rows * asynchronous operations (Java's ExecutorService is easy to use) Most NoSQL database products offer all 3 of these features in their clients, and so does any decent network based cache system like the open source memcache product. Relying only on key equality for data retrevial makes it simple for the storage engine to distribute across multiple machines. Traditional SQL systems could also be used with a JDBC based spi implementation. Before submitting this change I have implemented six storage systems for the spi layer: * Apache HBase[1] * Apache Cassandra[2] * Google Bigtable[3] * an in-memory implementation for unit testing * a JDBC implementation for SQL * a generic cache provider that can ride on top of memcache All six systems came in with an spi layer around 1000 lines of code to implement the above 7 interfaces. This is a huge reduction in size compared to prior attempts to implement a new JGit storage layer. As this package shows, a complete JGit storage implementation is more than 17,000 lines of fairly complex code. A simple cache is provided in storage.dht.spi.cache. Implementers can use CacheDatabase to wrap any other type of Database and perform fast reads against a network based cache service, such as the open source memcached[4]. An implementation of CacheService must be provided to glue this spi onto the network cache. [1] https://github.com/spearce/jgit_hbase [2] https://github.com/spearce/jgit_cassandra [3] http://labs.google.com/papers/bigtable.html [4] http://memcached.org/ Change-Id: I0aa4072781f5ccc019ca421c036adff2c40c4295 Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
13 年之前
Store Git on any DHT jgit.storage.dht is a storage provider implementation for JGit that permits storing the Git repository in a distributed hashtable, NoSQL system, or other database. The actual underlying storage system is undefined, and can be plugged in by implementing 7 small interfaces: * Database * RepositoryIndexTable * RepositoryTable * RefTable * ChunkTable * ObjectIndexTable * WriteBuffer The storage provider interface tries to assume very little about the underlying storage system, and requires only three key features: * key -> value lookup (a hashtable is suitable) * atomic updates on single rows * asynchronous operations (Java's ExecutorService is easy to use) Most NoSQL database products offer all 3 of these features in their clients, and so does any decent network based cache system like the open source memcache product. Relying only on key equality for data retrevial makes it simple for the storage engine to distribute across multiple machines. Traditional SQL systems could also be used with a JDBC based spi implementation. Before submitting this change I have implemented six storage systems for the spi layer: * Apache HBase[1] * Apache Cassandra[2] * Google Bigtable[3] * an in-memory implementation for unit testing * a JDBC implementation for SQL * a generic cache provider that can ride on top of memcache All six systems came in with an spi layer around 1000 lines of code to implement the above 7 interfaces. This is a huge reduction in size compared to prior attempts to implement a new JGit storage layer. As this package shows, a complete JGit storage implementation is more than 17,000 lines of fairly complex code. A simple cache is provided in storage.dht.spi.cache. Implementers can use CacheDatabase to wrap any other type of Database and perform fast reads against a network based cache service, such as the open source memcached[4]. An implementation of CacheService must be provided to glue this spi onto the network cache. [1] https://github.com/spearce/jgit_hbase [2] https://github.com/spearce/jgit_cassandra [3] http://labs.google.com/papers/bigtable.html [4] http://memcached.org/ Change-Id: I0aa4072781f5ccc019ca421c036adff2c40c4295 Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
13 年之前
Store Git on any DHT jgit.storage.dht is a storage provider implementation for JGit that permits storing the Git repository in a distributed hashtable, NoSQL system, or other database. The actual underlying storage system is undefined, and can be plugged in by implementing 7 small interfaces: * Database * RepositoryIndexTable * RepositoryTable * RefTable * ChunkTable * ObjectIndexTable * WriteBuffer The storage provider interface tries to assume very little about the underlying storage system, and requires only three key features: * key -> value lookup (a hashtable is suitable) * atomic updates on single rows * asynchronous operations (Java's ExecutorService is easy to use) Most NoSQL database products offer all 3 of these features in their clients, and so does any decent network based cache system like the open source memcache product. Relying only on key equality for data retrevial makes it simple for the storage engine to distribute across multiple machines. Traditional SQL systems could also be used with a JDBC based spi implementation. Before submitting this change I have implemented six storage systems for the spi layer: * Apache HBase[1] * Apache Cassandra[2] * Google Bigtable[3] * an in-memory implementation for unit testing * a JDBC implementation for SQL * a generic cache provider that can ride on top of memcache All six systems came in with an spi layer around 1000 lines of code to implement the above 7 interfaces. This is a huge reduction in size compared to prior attempts to implement a new JGit storage layer. As this package shows, a complete JGit storage implementation is more than 17,000 lines of fairly complex code. A simple cache is provided in storage.dht.spi.cache. Implementers can use CacheDatabase to wrap any other type of Database and perform fast reads against a network based cache service, such as the open source memcached[4]. An implementation of CacheService must be provided to glue this spi onto the network cache. [1] https://github.com/spearce/jgit_hbase [2] https://github.com/spearce/jgit_cassandra [3] http://labs.google.com/papers/bigtable.html [4] http://memcached.org/ Change-Id: I0aa4072781f5ccc019ca421c036adff2c40c4295 Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
13 年之前
Store Git on any DHT jgit.storage.dht is a storage provider implementation for JGit that permits storing the Git repository in a distributed hashtable, NoSQL system, or other database. The actual underlying storage system is undefined, and can be plugged in by implementing 7 small interfaces: * Database * RepositoryIndexTable * RepositoryTable * RefTable * ChunkTable * ObjectIndexTable * WriteBuffer The storage provider interface tries to assume very little about the underlying storage system, and requires only three key features: * key -> value lookup (a hashtable is suitable) * atomic updates on single rows * asynchronous operations (Java's ExecutorService is easy to use) Most NoSQL database products offer all 3 of these features in their clients, and so does any decent network based cache system like the open source memcache product. Relying only on key equality for data retrevial makes it simple for the storage engine to distribute across multiple machines. Traditional SQL systems could also be used with a JDBC based spi implementation. Before submitting this change I have implemented six storage systems for the spi layer: * Apache HBase[1] * Apache Cassandra[2] * Google Bigtable[3] * an in-memory implementation for unit testing * a JDBC implementation for SQL * a generic cache provider that can ride on top of memcache All six systems came in with an spi layer around 1000 lines of code to implement the above 7 interfaces. This is a huge reduction in size compared to prior attempts to implement a new JGit storage layer. As this package shows, a complete JGit storage implementation is more than 17,000 lines of fairly complex code. A simple cache is provided in storage.dht.spi.cache. Implementers can use CacheDatabase to wrap any other type of Database and perform fast reads against a network based cache service, such as the open source memcached[4]. An implementation of CacheService must be provided to glue this spi onto the network cache. [1] https://github.com/spearce/jgit_hbase [2] https://github.com/spearce/jgit_cassandra [3] http://labs.google.com/papers/bigtable.html [4] http://memcached.org/ Change-Id: I0aa4072781f5ccc019ca421c036adff2c40c4295 Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
13 年之前
Store Git on any DHT jgit.storage.dht is a storage provider implementation for JGit that permits storing the Git repository in a distributed hashtable, NoSQL system, or other database. The actual underlying storage system is undefined, and can be plugged in by implementing 7 small interfaces: * Database * RepositoryIndexTable * RepositoryTable * RefTable * ChunkTable * ObjectIndexTable * WriteBuffer The storage provider interface tries to assume very little about the underlying storage system, and requires only three key features: * key -> value lookup (a hashtable is suitable) * atomic updates on single rows * asynchronous operations (Java's ExecutorService is easy to use) Most NoSQL database products offer all 3 of these features in their clients, and so does any decent network based cache system like the open source memcache product. Relying only on key equality for data retrevial makes it simple for the storage engine to distribute across multiple machines. Traditional SQL systems could also be used with a JDBC based spi implementation. Before submitting this change I have implemented six storage systems for the spi layer: * Apache HBase[1] * Apache Cassandra[2] * Google Bigtable[3] * an in-memory implementation for unit testing * a JDBC implementation for SQL * a generic cache provider that can ride on top of memcache All six systems came in with an spi layer around 1000 lines of code to implement the above 7 interfaces. This is a huge reduction in size compared to prior attempts to implement a new JGit storage layer. As this package shows, a complete JGit storage implementation is more than 17,000 lines of fairly complex code. A simple cache is provided in storage.dht.spi.cache. Implementers can use CacheDatabase to wrap any other type of Database and perform fast reads against a network based cache service, such as the open source memcached[4]. An implementation of CacheService must be provided to glue this spi onto the network cache. [1] https://github.com/spearce/jgit_hbase [2] https://github.com/spearce/jgit_cassandra [3] http://labs.google.com/papers/bigtable.html [4] http://memcached.org/ Change-Id: I0aa4072781f5ccc019ca421c036adff2c40c4295 Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
