roundtripping mixed html and text results in html

[ChristianMurphy]

Subject of the issue

any text
    <div>more text

is stringified as:

any text
<div>more text

which changes the structure

Your environment

• OS: Ubuntu
• Packages: mdast-util-to-markdown 0.6.2
• Env: node v15.5.1, npm 7.3.0

Steps to reproduce

parse

any text
    <div>more text

which has the structure

{
    "type": "root",
    "children": [
        {
            "type": "paragraph",
            "children": [
                {
                    "type": "text"
                },
                {
                    "type": "html"
                },
                {
                    "type": "text"
                }
            ]
        }
    ]
}

and stringify it:

any text
<div>more text

the resulting markdown has a different structure than the original

{
    "type": "root",
    "children": [
        {
            "type": "paragraph",
            "children": [
                {
                    "type": "text"
                }
            ]
        },
        {
            "type": "html"
        }
    ]
}

📓 comparing how the two pieces of markdown text are being parsed with https://spec.commonmark.org/dingus it appears in both cases it is parsed as expected.

Expected behavior

structure is the same

Actual behavior

structure is different

[wooorm]

A similar case would probably be:

a
    ```b
    ```

a
```b
```

(as in, the overlap between things that could turn into blocks, and have a different meaning when found inline).

Escaping wouldn’t solve this, but perhaps the eol can be swapped for a space

[wooorm]

the code case seems to be prevented by cf7cf8f though, so I’m guessing this only happens to code

[wooorm]

Here are some more examples for all HTML cases (https://spec.commonmark.org/0.29/#html-blocks) that should have these problems:

c1
    <script></script>

c1
<script></script>

c2
    <!---->

c2
<!---->

c3
    <??>

c3
<??>

c4
    <!A b>

c4
<!A b>

c5
    <![CDATA[]]>

c5
<![CDATA[]]>

c6
    <div>

c6
<div>

c7 (cannot interrupt a paragraph)