2e1. Scripting in NeoHaskell
While coding the NeoHaskell CLI, I’ve found myself in a resistance while doing so. I think I’ve found the culprit:
A CLI build tool is essentially a script multiplexer, and writing a script (a series of steps) as a reactive event-based application doesn’t make much sense.
There are a couple solutions right now:
- Create an abstraction layer over the event-sourcing architecture, so it allows the user to define
Task
s orStep
s. Could take a closer look at how Roc does this or even Elm. - Drop using the event-based model for the CLI tool for now, and just rely on basic
IO
, with the regular tooling.
I think that the latter is the most sensible option for now, as the first one is an additional effort that would delay the CLI even more. And we can always add it on top later.