13 年之前
Store Git on any DHT jgit.storage.dht is a storage provider implementation for JGit that permits storing the Git repository in a distributed hashtable, NoSQL system, or other database. The actual underlying storage system is undefined, and can be plugged in by implementing 7 small interfaces: * Database * RepositoryIndexTable * RepositoryTable * RefTable * ChunkTable * ObjectIndexTable * WriteBuffer The storage provider interface tries to assume very little about the underlying storage system, and requires only three key features: * key -> value lookup (a hashtable is suitable) * atomic updates on single rows * asynchronous operations (Java's ExecutorService is easy to use) Most NoSQL database products offer all 3 of these features in their clients, and so does any decent network based cache system like the open source memcache product. Relying only on key equality for data retrevial makes it simple for the storage engine to distribute across multiple machines. Traditional SQL systems could also be used with a JDBC based spi implementation. Before submitting this change I have implemented six storage systems for the spi layer: * Apache HBase[1] * Apache Cassandra[2] * Google Bigtable[3] * an in-memory implementation for unit testing * a JDBC implementation for SQL * a generic cache provider that can ride on top of memcache All six systems came in with an spi layer around 1000 lines of code to implement the above 7 interfaces. This is a huge reduction in size compared to prior attempts to implement a new JGit storage layer. As this package shows, a complete JGit storage implementation is more than 17,000 lines of fairly complex code. A simple cache is provided in storage.dht.spi.cache. Implementers can use CacheDatabase to wrap any other type of Database and perform fast reads against a network based cache service, such as the open source memcached[4]. An implementation of CacheService must be provided to glue this spi onto the network cache. [1] https://github.com/spearce/jgit_hbase [2] https://github.com/spearce/jgit_cassandra [3] http://labs.google.com/papers/bigtable.html [4] http://memcached.org/ Change-Id: I0aa4072781f5ccc019ca421c036adff2c40c4295 Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
13 年之前
Store Git on any DHT jgit.storage.dht is a storage provider implementation for JGit that permits storing the Git repository in a distributed hashtable, NoSQL system, or other database. The actual underlying storage system is undefined, and can be plugged in by implementing 7 small interfaces: * Database * RepositoryIndexTable * RepositoryTable * RefTable * ChunkTable * ObjectIndexTable * WriteBuffer The storage provider interface tries to assume very little about the underlying storage system, and requires only three key features: * key -> value lookup (a hashtable is suitable) * atomic updates on single rows * asynchronous operations (Java's ExecutorService is easy to use) Most NoSQL database products offer all 3 of these features in their clients, and so does any decent network based cache system like the open source memcache product. Relying only on key equality for data retrevial makes it simple for the storage engine to distribute across multiple machines. Traditional SQL systems could also be used with a JDBC based spi implementation. Before submitting this change I have implemented six storage systems for the spi layer: * Apache HBase[1] * Apache Cassandra[2] * Google Bigtable[3] * an in-memory implementation for unit testing * a JDBC implementation for SQL * a generic cache provider that can ride on top of memcache All six systems came in with an spi layer around 1000 lines of code to implement the above 7 interfaces. This is a huge reduction in size compared to prior attempts to implement a new JGit storage layer. As this package shows, a complete JGit storage implementation is more than 17,000 lines of fairly complex code. A simple cache is provided in storage.dht.spi.cache. Implementers can use CacheDatabase to wrap any other type of Database and perform fast reads against a network based cache service, such as the open source memcached[4]. An implementation of CacheService must be provided to glue this spi onto the network cache. [1] https://github.com/spearce/jgit_hbase [2] https://github.com/spearce/jgit_cassandra [3] http://labs.google.com/papers/bigtable.html [4] http://memcached.org/ Change-Id: I0aa4072781f5ccc019ca421c036adff2c40c4295 Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
13 年之前
Store Git on any DHT jgit.storage.dht is a storage provider implementation for JGit that permits storing the Git repository in a distributed hashtable, NoSQL system, or other database. The actual underlying storage system is undefined, and can be plugged in by implementing 7 small interfaces: * Database * RepositoryIndexTable * RepositoryTable * RefTable * ChunkTable * ObjectIndexTable * WriteBuffer The storage provider interface tries to assume very little about the underlying storage system, and requires only three key features: * key -> value lookup (a hashtable is suitable) * atomic updates on single rows * asynchronous operations (Java's ExecutorService is easy to use) Most NoSQL database products offer all 3 of these features in their clients, and so does any decent network based cache system like the open source memcache product. Relying only on key equality for data retrevial makes it simple for the storage engine to distribute across multiple machines. Traditional SQL systems could also be used with a JDBC based spi implementation. Before submitting this change I have implemented six storage systems for the spi layer: * Apache HBase[1] * Apache Cassandra[2] * Google Bigtable[3] * an in-memory implementation for unit testing * a JDBC implementation for SQL * a generic cache provider that can ride on top of memcache All six systems came in with an spi layer around 1000 lines of code to implement the above 7 interfaces. This is a huge reduction in size compared to prior attempts to implement a new JGit storage layer. As this package shows, a complete JGit storage implementation is more than 17,000 lines of fairly complex code. A simple cache is provided in storage.dht.spi.cache. Implementers can use CacheDatabase to wrap any other type of Database and perform fast reads against a network based cache service, such as the open source memcached[4]. An implementation of CacheService must be provided to glue this spi onto the network cache. [1] https://github.com/spearce/jgit_hbase [2] https://github.com/spearce/jgit_cassandra [3] http://labs.google.com/papers/bigtable.html [4] http://memcached.org/ Change-Id: I0aa4072781f5ccc019ca421c036adff2c40c4295 Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
13 年之前
Store Git on any DHT jgit.storage.dht is a storage provider implementation for JGit that permits storing the Git repository in a distributed hashtable, NoSQL system, or other database. The actual underlying storage system is undefined, and can be plugged in by implementing 7 small interfaces: * Database * RepositoryIndexTable * RepositoryTable * RefTable * ChunkTable * ObjectIndexTable * WriteBuffer The storage provider interface tries to assume very little about the underlying storage system, and requires only three key features: * key -> value lookup (a hashtable is suitable) * atomic updates on single rows * asynchronous operations (Java's ExecutorService is easy to use) Most NoSQL database products offer all 3 of these features in their clients, and so does any decent network based cache system like the open source memcache product. Relying only on key equality for data retrevial makes it simple for the storage engine to distribute across multiple machines. Traditional SQL systems could also be used with a JDBC based spi implementation. Before submitting this change I have implemented six storage systems for the spi layer: * Apache HBase[1] * Apache Cassandra[2] * Google Bigtable[3] * an in-memory implementation for unit testing * a JDBC implementation for SQL * a generic cache provider that can ride on top of memcache All six systems came in with an spi layer around 1000 lines of code to implement the above 7 interfaces. This is a huge reduction in size compared to prior attempts to implement a new JGit storage layer. As this package shows, a complete JGit storage implementation is more than 17,000 lines of fairly complex code. A simple cache is provided in storage.dht.spi.cache. Implementers can use CacheDatabase to wrap any other type of Database and perform fast reads against a network based cache service, such as the open source memcached[4]. An implementation of CacheService must be provided to glue this spi onto the network cache. [1] https://github.com/spearce/jgit_hbase [2] https://github.com/spearce/jgit_cassandra [3] http://labs.google.com/papers/bigtable.html [4] http://memcached.org/ Change-Id: I0aa4072781f5ccc019ca421c036adff2c40c4295 Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
