Markdown to HTML and vice versa

[peter-oms]

Hi,

This is perhaps a question for @wooorm as I have been reading the "HTML and remark" recipe page and I saw your name against it.

Using the recipe code, convering HTML to Markdown and then converting the resulting Markdown back to HTML does not result in the original HTML.

See this CodeSandbox for an example of this behaviour - starting with the HTML string <p><strong>Hello </strong>world</p>, I end up with <p>**Hello **world</p> after the two conversions.

I think the problem comes from converting the initial HTML to Markdown - I end up with **Hello **world, but this is probably not valid Markdown (space inside the bold tag), so when converting back to HTML it does not think it should add the <strong> tag.

I couldn't see any relevant options that could be applied to any of the rehype or remark packages to fix this behaviour. Have you had a similar issue before? Any ideas how to fix it?

[wooorm]

The problem is that markdown can’t really represent that HTML. As you noted. There is a nice way around it: improve the HTML you have. Don’t put that space there.

There is one automatic way around it, but it is extremely ugly. It’s to turn:

<strong>Hello </strong>world

Into

**Hello **world

Which renders as: Hello world

I’ve laid the solution out in https://github.com/orgs/syntax-tree/discussions/60#discussioncomment-2111096. It’s quite complex to wrap your head around though. Solving it would go in https://github.com/syntax-tree/mdast-util-to-markdown.