Current implementation of pipeline cannot implement difficult recursion (only simple one-node problems). This feature will add ability to create auxiliary subpipelines, that can be launched from the arbitrary node (as blockable call or via loopbacks). All processed data will not be mixed, so each node can be ensured, that received data stream is processed input data stream, that was sent by node.
Subpipelines will be added to pipeline by constructor and will be accessed by its name. Creation of each subpipeline will be similar to pipeline creation (via Wrappers).
Closed Dec 29, 2009 at 10:21 AM by mak_
It works!