13 年之前
Store Git on any DHT jgit.storage.dht is a storage provider implementation for JGit that permits storing the Git repository in a distributed hashtable, NoSQL system, or other database. The actual underlying storage system is undefined, and can be plugged in by implementing 7 small interfaces: * Database * RepositoryIndexTable * RepositoryTable * RefTable * ChunkTable * ObjectIndexTable * WriteBuffer The storage provider interface tries to assume very little about the underlying storage system, and requires only three key features: * key -> value lookup (a hashtable is suitable) * atomic updates on single rows * asynchronous operations (Java's ExecutorService is easy to use) Most NoSQL database products offer all 3 of these features in their clients, and so does any decent network based cache system like the open source memcache product. Relying only on key equality for data retrevial makes it simple for the storage engine to distribute across multiple machines. Traditional SQL systems could also be used with a JDBC based spi implementation. Before submitting this change I have implemented six storage systems for the spi layer: * Apache HBase[1] * Apache Cassandra[2] * Google Bigtable[3] * an in-memory implementation for unit testing * a JDBC implementation for SQL * a generic cache provider that can ride on top of memcache All six systems came in with an spi layer around 1000 lines of code to implement the above 7 interfaces. This is a huge reduction in size compared to prior attempts to implement a new JGit storage layer. As this package shows, a complete JGit storage implementation is more than 17,000 lines of fairly complex code. A simple cache is provided in storage.dht.spi.cache. Implementers can use CacheDatabase to wrap any other type of Database and perform fast reads against a network based cache service, such as the open source memcached[4]. An implementation of CacheService must be provided to glue this spi onto the network cache. [1] https://github.com/spearce/jgit_hbase [2] https://github.com/spearce/jgit_cassandra [3] http://labs.google.com/papers/bigtable.html [4] http://memcached.org/ Change-Id: I0aa4072781f5ccc019ca421c036adff2c40c4295 Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
13 年之前
Store Git on any DHT jgit.storage.dht is a storage provider implementation for JGit that permits storing the Git repository in a distributed hashtable, NoSQL system, or other database. The actual underlying storage system is undefined, and can be plugged in by implementing 7 small interfaces: * Database * RepositoryIndexTable * RepositoryTable * RefTable * ChunkTable * ObjectIndexTable * WriteBuffer The storage provider interface tries to assume very little about the underlying storage system, and requires only three key features: * key -> value lookup (a hashtable is suitable) * atomic updates on single rows * asynchronous operations (Java's ExecutorService is easy to use) Most NoSQL database products offer all 3 of these features in their clients, and so does any decent network based cache system like the open source memcache product. Relying only on key equality for data retrevial makes it simple for the storage engine to distribute across multiple machines. Traditional SQL systems could also be used with a JDBC based spi implementation. Before submitting this change I have implemented six storage systems for the spi layer: * Apache HBase[1] * Apache Cassandra[2] * Google Bigtable[3] * an in-memory implementation for unit testing * a JDBC implementation for SQL * a generic cache provider that can ride on top of memcache All six systems came in with an spi layer around 1000 lines of code to implement the above 7 interfaces. This is a huge reduction in size compared to prior attempts to implement a new JGit storage layer. As this package shows, a complete JGit storage implementation is more than 17,000 lines of fairly complex code. A simple cache is provided in storage.dht.spi.cache. Implementers can use CacheDatabase to wrap any other type of Database and perform fast reads against a network based cache service, such as the open source memcached[4]. An implementation of CacheService must be provided to glue this spi onto the network cache. [1] https://github.com/spearce/jgit_hbase [2] https://github.com/spearce/jgit_cassandra [3] http://labs.google.com/papers/bigtable.html [4] http://memcached.org/ Change-Id: I0aa4072781f5ccc019ca421c036adff2c40c4295 Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
13 年之前
Store Git on any DHT jgit.storage.dht is a storage provider implementation for JGit that permits storing the Git repository in a distributed hashtable, NoSQL system, or other database. The actual underlying storage system is undefined, and can be plugged in by implementing 7 small interfaces: * Database * RepositoryIndexTable * RepositoryTable * RefTable * ChunkTable * ObjectIndexTable * WriteBuffer The storage provider interface tries to assume very little about the underlying storage system, and requires only three key features: * key -> value lookup (a hashtable is suitable) * atomic updates on single rows * asynchronous operations (Java's ExecutorService is easy to use) Most NoSQL database products offer all 3 of these features in their clients, and so does any decent network based cache system like the open source memcache product. Relying only on key equality for data retrevial makes it simple for the storage engine to distribute across multiple machines. Traditional SQL systems could also be used with a JDBC based spi implementation. Before submitting this change I have implemented six storage systems for the spi layer: * Apache HBase[1] * Apache Cassandra[2] * Google Bigtable[3] * an in-memory implementation for unit testing * a JDBC implementation for SQL * a generic cache provider that can ride on top of memcache All six systems came in with an spi layer around 1000 lines of code to implement the above 7 interfaces. This is a huge reduction in size compared to prior attempts to implement a new JGit storage layer. As this package shows, a complete JGit storage implementation is more than 17,000 lines of fairly complex code. A simple cache is provided in storage.dht.spi.cache. Implementers can use CacheDatabase to wrap any other type of Database and perform fast reads against a network based cache service, such as the open source memcached[4]. An implementation of CacheService must be provided to glue this spi onto the network cache. [1] https://github.com/spearce/jgit_hbase [2] https://github.com/spearce/jgit_cassandra [3] http://labs.google.com/papers/bigtable.html [4] http://memcached.org/ Change-Id: I0aa4072781f5ccc019ca421c036adff2c40c4295 Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
13 年之前
Store Git on any DHT jgit.storage.dht is a storage provider implementation for JGit that permits storing the Git repository in a distributed hashtable, NoSQL system, or other database. The actual underlying storage system is undefined, and can be plugged in by implementing 7 small interfaces: * Database * RepositoryIndexTable * RepositoryTable * RefTable * ChunkTable * ObjectIndexTable * WriteBuffer The storage provider interface tries to assume very little about the underlying storage system, and requires only three key features: * key -> value lookup (a hashtable is suitable) * atomic updates on single rows * asynchronous operations (Java's ExecutorService is easy to use) Most NoSQL database products offer all 3 of these features in their clients, and so does any decent network based cache system like the open source memcache product. Relying only on key equality for data retrevial makes it simple for the storage engine to distribute across multiple machines. Traditional SQL systems could also be used with a JDBC based spi implementation. Before submitting this change I have implemented six storage systems for the spi layer: * Apache HBase[1] * Apache Cassandra[2] * Google Bigtable[3] * an in-memory implementation for unit testing * a JDBC implementation for SQL * a generic cache provider that can ride on top of memcache All six systems came in with an spi layer around 1000 lines of code to implement the above 7 interfaces. This is a huge reduction in size compared to prior attempts to implement a new JGit storage layer. As this package shows, a complete JGit storage implementation is more than 17,000 lines of fairly complex code. A simple cache is provided in storage.dht.spi.cache. Implementers can use CacheDatabase to wrap any other type of Database and perform fast reads against a network based cache service, such as the open source memcached[4]. An implementation of CacheService must be provided to glue this spi onto the network cache. [1] https://github.com/spearce/jgit_hbase [2] https://github.com/spearce/jgit_cassandra [3] http://labs.google.com/papers/bigtable.html [4] http://memcached.org/ Change-Id: I0aa4072781f5ccc019ca421c036adff2c40c4295 Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
13 年之前
