The memcached support was extracted from PR #98 by @jbellis, with cleanups to
bring it in line with current APIs and style guide.

This PR also addresses issue #326.

* Handles missing "java" with an error message.
* Adds a check_output method for users running on Python 2.6.
* Prints an error when importing argparse fails (third party module on Python 2.6).

This reverts commit 0e5866eb.

Restores #372

Add a DB client for Google Cloud Datastore. This initial commit includes implementation of all CRUD operations (modulo scan), and integration into the rest of YCSB (pom.xml changes, etc)

Next few commits will include Scan/query support, single entity group support, and better documentation.

Currently doesn't include unit tests since we don't have mock objects for that.

get_classpath_from_maven(database) now does not alter the project name
before giving it to Maven. main() now calculates the db_dir with core
and uses the core project for Maven, otherwise things work as they did
before.

fixes #336

HBase 0.94 requires an additional dependency on hadoop-core, but has no
source-level changes to the client.

hbase/ now contains the HBase v0.9 binding code.
hbase1/ now contains the HBase 1.0 binding code.

1) fix handling of later binding versions, e.g. hbase-10
2) fix reading conf dir for binding-specific artifacts
3) allow running from source checkout by asking maven for artifact locations

YCSB now tracks what's in a given release using release notes.

ref:
https://github.com/brianfrankcooper/YCSB/releases

the classpath, rather than all of them.

classpath, rather than all of them.

http://tarantool.org and https://github.com/tarantool/tarantool
In "mail.ru" we wrote and widely use Tarantool key-value database.
It's key properties include:

* Defferent index types with iterators:
    - HASH (the fastest)
    - TREE (range and ordered retreival)
    - BITSET (bit mask search)
    - RTREE (geo search)
* multipart keys for HASH and TREE indexes
* Data persistence with by Write Ahead Log (WAL) and snapshots.
* asynchronous master-master replication, hot standby.
* coroutines and async. IO are used to implement high-performance lock-free access to data.
    - socket-io/file-io with yeilds from lua
* stored procedures in Lua (Using LuaJIT)
* supports plugins written on C/C++ (Have two basic plugins for working with MySQL and PostgreSQL)
* Authentication and access control

Move 'distribution target' to the end (it's needed for .jar to be in the .tar.gz)

closes #191

* Use argparse for argument parsing.
* Add a '-jvm-args' flag, to pass arbirary options to java.
* Exit with the return code from java.

Requires Python 2.7.

closes #136

Signed-off-by: Sean Busbey <sean.busbey@gmail.com>

This includes changes to main pom for the v1.2.9 cassandra driver.

closes #145

Allow setting the java version to use via the "JAVA_HOME" environment
variable. Fixes #252.

* Bumps `hbase.version` to 1.0.0
* Adds a new database, hbase-10, which targets the HBase 1.0 API.

Based on YCSB++

http://www.pdl.cmu.edu/ycsb++/

I have fixed a bug in insert() and delete() methods.There are also
some cosmetic changes to readme and properties files.