Post

Ooooh, the Stylelint team are removing some dependencies and in the process also making it easier to run Stylelint in the browser. github.com/stylelint/st...

Feel free to open an issue if there were specific dependencies that made it hard to run Stylelint in the browser :)

Thanks, I was going to give this another go after the next release! I think my main were that I'd like some specific non-node entry points to avoid all the globbing/IO etc. I can't remember exactly, but I will write it up.

You could try with the v18 branch (if you have time). Will be a while until we start planning a v19 so now is a good window to request such changes :)

Oooooh That's actually a pretty good idea.

I think I've nailed it down to these points (with a bunch of help from Claude): - Linting always starts from the `stylelint` global, no official way to skip I/O; also true for `createPlugin()` which leads ever - Several modules import Node API's to get EOL and process.env which trips the browser

What's the best way to approach this? Append finding to github.com/stylelint/st..., open 1 new issue or several issues?

Yeah, might be best to explain your use case there and list any new pain points that haven't been mentioned yet. Maybe some issues have already been resolved on the v18 branch? Stylelint will obviously never be browser first. The main use case is local/ci where we can assume a node env

So can't make any changes that make the primary case slower or worse. But maybe there are small affordance possible? Changes that aren't observable otherwise but do make it easier to run in a browser.

Yeah, I think some small affordances could make the public API more open to composition, pick-and-choose which entry points you need. But also, it could open up possibilities to run (parts of) stylelint.io/demo/ in the browser without web containers.