Don't call System.exit over uncaught exceptions from shutdown hooks

Fixes a broken link in documentation to issue tracker

Added a method to create PartitionPruningRDD.

Ganglia port change

Switch spark_ec2.py to use the new spark-ec2 scripts. 

Move ParallelCollection into spark.rdd package.

Make CoGroupedRDDs explicitly have the same key type.

If combineByKey is using the same partitioner, skip the shuffle.

include jobid in Executor commandline args

STREAMING-50: Support transform workaround in JavaPairDStream

Update issue tracker link in contributing guide

This ports a useful workaround (the `transform` function) to
JavaPairDStream. It is necessary to do things like sorting which
are not supported yet in the core streaming API.

SPARK-694: All references to [K, V] in JavaDStreamLike should be changed to [K2, V2] 

SPARK-696: sortByKey should use 'ascending' parameter

to lack of memoization