Store Git on any DHT jgit.storage.dht is a storage provider implementation for JGit that permits storing the Git repository in a distributed hashtable, NoSQL system, or other database. The actual underlying storage system is undefined, and can be plugged in by implementing 7 small interfaces: * Database * RepositoryIndexTable * RepositoryTable * RefTable * ChunkTable * ObjectIndexTable * WriteBuffer The storage provider interface tries to assume very little about the underlying storage system, and requires only three key features: * key -> value lookup (a hashtable is suitable) * atomic updates on single rows * asynchronous operations (Java's ExecutorService is easy to use) Most NoSQL database products offer all 3 of these features in their clients, and so does any decent network based cache system like the open source memcache product. Relying only on key equality for data retrevial makes it simple for the storage engine to distribute across multiple machines. Traditional SQL systems could also be used with a JDBC based spi implementation. Before submitting this change I have implemented six storage systems for the spi layer: * Apache HBase[1] * Apache Cassandra[2] * Google Bigtable[3] * an in-memory implementation for unit testing * a JDBC implementation for SQL * a generic cache provider that can ride on top of memcache All six systems came in with an spi layer around 1000 lines of code to implement the above 7 interfaces. This is a huge reduction in size compared to prior attempts to implement a new JGit storage layer. As this package shows, a complete JGit storage implementation is more than 17,000 lines of fairly complex code. A simple cache is provided in storage.dht.spi.cache. Implementers can use CacheDatabase to wrap any other type of Database and perform fast reads against a network based cache service, such as the open source memcached[4]. An implementation of CacheService must be provided to glue this spi onto the network cache. [1] https://github.com/spearce/jgit_hbase [2] https://github.com/spearce/jgit_cassandra [3] http://labs.google.com/papers/bigtable.html [4] http://memcached.org/ Change-Id: I0aa4072781f5ccc019ca421c036adff2c40c4295 Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
13 年之前
Store Git on any DHT jgit.storage.dht is a storage provider implementation for JGit that permits storing the Git repository in a distributed hashtable, NoSQL system, or other database. The actual underlying storage system is undefined, and can be plugged in by implementing 7 small interfaces: * Database * RepositoryIndexTable * RepositoryTable * RefTable * ChunkTable * ObjectIndexTable * WriteBuffer The storage provider interface tries to assume very little about the underlying storage system, and requires only three key features: * key -> value lookup (a hashtable is suitable) * atomic updates on single rows * asynchronous operations (Java's ExecutorService is easy to use) Most NoSQL database products offer all 3 of these features in their clients, and so does any decent network based cache system like the open source memcache product. Relying only on key equality for data retrevial makes it simple for the storage engine to distribute across multiple machines. Traditional SQL systems could also be used with a JDBC based spi implementation. Before submitting this change I have implemented six storage systems for the spi layer: * Apache HBase[1] * Apache Cassandra[2] * Google Bigtable[3] * an in-memory implementation for unit testing * a JDBC implementation for SQL * a generic cache provider that can ride on top of memcache All six systems came in with an spi layer around 1000 lines of code to implement the above 7 interfaces. This is a huge reduction in size compared to prior attempts to implement a new JGit storage layer. As this package shows, a complete JGit storage implementation is more than 17,000 lines of fairly complex code. A simple cache is provided in storage.dht.spi.cache. Implementers can use CacheDatabase to wrap any other type of Database and perform fast reads against a network based cache service, such as the open source memcached[4]. An implementation of CacheService must be provided to glue this spi onto the network cache. [1] https://github.com/spearce/jgit_hbase [2] https://github.com/spearce/jgit_cassandra [3] http://labs.google.com/papers/bigtable.html [4] http://memcached.org/ Change-Id: I0aa4072781f5ccc019ca421c036adff2c40c4295 Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
13 年之前
Store Git on any DHT jgit.storage.dht is a storage provider implementation for JGit that permits storing the Git repository in a distributed hashtable, NoSQL system, or other database. The actual underlying storage system is undefined, and can be plugged in by implementing 7 small interfaces: * Database * RepositoryIndexTable * RepositoryTable * RefTable * ChunkTable * ObjectIndexTable * WriteBuffer The storage provider interface tries to assume very little about the underlying storage system, and requires only three key features: * key -> value lookup (a hashtable is suitable) * atomic updates on single rows * asynchronous operations (Java's ExecutorService is easy to use) Most NoSQL database products offer all 3 of these features in their clients, and so does any decent network based cache system like the open source memcache product. Relying only on key equality for data retrevial makes it simple for the storage engine to distribute across multiple machines. Traditional SQL systems could also be used with a JDBC based spi implementation. Before submitting this change I have implemented six storage systems for the spi layer: * Apache HBase[1] * Apache Cassandra[2] * Google Bigtable[3] * an in-memory implementation for unit testing * a JDBC implementation for SQL * a generic cache provider that can ride on top of memcache All six systems came in with an spi layer around 1000 lines of code to implement the above 7 interfaces. This is a huge reduction in size compared to prior attempts to implement a new JGit storage layer. As this package shows, a complete JGit storage implementation is more than 17,000 lines of fairly complex code. A simple cache is provided in storage.dht.spi.cache. Implementers can use CacheDatabase to wrap any other type of Database and perform fast reads against a network based cache service, such as the open source memcached[4]. An implementation of CacheService must be provided to glue this spi onto the network cache. [1] https://github.com/spearce/jgit_hbase [2] https://github.com/spearce/jgit_cassandra [3] http://labs.google.com/papers/bigtable.html [4] http://memcached.org/ Change-Id: I0aa4072781f5ccc019ca421c036adff2c40c4295 Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
13 年之前
Store Git on any DHT jgit.storage.dht is a storage provider implementation for JGit that permits storing the Git repository in a distributed hashtable, NoSQL system, or other database. The actual underlying storage system is undefined, and can be plugged in by implementing 7 small interfaces: * Database * RepositoryIndexTable * RepositoryTable * RefTable * ChunkTable * ObjectIndexTable * WriteBuffer The storage provider interface tries to assume very little about the underlying storage system, and requires only three key features: * key -> value lookup (a hashtable is suitable) * atomic updates on single rows * asynchronous operations (Java's ExecutorService is easy to use) Most NoSQL database products offer all 3 of these features in their clients, and so does any decent network based cache system like the open source memcache product. Relying only on key equality for data retrevial makes it simple for the storage engine to distribute across multiple machines. Traditional SQL systems could also be used with a JDBC based spi implementation. Before submitting this change I have implemented six storage systems for the spi layer: * Apache HBase[1] * Apache Cassandra[2] * Google Bigtable[3] * an in-memory implementation for unit testing * a JDBC implementation for SQL * a generic cache provider that can ride on top of memcache All six systems came in with an spi layer around 1000 lines of code to implement the above 7 interfaces. This is a huge reduction in size compared to prior attempts to implement a new JGit storage layer. As this package shows, a complete JGit storage implementation is more than 17,000 lines of fairly complex code. A simple cache is provided in storage.dht.spi.cache. Implementers can use CacheDatabase to wrap any other type of Database and perform fast reads against a network based cache service, such as the open source memcached[4]. An implementation of CacheService must be provided to glue this spi onto the network cache. [1] https://github.com/spearce/jgit_hbase [2] https://github.com/spearce/jgit_cassandra [3] http://labs.google.com/papers/bigtable.html [4] http://memcached.org/ Change-Id: I0aa4072781f5ccc019ca421c036adff2c40c4295 Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
13 年之前
Store Git on any DHT jgit.storage.dht is a storage provider implementation for JGit that permits storing the Git repository in a distributed hashtable, NoSQL system, or other database. The actual underlying storage system is undefined, and can be plugged in by implementing 7 small interfaces: * Database * RepositoryIndexTable * RepositoryTable * RefTable * ChunkTable * ObjectIndexTable * WriteBuffer The storage provider interface tries to assume very little about the underlying storage system, and requires only three key features: * key -> value lookup (a hashtable is suitable) * atomic updates on single rows * asynchronous operations (Java's ExecutorService is easy to use) Most NoSQL database products offer all 3 of these features in their clients, and so does any decent network based cache system like the open source memcache product. Relying only on key equality for data retrevial makes it simple for the storage engine to distribute across multiple machines. Traditional SQL systems could also be used with a JDBC based spi implementation. Before submitting this change I have implemented six storage systems for the spi layer: * Apache HBase[1] * Apache Cassandra[2] * Google Bigtable[3] * an in-memory implementation for unit testing * a JDBC implementation for SQL * a generic cache provider that can ride on top of memcache All six systems came in with an spi layer around 1000 lines of code to implement the above 7 interfaces. This is a huge reduction in size compared to prior attempts to implement a new JGit storage layer. As this package shows, a complete JGit storage implementation is more than 17,000 lines of fairly complex code. A simple cache is provided in storage.dht.spi.cache. Implementers can use CacheDatabase to wrap any other type of Database and perform fast reads against a network based cache service, such as the open source memcached[4]. An implementation of CacheService must be provided to glue this spi onto the network cache. [1] https://github.com/spearce/jgit_hbase [2] https://github.com/spearce/jgit_cassandra [3] http://labs.google.com/papers/bigtable.html [4] http://memcached.org/ Change-Id: I0aa4072781f5ccc019ca421c036adff2c40c4295 Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
