Suppose one places an empty figure tag

<figure></figure>

I get an error

TypeError: Cannot read property 'type' of undefined
    at parseMarkdown (test\node_modules\@tufte-markdown\remark-figure-parser\dist\remark-figure-parser.cjs.js:15066:52)
    at Of.figureTokenizer (test\node_modules\@tufte-markdown\remark-figure-parser\dist\remark-figure-parser.cjs.js:15040:22)
    at Of.tokenize [as tokenizeBlock] (test\node_modules\remark-parse\lib\tokenizer.js:58:18)
    at Of.parse (test\node_modules\remark-parse\lib\parse.js:33:20)
    at Function.parse (test\node_modules\unified\index.js:271:45)
    at pipelineParse (\test\node_modules\unified\index.js:23:16)
    at wrapped (\node_modules\trough\wrap.js:25:19)
    at next (test\node_modules\trough\index.js:58:24)
    at Object.run (test\node_modules\trough\index.js:32:10)
    at executor (test\node_modules\unified\index.js:365:16)

I am not sure of how the AST works but I am guessing you need to test for result.children array not being empty????

By convention markdown syntax inside block elements is not processed. And remark respects those conventions (see rehypejs/rehype-raw#8).

A number of items such as <footer> inside blockquotes, epigraphs etc. do not work as expected when they contain markdown (for some reason beyond my understanding markdown inside <figure> tag works).

Even in your own tests some items are not being rendered correctly. See for example, output for section.md in tests:

> [It is] notable that the Feynman lectures (3 volumes) write about all of
> physics in 1800 pages, using only 2 levels of hierarchical headings: chapters
> and A-level heads in the text. It also uses the methodology of
> <em>sentences</em> which then cumulate sequentially into <em>paragraphs</em>,
> rather than the grunts of bullet points. Undergraduate Caltech physics is very
> complicated material, but it didn't require an elaborate hierarchy to
> organize.
>
> <footer>
> [Edward Tufte, forum post, 'Book design: advice and examples' thread][quote-cite]
> </footer>

[quote-cite]: http://www.edwardtufte.com/bboard/q-and-a-fetch-msg?msg_id=0000hB

is rendered as:

<blockquote>
<p>[It is] notable that the Feynman lectures (3 volumes) write about all of
physics in 1800 pages, using only 2 levels of hierarchical headings: chapters
and A-level heads in the text. It also uses the methodology of
<em>sentences</em> which then cumulate sequentially into <em>paragraphs</em>,
rather than the grunts of bullet points. Undergraduate Caltech physics is very
complicated material, but it didn’t require an elaborate hierarchy to
organize.</p>
<footer>
[Edward Tufte, forum post, 'Book design: advice and examples' thread][quote-cite]
</footer>
</blockquote>

Notice that the footer remains unprocessed!

We either need to change conventions in conversation with @jez or hope that remark supports non-standard syntax through a plugin.

Thanks for the great work on this library! I'm in the process of wiring up to the eleventy static site generator as a replacement for the built-in markdown parser, and everything seems to be working well. One hitch, though: Markdown documents with no second-level headings never get the necessary section tag as a wrapper, and the layout files. For example:

console.dir(parse('Opening paragraph\n## Heading **strong** *emphasis*'))

// Output:
// <section><p>Opening paragraph</p></section> <section><h2 id="heading-strong-emphasis">Heading <strong>strong</strong> <em>emphasis</em></h2></section>

console.dir(parse('Opening paragraph\n\nSecond paragraph'))

// Output:
// <p>Opening paragraph</p> <p>Second paragraph</p>

I'm not sure if this is intentional; in theory it's possible to check the output for a missing <section>, but it seems like adding the wrapper makes sense even if there's no second-level heading?