Move transformer into mdast-util-breaks

[remcohaszing]

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)

Problem

The transformer can work on the AST without unified. It may be useful to use this as a plain function.

Solution

Move the transformer into a new package mdast-util-breaks. This package would be part of syntax-tree.

Alternatives

I chose the name to match the name remark-breaks, but it may make sense to use something more verbose, e.g.: mdast-util-break-newlines.

[ChristianMurphy]

Sounds good to me, thanks @remcohaszing!
the more verbose name mdast-util-break-newlines sounds better to me, but interested to hear others thoughts.

[wooorm]

Yeah, sure.

Why not mdast-util-hard-break? mdast-util-automatic-break? mdast-util-auto-hard-break? mdast-util-newline-to-break? (optionally with s suffix)

I’m not sure what break-newlines means, it feels “inverted”.

I think mdast-util-newline-to-break is my preference.

[remcohaszing]

I’m not sure what break-newlines means, it feels “inverted”.

It works if you read “break” as a verb.

Anyway, I like mdast-util-newline-to-break too. I’ll go with that. I’ll get to it later this week.

What license should I use? Since it’s going to be mostly based on this repo, copy the license as-is?

[wooorm]

That could be an explanation, but then I think that’s a bit weird, wouldn’t you break the lines instead? break-lines?

What license should I use?

Yep, reuse the one seems fine.

---

I think you could also go with:

/**
 * @typedef {import('mdast').Content} Content
 * @typedef {import('mdast').Root} Root
 * @typedef {import('mdast-util-find-and-replace').ReplaceFunction} ReplaceFunction
 */

/**
 * @typedef {Content | Root} Node
 */

import {findAndReplace} from 'mdast-util-find-and-replace'

/**
 * Turn normal line endings into hard breaks.
 *
 * @param {Node} tree
 *   Tree to change.
 * @returns {void}
 *   Nothing.
 */
export function newlineToBreak(tree) {
  findAndReplace(tree, /\r?\n|\r/g, replace)
}

/**
 * Replace line endings.
 *
 * @type {ReplaceFunction}
 * @param {string} value
 */
function replace(value) {
  return [{type: 'break'}, {type: 'text', value}]
}

This is what I have locally now in rehype-github-*

AFAIK it is mostly equivalent: because line endings can never occur at the start or end of a paragraph, or next to each other (blank lines).

Though this findAndReplace method would split the line endings from the next text? Maybe findAndReplace should be updates to merge initial/final text node results?

Looking at the current algo, it actually drops all positional info on text nodes? 🤔