[brendankenny]

Right now, if there are list items on a page and the rendered font size is large, distill-page tries to turn the text into headings. So

<ol style="font-size: 26px;">
  <li>Text</li>
</ol>

turns into

1. ## Text

This gets really messed up with e.g. inline <code>, where

<ol style="font-size: 26px;">
  <li>The <code>distill-page</code> skill</li>
</ol>

gets turned into

1. ## The
  ## `
  ## distill-page
  ## `
  ## skill

To fix this, like in insideParagraph, this change makes it so large text in a list item isn't automatically a heading. You can still get headings from explicit heading elements in the list item, though.

To do that, another parser state variable—insideListItem—is added. AFAICT insideParagraph could have been successfully re-used insetad, and that would cut down on parser states, but overloading the state like that seemed hacky. It could also be renamed into some new variable (insideNonAutoHeadingContent), but maybe that's too generalized for what could end up being just two states to handle this.

Also added a test page for it. Kind of slow to start yet another chromium instance, but let me know if that should be done a different way.

(also, also, no idea what's going on with CDP Page.getAnnotatedPageContent on mock-page.html, it keeps overwriting with local paths and failing, but it was doing that before I made these changes too :)

[paulirish]

lovely thank yoU!