What is replacing soft line breaks with ensp?

[bwklein]

In the tree output from the toHast function there is a string like

lineBreak\u000bThis line contains a shift-enter soft line-break here\u000bThis bit is after the line break.

This string contains the "U+000B Line Tabulation" control character.

This is replaced by   in the HTML code, after running it through the following code:

let html = unified()
  .use(rehypeStringify, { collapseEmptyAttributes: true })
  .stringify(tree);

Do you know how I can preserve this control character or replace it with the HTML hex reference ?

[ka2n]

This is rehypeStringify behavior. You can replace the tree before passing to unified() process.

As a side node, my solution was replacing it with <br />:

function convertVerticalTabToBr(node: RootContent): void {
  visit(node, "text", function (node, index, parent) {
    if (node.type === "text") {
      const textNode = node as Text
      if (textNode.value.includes("\u000b")) {
        const texts = textNode.value
          .split("\u000b")
          .flatMap((item) => [{ ...textNode, value: item }, h("br")])
          .slice(0, -1)
        if (index !== null && parent) {
          // remove and insert
          parent.children.splice(index, 1, ...texts)
        }
      }
    }
  })
}