what exactly is the issue?

exports is fully supported, unless you are on an older version since it is pretty recent

and no, you don't need to enable it. it is on by default.

The required JS dependency "unist-util-visit-parents/do-not-use-color" is not available, it was required by "node_modules/unist-util-visit/node_modules/unist-util-visit-parents/lib/index.js". Happens with the latest version of shadow-cljs

The do-not-use-color is just specified in exports of the package-file. https://github.com/syntax-tree/unist-util-visit-parents/blob/58e20d8366363b8b6e22d2e9c82d2ded408ecf16/package.json#L34C19-L34C19 I'm not familiar with how that works tbh

This happens pretty much for all react libraries that work with markdown as they all use unist-util it seems

and which version of that package do you have installed?

as in node_modules/unist-util-visit-parents/package.json

I've tried a few. With the latest version of react-markdown, it takes in unist-util at "^5.0.0"

and which version is it currently?

Let me check, one sec

I mean if I look at it only v6 has this exports field

so if you have 5.x the error is completely legit, as that file doesn't exist

You are absolutely correct. It seems like a very widely used library, so I don't get how this is not a problem for ones in the JavaScript ecosystem

dependency conflicts are very common in JS world last time I checked 😉

Thank you for helping me out

actually, there seems to be a problem. There are just many dependencies with almost the same name, so it was a bit hard to spot. "unist-util-visit-parents": "^6.0.0" is a dependency of unist-util-visit. And unist-util-visit-parents is where the problem occurs

unist-util-visit is at version 5.0.0, so react-markdown is effectively using that either way with ^5.0.0

I'm not sure I follow

npm dependency conflict resolution is basically non existent and sometimes needs a hand

so your constellation of dependencies just might end up getting the wrong version somehow

this is nothing shadow-cljs can do anything about or work around, so you have so resolve it

@U0B1SDL67 did you find a solution? seems like it's an ESM related issue?

No. I wrote a parser myself instead