allows adding scheme code + ci testing

[Alan-Chen99]

part 1 of splitting #1

using .scm files instead have
++ allow using language servers and linters (one need to run them under lilypond; try my fork at https://github.com/Alan-Chen99/scheme-lsp-server)
++ better error messages and stack traces
+ better organization
+ namespaces
- no longer single file, harder to distribute

I think its overall a big improvement. what do you think?

[PaulMorris]

Hi, thanks for your PR(s) and thanks for splitting these changes out into their own PR. It will make them easier for me to review.

no longer single file, harder to distribute

This is a big consideration for me. For users it is really nice to have everything in a single file. Of course the benefits to these changes are also quite nice, but they are primarily for providing a better developer experience.

Might there be some way we could make these changes but also keep the single file workflow for users? Like a way to 'compile' the code down to a single file built artifact? It seems like that might be more possible with scheme code (where code is data)?

[Alan-Chen99]

I dont think a built artifact is easy to use in the perspective of the user. what if they tried to find documentation/ usage example of a command? thats going to be easiest if they have a copy of a sourcecode. it also wont be fun for the user to see that an error occurred in a nonexistent file.

If the user is running lilypond locally, then downloading a folder is not too different from downloading a file i think. if they are running lilybin or something online, afaik none of them allow including files yet.

What do you think is the way most users are using clairnote?