Raw datapoint output is quite useful and important for two reasons:
 1) Allows the user to do more thorough statistical analysis on the complete set of the data.
 2) Using timestamps that are logged with raw data points, it becomes much easier to investigate and correlate latency outliers at the backend service.

pull latest from master

merge from YCSB to my fork

[jdbc] added a README

Fixes #459

Adds a Cassandra 2 CQL binding.

[git] added standard intellij omissions to .gitignore

Add Kudu bindings

Fixes #451

Rename README file to README.md file for database bindings.

Rename README file to README.md file for the following DB bindings so the build process can generate binding specific tar balls for them. (currently they are broken because the build process looks for README.md)

dynamodb
hypertable
mapkeeper
nosqldb

[core] added configurable percentile values

Fixed #336. Classpath handling BasicDB from source

[docs] Update README for 0.4.0 release.

Closes #409

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

In OneMeasurementHdrHistogram, now reading hdrhistogram.percentiles
from workload configuration. This property is a comma delimited string
of Integers to use as percentile values.

This implements #437

[licensing] Added ASLv2 headers to files without them for v0.4.0. Removed old copy of Javadoc.

Fix 90th percentile reported as 95th

Fixes #438

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

[hbase10] Skip HBase MiniCluster tests on Windows

fixes #434 

[core] Update the AcknowledgedCounterGenerator to use a power of two and mask.

size for the window so a mask can be used instead of a modulo operation.
This avoids issues when there is overflow of the id integer values to
negative values.