Originally posted by @fregante in #3678 (comment):

I'd like to see the most lightweight Markdown parser being run directly to the description text to store HTML so the entire parseDescription function can be dropped.

Parser example: Holixus/nano-markdown
bundlephobia.com/result?p=nano-markdown@1.2.1

I've done a little comparison between multiple lightweight Markdown parsers, and I don't think we can directly use any of them 🙁

Features they should support:

1. code span with single backtick: `` ` ``
2. code span with a backticked word: `` `bla` ``
3. inline URLs: https://example.org or <https://example.org>
4. don't escape HTML tags: **bold** or <b>bold</b>

Notes:

1. not distributed via npm, 2.8kB unminified, about 1.5kB minified
2. not distributed via npm, 1.0kB unminified, 0.8kB minified