13 年之前
Store Git on any DHT jgit.storage.dht is a storage provider implementation for JGit that permits storing the Git repository in a distributed hashtable, NoSQL system, or other database. The actual underlying storage system is undefined, and can be plugged in by implementing 7 small interfaces: * Database * RepositoryIndexTable * RepositoryTable * RefTable * ChunkTable * ObjectIndexTable * WriteBuffer The storage provider interface tries to assume very little about the underlying storage system, and requires only three key features: * key -> value lookup (a hashtable is suitable) * atomic updates on single rows * asynchronous operations (Java's ExecutorService is easy to use) Most NoSQL database products offer all 3 of these features in their clients, and so does any decent network based cache system like the open source memcache product. Relying only on key equality for data retrevial makes it simple for the storage engine to distribute across multiple machines. Traditional SQL systems could also be used with a JDBC based spi implementation. Before submitting this change I have implemented six storage systems for the spi layer: * Apache HBase[1] * Apache Cassandra[2] * Google Bigtable[3] * an in-memory implementation for unit testing * a JDBC implementation for SQL * a generic cache provider that can ride on top of memcache All six systems came in with an spi layer around 1000 lines of code to implement the above 7 interfaces. This is a huge reduction in size compared to prior attempts to implement a new JGit storage layer. As this package shows, a complete JGit storage implementation is more than 17,000 lines of fairly complex code. A simple cache is provided in storage.dht.spi.cache. Implementers can use CacheDatabase to wrap any other type of Database and perform fast reads against a network based cache service, such as the open source memcached[4]. An implementation of CacheService must be provided to glue this spi onto the network cache. [1] https://github.com/spearce/jgit_hbase [2] https://github.com/spearce/jgit_cassandra [3] http://labs.google.com/papers/bigtable.html [4] http://memcached.org/ Change-Id: I0aa4072781f5ccc019ca421c036adff2c40c4295 Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
13 年之前
Store Git on any DHT jgit.storage.dht is a storage provider implementation for JGit that permits storing the Git repository in a distributed hashtable, NoSQL system, or other database. The actual underlying storage system is undefined, and can be plugged in by implementing 7 small interfaces: * Database * RepositoryIndexTable * RepositoryTable * RefTable * ChunkTable * ObjectIndexTable * WriteBuffer The storage provider interface tries to assume very little about the underlying storage system, and requires only three key features: * key -> value lookup (a hashtable is suitable) * atomic updates on single rows * asynchronous operations (Java's ExecutorService is easy to use) Most NoSQL database products offer all 3 of these features in their clients, and so does any decent network based cache system like the open source memcache product. Relying only on key equality for data retrevial makes it simple for the storage engine to distribute across multiple machines. Traditional SQL systems could also be used with a JDBC based spi implementation. Before submitting this change I have implemented six storage systems for the spi layer: * Apache HBase[1] * Apache Cassandra[2] * Google Bigtable[3] * an in-memory implementation for unit testing * a JDBC implementation for SQL * a generic cache provider that can ride on top of memcache All six systems came in with an spi layer around 1000 lines of code to implement the above 7 interfaces. This is a huge reduction in size compared to prior attempts to implement a new JGit storage layer. As this package shows, a complete JGit storage implementation is more than 17,000 lines of fairly complex code. A simple cache is provided in storage.dht.spi.cache. Implementers can use CacheDatabase to wrap any other type of Database and perform fast reads against a network based cache service, such as the open source memcached[4]. An implementation of CacheService must be provided to glue this spi onto the network cache. [1] https://github.com/spearce/jgit_hbase [2] https://github.com/spearce/jgit_cassandra [3] http://labs.google.com/papers/bigtable.html [4] http://memcached.org/ Change-Id: I0aa4072781f5ccc019ca421c036adff2c40c4295 Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
13 年之前
Store Git on any DHT jgit.storage.dht is a storage provider implementation for JGit that permits storing the Git repository in a distributed hashtable, NoSQL system, or other database. The actual underlying storage system is undefined, and can be plugged in by implementing 7 small interfaces: * Database * RepositoryIndexTable * RepositoryTable * RefTable * ChunkTable * ObjectIndexTable * WriteBuffer The storage provider interface tries to assume very little about the underlying storage system, and requires only three key features: * key -> value lookup (a hashtable is suitable) * atomic updates on single rows * asynchronous operations (Java's ExecutorService is easy to use) Most NoSQL database products offer all 3 of these features in their clients, and so does any decent network based cache system like the open source memcache product. Relying only on key equality for data retrevial makes it simple for the storage engine to distribute across multiple machines. Traditional SQL systems could also be used with a JDBC based spi implementation. Before submitting this change I have implemented six storage systems for the spi layer: * Apache HBase[1] * Apache Cassandra[2] * Google Bigtable[3] * an in-memory implementation for unit testing * a JDBC implementation for SQL * a generic cache provider that can ride on top of memcache All six systems came in with an spi layer around 1000 lines of code to implement the above 7 interfaces. This is a huge reduction in size compared to prior attempts to implement a new JGit storage layer. As this package shows, a complete JGit storage implementation is more than 17,000 lines of fairly complex code. A simple cache is provided in storage.dht.spi.cache. Implementers can use CacheDatabase to wrap any other type of Database and perform fast reads against a network based cache service, such as the open source memcached[4]. An implementation of CacheService must be provided to glue this spi onto the network cache. [1] https://github.com/spearce/jgit_hbase [2] https://github.com/spearce/jgit_cassandra [3] http://labs.google.com/papers/bigtable.html [4] http://memcached.org/ Change-Id: I0aa4072781f5ccc019ca421c036adff2c40c4295 Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
13 年之前
Store Git on any DHT jgit.storage.dht is a storage provider implementation for JGit that permits storing the Git repository in a distributed hashtable, NoSQL system, or other database. The actual underlying storage system is undefined, and can be plugged in by implementing 7 small interfaces: * Database * RepositoryIndexTable * RepositoryTable * RefTable * ChunkTable * ObjectIndexTable * WriteBuffer The storage provider interface tries to assume very little about the underlying storage system, and requires only three key features: * key -> value lookup (a hashtable is suitable) * atomic updates on single rows * asynchronous operations (Java's ExecutorService is easy to use) Most NoSQL database products offer all 3 of these features in their clients, and so does any decent network based cache system like the open source memcache product. Relying only on key equality for data retrevial makes it simple for the storage engine to distribute across multiple machines. Traditional SQL systems could also be used with a JDBC based spi implementation. Before submitting this change I have implemented six storage systems for the spi layer: * Apache HBase[1] * Apache Cassandra[2] * Google Bigtable[3] * an in-memory implementation for unit testing * a JDBC implementation for SQL * a generic cache provider that can ride on top of memcache All six systems came in with an spi layer around 1000 lines of code to implement the above 7 interfaces. This is a huge reduction in size compared to prior attempts to implement a new JGit storage layer. As this package shows, a complete JGit storage implementation is more than 17,000 lines of fairly complex code. A simple cache is provided in storage.dht.spi.cache. Implementers can use CacheDatabase to wrap any other type of Database and perform fast reads against a network based cache service, such as the open source memcached[4]. An implementation of CacheService must be provided to glue this spi onto the network cache. [1] https://github.com/spearce/jgit_hbase [2] https://github.com/spearce/jgit_cassandra [3] http://labs.google.com/papers/bigtable.html [4] http://memcached.org/ Change-Id: I0aa4072781f5ccc019ca421c036adff2c40c4295 Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
