Hi Gophers!
The API has solidified, so I've just released
github.com/splunk/pipelines at v1.0.0. This library provides a lightweight way to manage concurrency separately from your business logic by breaking a concurrent task down into stages in a pipeline, connected by nothing more than plain old channels.
Features:* Pipeline stages based on common functional programming primitives.
* Integrated with termination signals from the context package.
* Scale out individual pipeline stages without disrupting business logic.
* Introduce queues by buffering channels mid-stream.
* Split and combine pipeline stages as pairs.
* Handle fatal and non-fatal errors from any pipeline stage.
* Sinks for draining the result of a pipeline computation.
Some future enhancements are planned. Please leave any suggestions or issues in the issue tracker.