[Feature] Split ELisp core out, starting a general lisp interpreter

[wgxh-cc]

Sometimes maybe it's possible for a ELisp program to run separately from the Emacs runtime, and just use some general features, such as doom Emacs which uses a headless Emacs to manage its configuration. The question is, can there be a separate system, which just perform ELisp code as what Emacs ELisp core does, with some runtime-specified features removed or controlled.

The ideal construction is to extract the whole ELisp module out, making it independent of Emacs core features. That means the facility will become a fresh language from scratch, a base, and Emacs is the plug, customizing its behaviors. When returning to external environment, Emacs could also provide a slightly different one from the core.

This may do many goods in many aspects. First, decoupling means lower complexity and higher maintainability, contributing to a cleaner Emacs core. Then, a independent runtime increase the versatility, making ELisp available everywhere; this set up a consistent and comfortable environment for Emacs users. Last but not least, it's easier for developer to adjust their code, debug their programs and test it out.

That's just of my personal opinions. I'd appreciate that if it do any tricks!

[eval-exec]

Thanks for the suggestion — this is a good direction and it aligns with our NeoVM architecture.

Neomacs is already splitting the Elisp runtime into a separate core (Rust neovm-core) with a host boundary, so a headless/standalone host is possible in principle.

That said, our current priority is strict GNU Emacs compatibility first. We are not trying to fork Elisp into a new language with different semantics right now, and the standalone runtime surface is not yet stable.

So: yes to decoupling and headless use as a long-term goal; near-term we’ll continue focusing on compatibility and core parity.