Default hadoop version is now hadoop2
Generate specific pom avoids version collisions with dependencies
Updated poms for new release scripts

Conflicts:
	flink-addons/flink-avro/pom.xml
	flink-addons/flink-hadoop-compatibility/pom.xml
	flink-addons/flink-hbase/pom.xml
	flink-addons/flink-jdbc/pom.xml
	flink-addons/flink-spargel/pom.xml
	flink-addons/flink-tachyon/pom.xml
	flink-quickstart/flink-quickstart-java/pom.xml
	flink-quickstart/flink-quickstart-scala/pom.xml

 - Add missing entries for dependencies
 - Remove entries for removed dependencies
 - Change DEPENDENCIES file to point to src and bin LICENSE and NOTICE file
 - Add comments in some files to clarify whether the are Flink original,
   or from which third party file they are adapted
 - Change some links from incubator to TLP

  - Move sling JSON dependency to streaming connectors (prev streaming core)
  - Exclude YARN API from transitive mapreduce-core dependencies we use the dependency
    only for .mapred and .mapreduce interfaces
  - Remove some unnecessary dependencies (junit in java8)
  - Manage Kryo dependency
  - Cleanup minor POM warnings.

 - Remove old unused figure image files
 - Add a file clarifying that figures are licensed under ASL 2.0

[runtime tests] Fixes TaskManagerTest.testRunJobWithForwardChannel which caused spurious errors. Doubles the timeout for tests.

This closes #296.

This closes #209

This closes #287

- Fix for paths with blank characters
- Fix for single entry wildcard classpaths (http://bugs.java.com, bugID: 7146578)

Remove dup code by delegating call from RemoteExecutor#executePlan to RemoteExecutor#executePlanWithJars.

Author: Henry Saputra <henry.saputra@gmail.com>

Closes #307 from hsaputra/reduce_dup_code_in_executePlan and squashes the following commits:

a98f3b6 [Henry Saputra] Remove dup code by delegating call from RemoteExecutor#executePlan to RemoteExecutor#executePlanWithJars.

This would print a warning if a type could not be analyzed by the Scala
TypeAnalyzer. In such a case analysis is done using the Java
TypeExtractor but the warning seems to confuse people.

[FLINK-1380] [streaming] Updated SplitDataStream to extend DataStream to get rid of selectAll method for splits

This closes #254.

This closes #293

Before, something like:
def f[T: TypeInformation](data: DataSet[T]) = {
  val tpe = createTypeInformation[(T, Seq[T])]
  println("Type: " + tpe)
}

f(Seq(1.0f, 2.0f)

would fail because the type extractor could not re-use existing
TypeInformation for nested types.

This closes #279

This closes #268

to HadoopFileSystem wrapper, which supports all subclasses of org.apache.hadoop.fs.FileSystem.
This allows us to let users use all file systems with support for HDFS.
The change has been tested with Tachyon, Google Cloud Storage Hadoop Adapter and HDFS.

The change also cleans up the Hadoop dependency exclusions.

Also contains fixes by sewen@apache.org
 - Make MissingTypeInfo optional in TypeExtractor (by default still throws exception)
 - Simplified deferred evaluation of type dependend code by making evaluations lazy
 - Add call location function names to MissingTypeInfo error messages.
 - Improvements on other error messages.

This closes #270