13 年之前
Store Git on any DHT jgit.storage.dht is a storage provider implementation for JGit that permits storing the Git repository in a distributed hashtable, NoSQL system, or other database. The actual underlying storage system is undefined, and can be plugged in by implementing 7 small interfaces: * Database * RepositoryIndexTable * RepositoryTable * RefTable * ChunkTable * ObjectIndexTable * WriteBuffer The storage provider interface tries to assume very little about the underlying storage system, and requires only three key features: * key -> value lookup (a hashtable is suitable) * atomic updates on single rows * asynchronous operations (Java's ExecutorService is easy to use) Most NoSQL database products offer all 3 of these features in their clients, and so does any decent network based cache system like the open source memcache product. Relying only on key equality for data retrevial makes it simple for the storage engine to distribute across multiple machines. Traditional SQL systems could also be used with a JDBC based spi implementation. Before submitting this change I have implemented six storage systems for the spi layer: * Apache HBase[1] * Apache Cassandra[2] * Google Bigtable[3] * an in-memory implementation for unit testing * a JDBC implementation for SQL * a generic cache provider that can ride on top of memcache All six systems came in with an spi layer around 1000 lines of code to implement the above 7 interfaces. This is a huge reduction in size compared to prior attempts to implement a new JGit storage layer. As this package shows, a complete JGit storage implementation is more than 17,000 lines of fairly complex code. A simple cache is provided in storage.dht.spi.cache. Implementers can use CacheDatabase to wrap any other type of Database and perform fast reads against a network based cache service, such as the open source memcached[4]. An implementation of CacheService must be provided to glue this spi onto the network cache. [1] https://github.com/spearce/jgit_hbase [2] https://github.com/spearce/jgit_cassandra [3] http://labs.google.com/papers/bigtable.html [4] http://memcached.org/ Change-Id: I0aa4072781f5ccc019ca421c036adff2c40c4295 Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
13 年之前
Store Git on any DHT jgit.storage.dht is a storage provider implementation for JGit that permits storing the Git repository in a distributed hashtable, NoSQL system, or other database. The actual underlying storage system is undefined, and can be plugged in by implementing 7 small interfaces: * Database * RepositoryIndexTable * RepositoryTable * RefTable * ChunkTable * ObjectIndexTable * WriteBuffer The storage provider interface tries to assume very little about the underlying storage system, and requires only three key features: * key -> value lookup (a hashtable is suitable) * atomic updates on single rows * asynchronous operations (Java's ExecutorService is easy to use) Most NoSQL database products offer all 3 of these features in their clients, and so does any decent network based cache system like the open source memcache product. Relying only on key equality for data retrevial makes it simple for the storage engine to distribute across multiple machines. Traditional SQL systems could also be used with a JDBC based spi implementation. Before submitting this change I have implemented six storage systems for the spi layer: * Apache HBase[1] * Apache Cassandra[2] * Google Bigtable[3] * an in-memory implementation for unit testing * a JDBC implementation for SQL * a generic cache provider that can ride on top of memcache All six systems came in with an spi layer around 1000 lines of code to implement the above 7 interfaces. This is a huge reduction in size compared to prior attempts to implement a new JGit storage layer. As this package shows, a complete JGit storage implementation is more than 17,000 lines of fairly complex code. A simple cache is provided in storage.dht.spi.cache. Implementers can use CacheDatabase to wrap any other type of Database and perform fast reads against a network based cache service, such as the open source memcached[4]. An implementation of CacheService must be provided to glue this spi onto the network cache. [1] https://github.com/spearce/jgit_hbase [2] https://github.com/spearce/jgit_cassandra [3] http://labs.google.com/papers/bigtable.html [4] http://memcached.org/ Change-Id: I0aa4072781f5ccc019ca421c036adff2c40c4295 Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
13 年之前
Store Git on any DHT jgit.storage.dht is a storage provider implementation for JGit that permits storing the Git repository in a distributed hashtable, NoSQL system, or other database. The actual underlying storage system is undefined, and can be plugged in by implementing 7 small interfaces: * Database * RepositoryIndexTable * RepositoryTable * RefTable * ChunkTable * ObjectIndexTable * WriteBuffer The storage provider interface tries to assume very little about the underlying storage system, and requires only three key features: * key -> value lookup (a hashtable is suitable) * atomic updates on single rows * asynchronous operations (Java's ExecutorService is easy to use) Most NoSQL database products offer all 3 of these features in their clients, and so does any decent network based cache system like the open source memcache product. Relying only on key equality for data retrevial makes it simple for the storage engine to distribute across multiple machines. Traditional SQL systems could also be used with a JDBC based spi implementation. Before submitting this change I have implemented six storage systems for the spi layer: * Apache HBase[1] * Apache Cassandra[2] * Google Bigtable[3] * an in-memory implementation for unit testing * a JDBC implementation for SQL * a generic cache provider that can ride on top of memcache All six systems came in with an spi layer around 1000 lines of code to implement the above 7 interfaces. This is a huge reduction in size compared to prior attempts to implement a new JGit storage layer. As this package shows, a complete JGit storage implementation is more than 17,000 lines of fairly complex code. A simple cache is provided in storage.dht.spi.cache. Implementers can use CacheDatabase to wrap any other type of Database and perform fast reads against a network based cache service, such as the open source memcached[4]. An implementation of CacheService must be provided to glue this spi onto the network cache. [1] https://github.com/spearce/jgit_hbase [2] https://github.com/spearce/jgit_cassandra [3] http://labs.google.com/papers/bigtable.html [4] http://memcached.org/ Change-Id: I0aa4072781f5ccc019ca421c036adff2c40c4295 Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
13 年之前
Store Git on any DHT jgit.storage.dht is a storage provider implementation for JGit that permits storing the Git repository in a distributed hashtable, NoSQL system, or other database. The actual underlying storage system is undefined, and can be plugged in by implementing 7 small interfaces: * Database * RepositoryIndexTable * RepositoryTable * RefTable * ChunkTable * ObjectIndexTable * WriteBuffer The storage provider interface tries to assume very little about the underlying storage system, and requires only three key features: * key -> value lookup (a hashtable is suitable) * atomic updates on single rows * asynchronous operations (Java's ExecutorService is easy to use) Most NoSQL database products offer all 3 of these features in their clients, and so does any decent network based cache system like the open source memcache product. Relying only on key equality for data retrevial makes it simple for the storage engine to distribute across multiple machines. Traditional SQL systems could also be used with a JDBC based spi implementation. Before submitting this change I have implemented six storage systems for the spi layer: * Apache HBase[1] * Apache Cassandra[2] * Google Bigtable[3] * an in-memory implementation for unit testing * a JDBC implementation for SQL * a generic cache provider that can ride on top of memcache All six systems came in with an spi layer around 1000 lines of code to implement the above 7 interfaces. This is a huge reduction in size compared to prior attempts to implement a new JGit storage layer. As this package shows, a complete JGit storage implementation is more than 17,000 lines of fairly complex code. A simple cache is provided in storage.dht.spi.cache. Implementers can use CacheDatabase to wrap any other type of Database and perform fast reads against a network based cache service, such as the open source memcached[4]. An implementation of CacheService must be provided to glue this spi onto the network cache. [1] https://github.com/spearce/jgit_hbase [2] https://github.com/spearce/jgit_cassandra [3] http://labs.google.com/papers/bigtable.html [4] http://memcached.org/ Change-Id: I0aa4072781f5ccc019ca421c036adff2c40c4295 Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
13 年之前
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497
  1. /*
  2. * Copyright (C) 2011, Google Inc.
  3. * and other copyright owners as documented in the project's IP log.
  4. *
  5. * This program and the accompanying materials are made available
  6. * under the terms of the Eclipse Distribution License v1.0 which
  7. * accompanies this distribution, is reproduced below, and is
  8. * available at http://www.eclipse.org/org/documents/edl-v10.php
  9. *
  10. * All rights reserved.
