Emphasis takes precedence over inline code span

[eush77]

Input:

_`_`_

Actual AST:

{
  "type": "paragraph",
  "children": [
    {
      "type": "emphasis",
      "children": [
        {
          "type": "text",
          "value": "`"
        }
      ]
    },
    {
      "type": "text",
      "value": "`_"
    }
  ]
}

Expected AST:

{
  "type": "paragraph",
  "children": [
    {
      "type": "emphasis",
      "children": [
        {
          "type": "inlineCode",
          "value": "_"
        }
      ]
    }
  ]
}

Commonmark:

Code span backticks have higher precedence than any other inline constructs except HTML tags and autolinks. Thus, for example, this is not parsed as emphasized text, since the second * is part of a code span: *foo*``.

[wooorm]

This kinda sucks; due to the legacy of marked this isn’t really do-able currently.

I’d love to make this happen but it will need a rewrite of the parsing algorithm; stepping away from regexes and to a real algorithmic tokeniser, e.g., like markdown-it 😦

[eush77]

I came across the test that freezes the wrong precedence of inlineCode against linkReference.

In Commonmark,

[the `]` character](/url)

should parse as

<a href="/url">the <code>]</code> character</a>

[wooorm]

Nice catch @eush77! In fact, I think all of the “pending” tests in mdast-html have to do with this (I wrote a bit about it here).

I think the only way to fix this is by rewriting the engine to use code tokenisers in favour of regexes; it’ll take a while to rewrite everything though!

[wooorm]

Let’s track this in GH-306.