I've recently split oniguruma-to-es's parser/tokenizer/traverser into the new dedicated library oniguruma-parser. It includes lots of improvements, including a new Oniguruma optimizer that I think is a perfect fit for this library (in fact I built it primarily for Shiki). 😊

As of now, the optimizer includes 18 optimization transforms, but many more can be added in the future. tm-grammars will get continually-improving regex minification when bumping dependencies.

Here's how you'd want to call it in this library:

import { optimize } from 'oniguruma-parser/optimizer'

const pattern = '...'
const optimized = optimize(pattern, {
  rules: {
    // Follow `vscode-oniguruma` which enables this Oniguruma option by default
    captureGroup: true,
  },
}).pattern

Some notes:

• Although oniguruma-parser supports a few more Oniguruma features than oniguruma-to-es, there are still a handful of features it doesn't yet support. So if optimize throws an error, you should probably ignore it and leave the original pattern unaltered. As of now, I think this will only happen for one regex in the Swift grammar.
• Any existing code in this library that is changing the regexes can/should be removed when you start using optimize. It will take care of existing minification like removing comments and free-spacing with flag x.
• Since the optimizer and the Oniguruma code generator it uses are both brand new, it probably makes sense to introduce this in a PR so I can review the regex changes before you land them.