I want to support the project 100%, but the friction (unclear tooling) right now is a challenge

@richiardiandrea: if you use yarn/`npm` scripts as your task runner -- how do you kick off a lumo build without it being a global dep?

do you git clone ... && yarn install and then install lumo as a global dep and then you run yarn scripts?

sorry if it sounds like I'm just complaining 😕

I have a global lumo yes, but you can yarn install and use it locally with no issue... can't you?

@johnjelinek npm install lumo-cljs?

does that just download the binary or does it run through the boot build, which depends on the JDK?

I'll give it a shot though 🙂

Downlods the binary

There’s no JVM involved at all

ok cool, so @anmonteiro for apps you build in lumo, do you use npm/`yarn` tasks in place of boot too?

Uh

I actually just use lumo

Yeah, and you can rely on caching compiled code 🙂

You can write Lumo scripts to compile Lumo apps

Or Lumo scripts for basically anything

But you can put those in your package.json scripts entry too

so you built a task runner in lumo to build/test/package your lumo app?

not a “task runner”

just a plain CLJS script

a plain cljs script that has a collection of tasks like gulp or boot would?

that’s an option

but you’re overcomplicating IMO

so, here's what I have: I have a features directory with a bunch of Gherkin I have a step_definitions folder with cljs that map to the Gherkin features what I want is to run a lumo task to temporarily build the cljs and then run ./node_modules/bin/cucumber on the directory to see if tests pass or fail

so, would I make a separate test.clj for this task and have lumo execute it? or add a yarn script? or use boot because I'm sure in the future there will be other preprocessing devdependencies being used

to you, which path provides the least complications?