Add retext-case-police to list of plugins

[JulianCataldo]

Initial checklist

• I read the support docs
• I read the contributing guide
• I agree to follow the code of conduct
• I searched issues and couldn’t find anything (or linked relevant results below)
• If applicable, I’ve added docs and tests

Description of changes

Repository

A retext plugin for checking popular names casing.

Examples:

• ⚠️ macbook → ✅ MacBook
• ⚠️ MacOS → ✅ macOS

For:

• Abbreviates
• Brands
• General
• Products
• Softwares

---

Dictionaries are from case-police.

[wooorm]

Cool, looks useful!

• I recommend a message of the form Expected casing of `${correctOne}` instead of `${original}`
  • Puts what’s expected in front
  • Uses markdown (code) for “literal” values
• I recommend adding actual, expected, and url fields to messages.
  • actual is supposed to hold the actual value in the document: toString(nodes)
  • expected is supposed to hold the suggestion (an array of strings or one string): correctOne
  • url can go to documentation, preferrably specific to the ruleId
• I’m not sure why you’re checking if (nodes[0].type === 'WordNode') {? I don’t think that’s every false?

[wooorm]

good catch :)

[wooorm]

• You might also want to support options.ignore, which similar plugins supports too. E.g., https://github.com/retextjs/retext-passive#optionsignore
• case-police seems to do some fancy things on which casings they include and which they ignore: https://github.com/antfu/case-police/blob/6cfa330a55e987038d1a618cb163ebadfa8fbbae/src/utils.ts#L46-L54. Maybe good to add some tests for that?

[JulianCataldo]

• I’m not sure why you’re checking if (nodes[0].type === 'WordNode') {? I don’t think that’s every false?

Good question.
I've added it after trial and error, for making ESLint happy 🤔 Even if it feels redundant. Smash that squiggles 😄

[wooorm]

This can likely be solved by using nodes[0] as the place, instead of delving into it.

A word like don’t, consists of text("don"), punctuation("’"), text("t"). Right now you are warning about don, but the warning is instead about the whole word?

[JulianCataldo]

Oh, neat! I wasn't conscious of what you are describing.
So ESLint was right bothering me, even if for me, it wasn't clear why ^^

Did a quick try, and it does address this, indeed, while being more robust, even if the children[0] didn't seems to cause side-effects. I always observed singleton in node[0], maybe because case-police dictionaries are alphabetical only, without any embedded punctuation in tokens.

thx 👍