Subprocess pipelines

Create complex pipeline stages and run the pipeline configuration
using a single I/O configuration based on the usual Input, Output,
and Error options provided by the existing run() functions.

Each stage can be either a process or a Swift function. For processes
there are the typical executable, arguments, or configuration
parameters. There is an additional process option to rewire standard
output and standard error in typical ways, such as merging the two
to standard output, or replacing standard output with standard error.

A Swift function stage can read from its input, and write to its
output/error streams much like a process, but provides the full
Swift power in the current process to parse JSON and do other powerful
things. Similarly, an exit code can be returned.

Provide a rich set of top-level functions, methods and overloaded
operators to construct and execute the pipelines.

Author: cmcgee1024