Markdown inside <p></p> is not parsed

[vedmant]

Describe the bug

Markdown inside <p></p> is not parsed

To Reproduce
Steps to reproduce the behavior:

1. Marked Demo
2. CommonMark Demo

Expected behavior

Should parse markdown inside html <p> tag.

[UziTech]

This follows the CommonMark spec:

An HTML block is a group of lines that is treated as raw HTML (and will not be escaped in HTML output).

If you put a blank like between the <p> tags and the content it will be parsed as markdown.

demo

<p>

fdsfsd **test** sdfsdfs

</p>

Result:

fdsfsd test sdfsdfs

Or you can continue using HTML

<p>fdsfsd <strong>test</strong> sdfsdfs</p>

[vedmant]

@UziTech I see, thanks. Would be nice to have an option to force parse markdown inside html tags. It will solve lots of issues with markdown in WYSIWYG editors.

[UziTech]

You could create a renderer for that.

Something like:

const marked = require('marked');

// Override html renderer
const renderer = {
  html(html) {
    // pull out inner markdown
    const match = html.match(/^(<[^>]+>)(.*)(<[^>]+>)$/);
    // run markdown through marked again
    return match[1] + marked(match[2]) + match[3];
  }
};

marked.use({ renderer });

console.log(marked('<p>asdf **test** asdf</p>'));
// <p><p>asdf <strong>test</strong> asdf</p></p>

[neopostmodern]

@UziTech I'm trying to create a render like you mentioned, but can't figure it out. Consider the following input:

<div>
  <h1>Content</h1>
  like, *markdown*;  
  or, [links](//example.com)
</div>

The first iteration of your modified renderer runs fine and strips away the <div> and </div>. But the remaining content

  <h1>Content</h1>
  like, *markdown*;  
  or, [links](//example.com)
  

gets passed into the HTML renderer override — and can't be matched since it's a mix of HTML and markdown and there is no wrapping HTML tag. I would have thought the tokenizer takes care of this?

Any hints?

The last version where HTML can be mixed seamlessly with markdown seems to be 0.3.1, so I'll have to downgrade for now.

[UziTech]

@neopostmodern my code above is just a proof of concept, you will have to change it to work with your use case. There are a lot of edge cases that need to be thought of to make it work with any markdown.