  11. *
  12. * Redistribution and use in source and binary forms, with or
  13. * without modification, are permitted provided that the following
  14. * conditions are met:
  15. *
  16. * - Redistributions of source code must retain the above copyright
  17. * notice, this list of conditions and the following disclaimer.
  18. *
  19. * - Redistributions in binary form must reproduce the above
  20. * copyright notice, this list of conditions and the following
  21. * disclaimer in the documentation and/or other materials provided
  22. * with the distribution.
  23. *
  24. * - Neither the name of the Eclipse Foundation, Inc. nor the
  25. * names of its contributors may be used to endorse or promote
  26. * products derived from this software without specific prior
  27. * written permission.
  28. *
  29. * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND
  30. * CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES,
  31. * INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
  32. * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
  33. * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
  34. * CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
  35. * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
  36. * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
  37. * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
  38. * CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
  39. * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
  40. * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF
  41. * ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
  42. */
  43. package org.eclipse.jgit.storage.dht;
  44. import java.security.MessageDigest;
  45. import java.util.ArrayList;
  46. import java.util.Collections;
  47. import java.util.Comparator;
  48. import java.util.HashMap;
  49. import java.util.List;
  50. import java.util.Map;
  51. import java.util.zip.Deflater;
  52. import org.eclipse.jgit.generated.storage.dht.proto.GitStore;
  53. import org.eclipse.jgit.generated.storage.dht.proto.GitStore.ChunkMeta;
  54. import org.eclipse.jgit.generated.storage.dht.proto.GitStore.ChunkMeta.BaseChunk;
  55. import org.eclipse.jgit.generated.storage.dht.proto.GitStore.ObjectInfo.ObjectType;
  56. import org.eclipse.jgit.lib.AnyObjectId;
  57. import org.eclipse.jgit.lib.Constants;
  58. import org.eclipse.jgit.lib.ObjectId;
  59. import org.eclipse.jgit.storage.dht.spi.Database;
  60. import org.eclipse.jgit.storage.dht.spi.WriteBuffer;
  61. import org.eclipse.jgit.transport.PackedObjectInfo;
  62. import org.eclipse.jgit.util.NB;
  63. /**
  64. * Formats one {@link PackChunk} for storage in the DHT.
  65. * <p>
  66. * Each formatter instance can be used only once.
  67. */
  68. class ChunkFormatter {
  69. static final int TRAILER_SIZE = 4;
  70. private final RepositoryKey repo;
  71. private final DhtInserterOptions options;
  72. private final byte[] varIntBuf;
  73. private final int maxObjects;
  74. private Map<ChunkKey, BaseChunkInfo> baseChunks;
  75. private List<StoredObject> objectList;
  76. private byte[] chunkData;
  77. private int ptr;
  78. private int mark;
  79. private int currentObjectType;
  80. private BaseChunkInfo currentObjectBase;
  81. private PackChunk.Members builder;
  82. private GitStore.ChunkInfo.Source source;
  83. private boolean fragment;
  84. private int objectType;
  85. private int objectsTotal, objectsWhole, objectsRefDelta, objectsOfsDelta;
  86. private ChunkInfo chunkInfo;
  87. ChunkFormatter(RepositoryKey repo, DhtInserterOptions options) {
  88. this.repo = repo;
  89. this.options = options;
  90. this.varIntBuf = new byte[32];
  91. this.chunkData = new byte[options.getChunkSize()];
  92. this.maxObjects = options.getMaxObjectCount();
  93. this.objectType = -1;
  94. }
  95. void setSource(GitStore.ChunkInfo.Source src) {
  96. source = src;
  97. }
  98. void setObjectType(int type) {
  99. objectType = type;
  100. }
  101. void setFragment() {
  102. fragment = true;
  103. }
  104. ChunkKey getChunkKey() {
  105. return getChunkInfo().getChunkKey();
  106. }
  107. ChunkInfo getChunkInfo() {
  108. return chunkInfo;
  109. }
  110. ChunkMeta getChunkMeta() {
  111. return builder.getMeta();
  112. }
  113. PackChunk getPackChunk() throws DhtException {
  114. return builder.build();
  115. }
  116. void setChunkIndex(List<PackedObjectInfo> objs) {
  117. builder.setChunkIndex(ChunkIndex.create(objs));
  118. }
  119. ChunkKey end(MessageDigest md) {
  120. if (md == null)
  121. md = Constants.newMessageDigest();
  122. // Embed a small amount of randomness into the chunk content,
  123. // and thus impact its name. This prevents malicious clients from
  124. // being able to predict what a chunk is called, which keeps them
  125. // from replacing an existing chunk.
  126. //
  127. chunkData = cloneArray(chunkData, ptr + TRAILER_SIZE);
  128. NB.encodeInt32(chunkData, ptr, options.nextChunkSalt());
  129. ptr += 4;
  130. md.update(chunkData, 0, ptr);
  131. ChunkKey key = ChunkKey.create(repo, ObjectId.fromRaw(md.digest()));
  132. GitStore.ChunkInfo.Builder info = GitStore.ChunkInfo.newBuilder();
  133. info.setSource(source);
  134. info.setObjectType(GitStore.ChunkInfo.ObjectType.valueOf(objectType));
  135. if (fragment)
  136. info.setIsFragment(true);
  137. info.setChunkSize(chunkData.length);
  138. GitStore.ChunkInfo.ObjectCounts.Builder cnts = info.getObjectCountsBuilder();
  139. cnts.setTotal(objectsTotal);
  140. if (objectsWhole > 0)
  141. cnts.setWhole(objectsWhole);
  142. if (objectsRefDelta > 0)
  143. cnts.setRefDelta(objectsRefDelta);
  144. if (objectsOfsDelta > 0)
  145. cnts.setOfsDelta(objectsOfsDelta);
  146. builder = new PackChunk.Members();
  147. builder.setChunkKey(key);
  148. builder.setChunkData(chunkData);
  149. if (baseChunks != null) {
  150. List<BaseChunk> list = new ArrayList<BaseChunk>(baseChunks.size());
  151. for (BaseChunkInfo b : baseChunks.values()) {
  152. if (0 < b.useCount) {
  153. BaseChunk.Builder c = BaseChunk.newBuilder();
  154. c.setRelativeStart(b.relativeStart);
  155. c.setChunkKey(b.key.asString());
  156. list.add(c.build());
  157. }
  158. }
  159. Collections.sort(list, new Comparator<BaseChunk>() {
  160. public int compare(BaseChunk a, BaseChunk b) {
  161. return Long.signum(a.getRelativeStart()
  162. - b.getRelativeStart());
  163. }
  164. });
  165. ChunkMeta.Builder b = ChunkMeta.newBuilder();
  166. b.addAllBaseChunk(list);
  167. ChunkMeta meta = b.build();
  168. builder.setMeta(meta);
  169. info.setMetaSize(meta.getSerializedSize());
  170. }
  171. if (objectList != null && !objectList.isEmpty()) {
  172. byte[] index = ChunkIndex.create(objectList);
  173. builder.setChunkIndex(index);
  174. info.setIndexSize(index.length);
  175. }
  176. chunkInfo = new ChunkInfo(key, info.build());
  177. return getChunkKey();
  178. }
  179. /**
  180. * Safely put the chunk to the database.
  181. * <p>
  182. * This method is slow. It first puts the chunk info, waits for success,
  183. * then puts the chunk itself, waits for success, and finally queues up the
  184. * object index with its chunk links in the supplied buffer.
