Markdown incorrectly processing HTML

[dkharrat]

Not sure if this is another bug that popped up after the fix for issue #96, but looks like I found another bug in processing inline HTML. The following HTML snippet reproduces the bug:

This line is correctly interpreted as regular markdown

<div>
<div>
test
</div>

<h2>Header</h2>
<ol>
   <li>
      <p>First paragraph</p>
   </li>
   <li>
      <p>Second paragaph</p>
   </li>
</ol>
</div>

generates:

<p>This line is correctly interpreted as regular markdown</p>

<div>
<div>
test
</div>

<p><h2>Header</h2>
<ol>
   <li>
      <p>First paragraph</p>
   </li>
   <li>
      <p>Second paragaph</p>
   </li>
</ol>
</div></p>

Notice that an extra <p> tag is generated before the <h2> tag and after the last </div> tag, resulting in an invalid HTML.

[dkharrat]

Any update on this one?

[mildsunrise]

For the record: this isn't a bug, it's a consequence of how Sundown looks for the ending tag.

In order to tell Sundown where your tag really ends, you should indent the HTML, or at least the first </div>:

This line is correctly interpreted as regular markdown

<div>
<div>
test
  </div>

<h2>Header</h2>
<ol>
   <li>
      <p>First paragraph</p>
   </li>
   <li>
      <p>Second paragaph</p>
   </li>
</ol>
</div>

[dkharrat]

Good to know about the workaround. But from the perspective of the library user, I'd consider this a bug as valid markdown input is producing invalid html output. Shouldn't Sundown just not process markdown inside HTML tags instead of relying on indentation? According to the Markdown spec:

Note that Markdown formatting syntax is not processed within block-level HTML tags. E.g., you can’t use Markdown-style emphasis inside an HTML block.

[mildsunrise]

Well, that's actually from the spec. The spec says the ending tag should be on an unindented line, and be followed by empty lines.

So, Sundown searches for the first </div> tag matching these requeriments (which is the wrong tag, but Sundown doesn't actually parse the HTML).