A project-wide js-linting tool
yarn add stricter --dev
yarn stricter
You can run yarn stricter --help
for help.
Stricter uses stricter.config.js
to read configuration.
The configuration file will be resolved starting from the current working directory location, and searching up the file tree until a config file is (or isn't) found.
module.exports = {
root: 'src',
rulesDir: 'rules',
exclude: /\.DS_Store/,
plugins: ['tangerine'],
resolve: {
modules: [path.resolve(__dirname, 'src'), 'node_modules'],
},
rules: {
'hello-world-project': {
level: 'error'
},
'stricter/unused-files': [{
level: 'warning',
include : [/foo\.*/, /bar\.*/],
exclude : (i) => i.includes('testFolder'),
config: {
entry: [
/foo\.eslintrc\.js/,
/foo\.*\.md/,
/foo\/bar\/index\.js/,
/foo\/baz\/index\.js/,
],
relatedEntry: [
/foo\.*spec\.js/,
/foo\.*test\.js/,
/foo\.*story\.js/,
]
}
}],
'stricter/circular-dependencies': [{
level: 'error',
config: {
checkSubTreeCycle: true,
registries: ['**/foo/bar', 'baz'],
},
}],
'tangerine/project-structure': {
level: 'error',
config: { ... },
}
}
}
root
- string
required, root folder for the project.
rulesDir
- string | string[]
, folder(s), containing custom rules. Rule files need to follow naming convention <rulename>.rule.js
. They will be available for configuration as <rulename>
.
exclude
- RegExp | RegExp[] | Function
, regular expressions to exclude files, uses relative path from root or function accepting relative path and returning boolean
plugins
- string[]
, packages that contain third-party rule definitions that you can use in rules
. See Plugins for more details.
resolve
- Object
, if you are using webpack, and you want to pass custom resolution options to stricter
, the options are passed from the resolve
key of your webpack configuration.
rules
- required, an object containing configuration for rules.
The keys should be rule names and values should be an object, array of objects or a function. Arrays will result in the rule being executed once per each entry in the array, see rule functions for more info on that syntax. The objects (RuleObject
) should be of the form:
level
- error | warning | off
, log levelinclude
- RegExp | RegExp[] | Function
, regular expressions to match files, uses relative path from root or function accepting relative path and returning booleanexclude
- RegExp | RegExp[] | Function
, regular expressions to exclude from matched files, uses relative path from root or function accepting relative path and returning booleanconfig
- any
, config to be passed into rulepackages
- string[]
, an array of globs that match paths to packages if you are in a multi-package repo. This can be used to override the default list of packages that are provided to rules configured using a function.
The default way of configuring rules is to provide objects, however, each rule value may also be a function that returns an object instead. This provides an easy way to configure rules designed to be executed against each package separately in a monorepo.
Signature: (args: { packages: string[] }) => RuleObject | RuleObject[]
where packages
is a list of package directory paths in your project that are automatically detected by searching for package.json's in sub-directories, i.e. */**/package.json
.
To override where packages
are searched, you can use the top-level packages
config to provide an array of globs instead. For example, this could be sourced from the yarn workspaces field in your project's root package.json.
E.g.
module.exports = {
...
rules: {
'package-structure': ({ packages }) => packages.map(pkg => ({
level: 'error',
config: {
pkgRoot: pkg,
},
})),
'rule-that-does-not-need-to-execute-multiple-times': {
level: 'error',
...
}
},
...
}
Here, the package-structure
rule enforces a specific structure for a package and takes the root path of the package as a config argument. In a single-package repo, the rule
can just use the object syntax and specify the root path of the project as the package root.
However, in a multi-package repo this rule should be executed against each package separately rather than once at the root of the project so the function syntax can be used.
Checks for circular dependencies in the code. Has a configuration to additionally check for cycles on folder level with ability to exclude particular directory from check by providing path to it in registries.
'stricter/circular-dependencies': {
config: {
checkSubTreeCycle: Boolean, // true to check for folder-lever cycles
registries?: string[] | string, // Optional: values should be a glob
}
}
Checks for unused files.
entry
- application entry points. Usually these files are mentioned in entry
part of webpack config or they are non-js files you want to keep (configs, markdown, etc.)
relatedEntry
- related entry points, they are considered used only if any of its dependencies are used by an entry
or its transitive dependencies. Usually these are tests and storybooks.
