Performance: larger MDX files are unmanagably slow to parse about micromark

[robsimmons]

Just to close the loop, here are those two graphs again with a fresh npm install that grabs the updated subcomponent-2.0.1:

[robsimmons]

It's actually MDX - the content is all in a <DummyComponent />. The MDX usage is probably a red herring for the performance issues, but the place where I was encountering 60-minute parse times came from a file that had the structure of >90% of the content in large template strings passed to MDX components.

[robsimmons]

I suspect this is related to #23 - the code spends almost 100% of its time in this while loop.

While this is synthetic data, but I came across this issue through Astro, as the MDX-parsing step for twelf.org was taking 60 minutes due to a 500kb generated markdown file. We have a workaround for that, but even on much smaller files MDX parsing is the slowest part of the site building process.

[wooorm]

Why do you generate a 500kb markdown file, and call it .mdx instead of .md 🤔

[github-actions]

Hi! This was closed. Team: If this was fixed, please add phase/solved. Otherwise, please add one of the no/* labels.