You can turn off a rule by setting its config value to null
.
For example, to use stylelint-config-standard
without the at-rule-empty-line-before
rule:
{
"extends": "stylelint-config-standard",
"rules": {
"at-rule-empty-line-before": null
}
}
You can also turn off a rule for specific sections of your CSS. Refer to the rules section of the configuration guide for more information.
Refer to the CLI section of the docs.
The CLI can also be used from within npm run scripts to use a non-global installation of stylelint.
lint-staged is a NodeJS script that supports running stylelint against Git staged files.
The stylelint community maintains a handful of plugins for popular task runners, including ones for gulp, webpack, Broccoli and Grunt. Refer to their individual READMEs to get started.
If there isn't a dedicated stylelint plugin for your task runner of choice, you can use stylelint as a PostCSS plugin and make use of PostCSS's numerous task runner plugins.
There are also examples of using the PostCSS plugin via the PostCSS JS API within the docs.
However, using stylelint as a PostCSS plugin limits your reporting options to postcss-reporter. We recommend using the stylelint CLI or Node API, instead, for better reporting.
The stylelint community also maintains a handful of plugins for popular editors. Refer to their individual READMEs to get started.
stylelint can parse any the following non-standard syntaxes by default: SCSS, Less and SugarSS. Non-standard syntaxes can automatically be inferred from the following file extensions .scss
, .less
, and .sss
; or else you can specify the syntax yourself.
Additionally, stylelint can accept any PostCSS-compatible syntax when using the CLI or Node API. Note, however, that stylelint can provide no guarantee that core rules will work with syntaxes other than the defaults listed above.
Refer to the docs on how to configure stylelint to parse non-standard syntaxes.
Should I lint before or after processing my stylesheets through PostCSS plugins or other processors?
We recommend linting your source files before any transformations.
Create a processor or use an existing one that extracts CSS from your HTML's <style>
tags and feeds it into stylelint.
Use the --fix
CLI flag or the fix
Node API option to fix a number of stylistic violations with this experimental feature.
Each rule stands alone, so sometimes it's possible to configure rules such that they conflict with one another. For example, you could turn on two conflicting blacklist and whitelist rules, e.g. unit-blacklist
and unit-whitelist
.
It's your responsibility as the configuration author to resolve these conflicts.
A rule must meet the criteria set out in the developer guide, including being applicable to only standard CSS syntax, and having a clear and unambiguous finished state. Whereas a plugin is a rule or sets of rules built by the community that don't adhere to the criteria. It might support a particular methodology or toolset, or apply to non-standard constructs and features, or be for specific use cases.
For example, we've found that rules to enforce property order, property groupings, etc., work better as plugins, because there are so many different opinions about what to enforce, and how. When you write or use a plugin, you can make sure to enforce your own particular preferences, exactly; but a rule that tries to address too many divergent use-cases becomes a mess.
Plugins are easy to incorporate into your configuration.
Yes, you can either use the message
secondary option or write your own formatter.
Use the stylelint-selector-bem-pattern plugin to ensure your selectors conform to a chosen BEM-flavor pattern.
You can also take advantage of the selector-*
rules to ban certain types of selectors (e.g. id selectors) and control specificity.
If you're using SUITCSS, you might want to use their shareable config.
a { color: red; }
/** ↑
* Declaration blocks like this */
Use the block-opening-brace-newline-after
and block-opening-brace-newline-before
rules together. For example, this config:
{
"block-opening-brace-newline-after": ["always"],
"block-closing-brace-newline-before": ["always"]
}
Would allow:
a {
color: red;
}
But not these patterns:
a { color: red;
}
a {
color: red; }
a { color: red; }
To allow single-line blocks but enforce newlines with multi-line blocks, use the "always-multi-line"
option for both rules.
Use the regex that corresponds to your chosen convention:
- kebab-case:
^([a-z][a-z0-9]*)(-[a-z0-9]+)*$
- lowerCamelCase:
^[a-z][a-zA-Z0-9]+$
- snake_case:
^([a-z][a-z0-9]*)(_[a-z0-9]+)*$
- UpperCamelCase:
^[A-Z][a-zA-Z0-9]+$
e.g. for lowerCamelCase class selectors, use "selector-class-pattern": "^[a-z][a-zA-Z0-9]+$"
.
All these patterns disallow CSS identifiers that start with a digit, two hyphens, or a hyphen followed by a digit.
Use the defaultSeverity
configuration option.
A user can require()
any file in your npm package, so all you need to do is document which paths point to configs (e.g. require('my-package/config-2')
).
Refer to this section of the docs that explains how the *-empty-line-before
rules work.
If I use extends
within my configuration object, will the options for each rule be merged or overridden?
They will be overridden.
The extends
mechanism is detailed within the configuration docs:
When one configuration extends another, it starts with the other's properties then adds to and overrides what's there.
The reason for this design is documented in #1646.