link rel= detection misses some links

[tmcw]

I think this is the code that's looking for a rel=me twitter account:

shovel/lib/parsers/html.ts

Line 61 in a847935

const match2 = html.match(

My website does have such a meta tag, but it's all HTML5'd out - uses optional quotes and an implied self-closing tag, so this doesn't get caught:

<link href=https://mastodon.social/@tmcw rel=me>

Probably this is one of the more tricky variations to catch, but ideally different attribute order is okay with shovel's parser.

[jmduke]

Yup, we should absolutely use node-html-parser (which we already have depped) for this. T/Y for flag!