'stricter/unused-files': {
config: {
entry: RegExp | RegExp[] | Function; // if function, will get file path as an argument
relatedEntry: RegExp | RegExp[] | Function; // if function, will get file path as an argument
}
}
A rule is a javascript module that exports an object that implements the following interface
interface RuleDefinition {
onProject: ({
config?: { [prop: string]: any; };
dependencies: {
[fileName: string]: string[];
};
files: {
[fileName: string]: {
ast?: () => any;
source?: string;
};
};
rootPath: string;
include?: RegExp | RegExp[] | Function;
exclude?: RegExp | RegExp[] | Function;
}) => (string | { message: string; fix?: () => void; })[];
}
onProject
will be called once with files
and dependencies
calculated for current project.
rootPath
is an absolute path to project root.
config
is an optional object that may be specifified in configuration.
onProject
should return an array of objects, describing violations and their fixes, or an empty array if there is none.
include
value of include
from the rule
exclude
value of exclude
from the rule
Options:
--help Show help [boolean]
--version Show version number [boolean]
--config, -c Specify config location [string]
--reporter, -r Specify reporter [choices: "console", "mocha", "junit"]
--rule Verify particular rule [array]
--clearCache Clears cache
--fix Apply fixes for rule violations
Stricter supports consuming rule definitions from other packages by specifying them in the plugins
field of your stricter config.
The package names of plugins must be named stricter-plugin-<name>
, e.g. stricter-plugin-tangerine
. This guarantees unique
rule names across different plugins.
In the plugins
field, you can specify the plugin using its short name <name>
or its long form stricter-plugin-<name>
. You can then
enable and configure rules from a plugin by specifying the rules in the rules
field.
When configuring rules from a plugin, they must be prefixed by their short plugin name <name>/<ruleName>
, e.g. tangerine/project-structure
.
e.g.
// stricter.config.js
module.exports = {
root: '.',
plugins: ['tangerine'],
rules: {
'tangerine/project-structure': {
level: 'error',
config: {...},
}
}
}
To create a stricter plugin, ensure the package name is of the format stricter-plugin-<name>
.
The main file of the package should then export a rules
key that contains the rule definitions you wish to provide.
e.g.
module.exports = {
rules: {
'project-structure': {
onProject: (...) => {...}
},
'another-rule': {
onProject: (...) => {...}
}
}
}
Note that the rule names should not be prefixed when defining them inside the plugin, they are only prefixed when specifying them in configuration.
Rules provided by a plugin are not enabled by default, they must be configured by the end-user. If you would like to provide a preset configuration of rules provided by your plugin, simply export your preset configuration under a certain key. Consumers can then import that configuration and spread it into the rules
field of their stricter config.
E.g.
// stricter-plugin-tangerine/index.js
module.exports = {
// This key can be arbitrarily named
config: {
'tangerine/project-structure': {
level: 'error',
config: {
'.': {
'package.json': { type: 'file' },
'src': { type: 'dir' }
}
}
}
},
rules: {
'project-structure': {
onProject: (...) => {...}
},
}
}
// stricter.config.js
// This import key `config` must match what is exported by the plugin
const { config: tangerineConfig } = require('stricter-plugin-tangerine');
module.exports = {
root: '.',
plugins: ['tangerine'],
rules: {
...tangerineConfig,
},
};
Parses files and returns an object, containing filenames as keys and array of their dependencies as values.
const parseDependencies = (
files: string[],
{ useCache = false, resolve = {} } = { useCache: false, resolve: {} }
): {
[fileName: string]: string[];
}
useCache
- pass true
to leverage stricter
filesystem cache
resolve
- Object
, if you are using webpack, and you want to pass custom resolution options to stricter
, the options are passed from the resolve
key of your webpack configuration
usage:
const { parseDependencies } = require('stricter');
const dependencies = parseDependencies(['foo.js', 'bar.js']);
It helps to use src/debug.ts
as an entry point for debugging.
A sample launch.json for VS Code might look like
{
"version": "0.2.0",
"configurations": [
{
"name": "Current TS File",
"type": "node",
"request": "launch",
"args": ["${relativeFile}"],
"env": { "TS_NODE_FILES": "true" },
"runtimeArgs": ["-r", "ts-node/register"],
"cwd": "${workspaceRoot}"
}
]
}