Files
John MacFarlane 13e1b49224 HTML reader: Fix handling of invalidly nested sublists.
This revises the fix to #8150 (and the test case) and closes #9187.

HTML in the (invalid) form:

    <ul>
      <li>L1</li>
      <ul>
        <li>L1.1</li>
      </ul>
    </ul>

is treated by browsers like

    <ul>
      <li>L1
        <ul>
          <li>L1.1</li>
        </ul>
      </li>
    </ul>

not

    <ul>
      <li>L1
      <li><ul>
          <li>L1.1</li>
        </ul>
      </li>
    </ul>

as pandoc previously assumed.

This change will give a similar treatment to

    <ul>
      <li>L1</li>
      <p>foobar</p>
    </ul>

which also seems to match browser behavior.
2023-11-15 14:40:00 -08:00

594 B

Nested bullet lists

% pandoc -f html -t markdown
<ul>
	<li>L1</li>
	<li>L2</li>
	<ul>
		<li>L3.1</li>
		<li>L3.2</li>
	</ul>
  <li>L4</li>
</ul>
^D
-   L1
-   L2
    -   L3.1
    -   L3.2
-   L4

Nested ordered lists

% pandoc -f html -t markdown
<ol>
	<li>L1</li>
	<li>L2</li>
	<ol>
		<li>L3.1</li>
		<li>L3.2</li>
	</ol>
</ol>
^D
1.  L1
2.  L2
    1.  L3.1
    2.  L3.2

Ordered list nested below an unordered list

% pandoc -f html -t markdown
<ul>
	<li>L1</li>
	<li>L2</li>
	<ol>
		<li>L3.1</li>
		<li>L3.2</li>
	</ol>
</ul>
^D
-   L1
-   L2
    1.  L3.1
    2.  L3.2