Discussion: drop csslint support

[thetric]

Stylus supports Stylelint and CSSLint as CSS Linter. While the former seems actively maintained (releases, contributions), CSSLint seems to be dead (no release since December 2016, no commits since February 2019, no status updates from the maintainers, older question for project status). Does it make sense to drop support for the unmaintained CSSLint?

[tophf]

We're using our own fork of CSSLint. You can report issues with csslint in this repo, but first search for an existing report.

[eight04]

We are still using the parser from CSSLint to parse the code so it can't be dropped directly. Do stylelint have a parser? I've heard that it uses postcss under the hood but I never learn how it works.

[tophf]

Even if we won't use csslint's parser, our fork of csslint itself is useful for standard CSS because it's several times faster than stylelint and provides a minimum level of linting to prevent typos or to check parentheses and other basic stuff. Actually, if I weren't so lazy I would add the ability to check stylus/less syntax to our csslint and then try to convince everyone to get rid of stylelint.

Edit: turns out our csslint can be 10+ times faster than stylelint thanks to reusing the previous run's results so continuously making trivial edits (typing, changing colors) in GitHub-Dark usercss will take ~600ms in stylelint (blergh) but only ~60ms in csslint.