  185. *
  186. * @param db
  187. * @param dbWriteBuffer
  188. * @throws DhtException
  189. */
  190. void safePut(Database db, WriteBuffer dbWriteBuffer) throws DhtException {
  191. WriteBuffer chunkBuf = db.newWriteBuffer();
  192. db.repository().put(repo, getChunkInfo(), chunkBuf);
  193. chunkBuf.flush();
  194. db.chunk().put(builder, chunkBuf);
  195. chunkBuf.flush();
  196. linkObjects(db, dbWriteBuffer);
  197. }
  198. void unsafePut(Database db, WriteBuffer dbWriteBuffer) throws DhtException {
  199. db.repository().put(repo, getChunkInfo(), dbWriteBuffer);
  200. db.chunk().put(builder, dbWriteBuffer);
  201. linkObjects(db, dbWriteBuffer);
  202. }
  203. private void linkObjects(Database db, WriteBuffer dbWriteBuffer)
  204. throws DhtException {
  205. if (objectList != null && !objectList.isEmpty()) {
  206. for (StoredObject obj : objectList) {
  207. db.objectIndex().add(ObjectIndexKey.create(repo, obj),
  208. obj.link(getChunkKey()), dbWriteBuffer);
  209. }
  210. }
  211. }
  212. boolean whole(Deflater def, int type, byte[] data, int off, final int size,
  213. ObjectId objId) {
  214. if (free() < 10 || maxObjects <= objectsTotal)
  215. return false;
  216. header(type, size);
  217. objectsWhole++;
  218. currentObjectType = type;
  219. int endOfHeader = ptr;
  220. def.setInput(data, off, size);
  221. def.finish();
  222. do {
  223. int left = free();
  224. if (left == 0) {
  225. rollback();
  226. return false;
  227. }
  228. int n = def.deflate(chunkData, ptr, left);
  229. if (n == 0) {
  230. rollback();
  231. return false;
  232. }
  233. ptr += n;
  234. } while (!def.finished());
  235. if (objectList == null)
  236. objectList = new ArrayList<StoredObject>();
  237. final int packedSize = ptr - endOfHeader;
  238. objectList.add(new StoredObject(objId, type, mark, packedSize, size));
  239. if (objectType < 0)
  240. objectType = type;
  241. else if (objectType != type)
  242. objectType = ChunkInfo.OBJ_MIXED;
  243. return true;
  244. }
  245. boolean whole(int type, long inflatedSize) {
  246. if (free() < 10 || maxObjects <= objectsTotal)
  247. return false;
  248. header(type, inflatedSize);
  249. objectsWhole++;
  250. currentObjectType = type;
  251. return true;
  252. }
  253. boolean ofsDelta(long inflatedSize, long negativeOffset) {
  254. final int ofsPtr = encodeVarInt(negativeOffset);
  255. final int ofsLen = varIntBuf.length - ofsPtr;
  256. if (free() < 10 + ofsLen || maxObjects <= objectsTotal)
  257. return false;
  258. header(Constants.OBJ_OFS_DELTA, inflatedSize);
  259. objectsOfsDelta++;
  260. currentObjectType = Constants.OBJ_OFS_DELTA;
  261. currentObjectBase = null;
  262. if (append(varIntBuf, ofsPtr, ofsLen))
  263. return true;
  264. rollback();
  265. return false;
  266. }
  267. boolean refDelta(long inflatedSize, AnyObjectId baseId) {
  268. if (free() < 30 || maxObjects <= objectsTotal)
  269. return false;
  270. header(Constants.OBJ_REF_DELTA, inflatedSize);
  271. objectsRefDelta++;
  272. currentObjectType = Constants.OBJ_REF_DELTA;
  273. baseId.copyRawTo(chunkData, ptr);
  274. ptr += 20;
  275. return true;
  276. }
  277. void useBaseChunk(long relativeStart, ChunkKey baseChunkKey) {
  278. if (baseChunks == null)
  279. baseChunks = new HashMap<ChunkKey, BaseChunkInfo>();
  280. BaseChunkInfo base = baseChunks.get(baseChunkKey);
  281. if (base == null) {
  282. base = new BaseChunkInfo(relativeStart, baseChunkKey);
  283. baseChunks.put(baseChunkKey, base);
  284. }
  285. base.useCount++;
  286. currentObjectBase = base;
  287. }
  288. void appendDeflateOutput(Deflater def) {
  289. while (!def.finished()) {
  290. int left = free();
  291. if (left == 0)
  292. return;
  293. int n = def.deflate(chunkData, ptr, left);
  294. if (n == 0)
  295. return;
  296. ptr += n;
  297. }
  298. }
  299. boolean append(byte[] data, int off, int len) {
  300. if (free() < len)
  301. return false;
  302. System.arraycopy(data, off, chunkData, ptr, len);
  303. ptr += len;
  304. return true;
  305. }
  306. boolean isEmpty() {
  307. return ptr == 0;
  308. }
  309. int getObjectCount() {
  310. return objectsTotal;
  311. }
  312. int position() {
  313. return ptr;
  314. }
  315. int size() {
  316. return ptr;
  317. }
  318. int free() {
  319. return (chunkData.length - TRAILER_SIZE) - ptr;
  320. }
  321. byte[] getRawChunkDataArray() {
  322. return chunkData;
  323. }
  324. int getCurrentObjectType() {
  325. return currentObjectType;
  326. }
  327. void rollback() {
  328. ptr = mark;
  329. adjustObjectCount(-1, currentObjectType);
  330. }
  331. void adjustObjectCount(int delta, int type) {
  332. objectsTotal += delta;
  333. switch (type) {
  334. case Constants.OBJ_COMMIT:
  335. case Constants.OBJ_TREE:
  336. case Constants.OBJ_BLOB:
  337. case Constants.OBJ_TAG:
  338. objectsWhole += delta;
  339. break;
  340. case Constants.OBJ_OFS_DELTA:
  341. objectsOfsDelta += delta;
  342. if (currentObjectBase != null && --currentObjectBase.useCount == 0)
  343. baseChunks.remove(currentObjectBase.key);
  344. currentObjectBase = null;
  345. break;
  346. case Constants.OBJ_REF_DELTA:
  347. objectsRefDelta += delta;
  348. break;
  349. }
  350. }
  351. private void header(int type, long inflatedSize) {
  352. mark = ptr;
  353. objectsTotal++;
  354. long nextLength = inflatedSize >>> 4;
  355. chunkData[ptr++] = (byte) ((nextLength > 0 ? 0x80 : 0x00) | (type << 4) | (inflatedSize & 0x0F));
  356. inflatedSize = nextLength;
  357. while (inflatedSize > 0) {
  358. nextLength >>>= 7;
  359. chunkData[ptr++] = (byte) ((nextLength > 0 ? 0x80 : 0x00) | (inflatedSize & 0x7F));
  360. inflatedSize = nextLength;
  361. }
  362. }
  363. private int encodeVarInt(long value) {
  364. int n = varIntBuf.length - 1;
  365. varIntBuf[n] = (byte) (value & 0x7F);
  366. while ((value >>= 7) > 0)
  367. varIntBuf[--n] = (byte) (0x80 | (--value & 0x7F));
  368. return n;
  369. }
  370. private static byte[] cloneArray(byte[] src, int len) {
  371. byte[] dst = new byte[len];
  372. System.arraycopy(src, 0, dst, 0, len);
  373. return dst;
  374. }
  375. private static class BaseChunkInfo {
  376. final long relativeStart;
  377. final ChunkKey key;
  378. int useCount;
  379. BaseChunkInfo(long relativeStart, ChunkKey key) {
  380. this.relativeStart = relativeStart;
  381. this.key = key;
  382. }
  383. }
  384. private static class StoredObject extends PackedObjectInfo {
  385. private final int type;
  386. private final int packed;
  387. private final int inflated;
  388. StoredObject(AnyObjectId id, int type, int offset, int packed, int size) {
  389. super(id);
  390. setOffset(offset);
  391. this.type = type;
  392. this.packed = packed;
  393. this.inflated = size;
  394. }
  395. ObjectInfo link(ChunkKey key) {
  396. GitStore.ObjectInfo.Builder b = GitStore.ObjectInfo.newBuilder();
  397. b.setObjectType(ObjectType.valueOf(type));
  398. b.setOffset((int) getOffset());
  399. b.setPackedSize(packed);
  400. b.setInflatedSize(inflated);
  401. return new ObjectInfo(key, b.build());
  402. }
  403. }
  404. }