v1.1.2

This is mainly a bug fix release that fixes #57, but it also contains a few small optimizations in the frame decoding.

v1.1.1

Bug fix release:

• Fixes #55 (you couldn't connect if the hostname you specified resolved to multiple addresses and Cassandra didn't bind to the first one).

v1.1.0

v1.1 contains most of the features that didn't make it into v1.0, but that you expect from a driver for a distributed database. It makes it possible for the application to carry on working in the face of node failures, and use new nodes as they are added to the cluster.

These are the major new features added since v1.0:

• Automatic peer discovery: when connecting the client will automatically discover the whole Cassandra cluster and connect to all nodes in the same data centers as the seed nodes.
• Automatic reconnection: when nodes are restarted, fail and come back again, or new nodes join the cluster, the client will automatically connect to them.
• Configurable number of connections per node: a client can optionally connect multiple times to each node, in order to increase the maximum number of concurrent requests.
• Optional per-request timeouts.
• Configurable default consistency per client.
• Prepared statements now use all connections and will automatically be prepared on new connections.
• The client can log significant events to a user-defined logger.
• Official support for Rubinius 2.0.

There has also been many bug fixes, especially in the serialization and deserialization code, and optimizations and general speed improvements.

v1.1.0.rc0

These are the significant new features that have been added over the last nine prereleases:

• Automatic peer discovery: when connecting the client will automatically discover the whole Cassandra cluster and connect to all nodes in the same data centers as the seed nodes.
• Automatic reconnection: when nodes are restarted, fail and come back again, or new nodes join the cluster, the client will automatically connect to them.
• Configurable number of connections per node: a client can optionally connect multiple times to each node, in order to increase the maximum number of concurrent requests.
• Optional per-request timeouts.
• Configurable default consistency per client.
• Prepared statements now use all connections and will automatically be prepared on new connections.
• The client can log significant events to a user-defined logger.

There has also been many bug fixes, especially in the serialization and deserialization code, and optimizations and general speed improvements.

v1.1.0.pre8

Fixes a deadlock in Future#value in JRuby (unknown if it was present in MRI).

v1.0.6

• Fixes an issue with UUIDs with leading zeroes and prepared statements.

v1.1.0.pre7

• Optionally connect multiple times to each node
• Optional per-request timeouts

v1.1.0.pre6

• Prepared statements now use all connections.
• The default consistency level is now configurable per client instance.
• Overhaul and optimizations in the (private) futures implementation.
• Fixes an issue with UUIDs with leading zeroes and prepared statements.
• Applied a fix for serializing counter updates that was applied to v1.0 but never to v1.1.

v1.1.0.pre5

This release has been removed, it was accidentally created from the wrong branch.

See v1.1.0.pre6.

v1.0.5

• UUIDs no longer raises an error when compared to objects of other types (thanks to @christophersansone)
• Fixes an issue where a request could deadlock under special circumstances
• Fixes an issue that prevented updating counters in prepared statements (thanks to @brackxm)