Should we add other types of downstream plugins?

[Mouvedia]

e.g. stylelint-webpack-plugin, vite-plugin-stylelint, esbuild-plugin-stylelint, etc.

---

• stylelint-webpack-plugin
• nx-stylelint

[romainmenke]

These are packages that run Stylelint for you, integrated into some other tool or task runner, right?

If these are testable and add meaningful coverage, then sure :)
Always good to have more insights before releasing.

[Mouvedia]

nx is not a bundler but maybe nx-stylelint could be considered?

[romainmenke]

Same answer I guess :)

The goal of the project (as I understood it) is to make rolling out releases safer and more controlled.

We try to attain that goal by testing downstream dependents so that unexpected side-effects of changes can surface before a release.

I don't think we need to test every or even many downstream dependents.
That is both unrealistic and unsustainable.

By testing a few downstream dependents, those that are simple and easy to test, we gain a lot of assurances before a release. But the returns diminish sharply with each extra package we test.

If something can be trivially added to the existing list and adds some meaningful test coverage, then that is always nice.

If something is complex then I think we might want to avoid adding it here.

[Mouvedia]

But the returns diminish sharply with each extra package we test.

Then let's add the ones that have tests for now.
i.e. stylelint-webpack-plugin (for rspack and webpack) and nx-stylelint

---

off-topic

I am kind of surprised that we don't have an unplugin-stylelint.

[Mouvedia]

I wish we had somehow added support for vite-plugin-stylelint; it would have probably had picked up ModyQyW/vite-plugin-stylelint#61 earlier.