I'm somewhat following along with your work here. However, without a PureScript source file to use as my mental model for what's being input into the codegen, I'm not sure what the corefn you have stored in a String represents.

Thoughts on duplicating what's done in purs-backend-es to test the codegen of this backend?

Hello Mr functor.

I'm currently also working on a Chez Scheme backend and wonder if it might be worth combining efforts.

The way I've structured my project is to basically keep the CoreFn -> Chez Scheme implementation written in Chez Scheme to maintain simplicity and to take advantage of Lisp's superior code generation capabilities. A downside of this is that Chez Scheme has a barebones echo system of libraries so I've worked to implement everything I need to make the transpiler easy to implement; I probably won't need more than chezscheme-match and chezscheme-json (currently being cleaned up after a refactor) though which should be ready to be used.

A separate project I'm working on is a Chez Scheme IDE extension for Visual Studio Code. Two of the main features are interactive debugging and unit test exploration. Anyway, one idea I had was to allow a CoreFn -> Chez Scheme transpiler to output instrumented code such that the Chez Scheme IDE debugger could work seamlessly for PureScript as well.

I'm looking forward to hear more about what your plans are for this project and to see if there's room for cooperation.

// Regards
Björn Kihlberg