Nextflow

- Added support for Java 8
- Upgraded to Groovy 2.3.0
- Upgraded to Kryo 2.24.0
- New splitter operators: 'splitText', 'splitFasta', 'splitFastq', 'splitCsv'
- New operator 'collectFile' to gather file chunks
- New operator Channel.fromPath which allows fine control on paths traversal
- Operator Channel.from has been deprecated in favor of Channel.fromPath
- Operator Channel.watch has been deprecated in favor of Channel.watchPath
- Clean-up API for operators: 'into', 'choice', 'separate', 'filter'
- Operator 'grep' has been removed and features included in 'filter' operator
- Deprecated 'chopXxx' operators
- Deprecated 'chunkXxx' operators
- Deprecated 'each' operator in favor of 'subscribe'
- Deprecated process 'merge' directive
- Executor service now uses a fixed thread pool
- Fixed issue on 'maxForks' process directive
- Added 'info' cli parameter
- Improved documentation

- Enhanced stability and errors reporting

- Enhanced support for Docker containers:
Now the current user id is added on the Docker command line
and /tmp folder in the container is mounted on a temporary path in the host disk

- Fixed an issue with 'scratch' directive i.e. Fixed issue when copying a symlink
- Added 'Retry' error handling strategy and 'maxRetries' and 'maxErrors' process directives
- Added GridGain cluster sub-module
- Added TaskDispatcher submit, start, complete and error listeners
- Added new ServiceDiscovery to handle sub-modules executors
- Added MapgetOrCreate() extension method
- Using CompileStatic for AST transformation
- Removing unused APP_TMP_DIR const
- Renamed class 'AbstractExecutor' to 'Executor'
- Added ProcessFactory class
- Refactored sub-modules directory structure
- Define the default 'lib' path relative to the pipeline 'baseDir' i.e. relative to the main script location
- Save the history entry just before launch the execution

- New self-downloadable and self-updatable script launcher
- New cluster mode, Nextflow can run as a daemon node by using Hazelcast cluster (incubating)
- New support for Docker containers. Any process can be executed through Docker adding 'container = name' in the process definition (incubating)
- New native processes (groovy closure) can executed in the cluster.
- Enhanced task polling monitor
- Use 'rsync' to upstage task result files to target directory
- Pipeline execution can be launched in background by adding the option '-bg- to the command line
- Updated documentation

- Added 'concat' operator to concatenate two or more channels
- Added 'into(ch1,ch2...)' operator which maps tuple elements into the corresponding target channel
- Implemented queue status check for SLURM grid executor
- Implemented queue status check for LSF grid executor
- Improved grid executors stability
- Fixed a bug when setting setting the monitor queue size