Linux OS X Windows Coverage Downloads
Build Status Windows Build Status Coverage Status npm module downloads per month
# ignore `ignore` is a manager, filter and parser which implemented in pure JavaScript according to the [.gitignore spec 2.22.1](http://git-scm.com/docs/gitignore). `ignore` is used by eslint, gitbook and [many others](https://www.npmjs.com/browse/depended/ignore). Pay **ATTENTION** that [`minimatch`](https://www.npmjs.org/package/minimatch) (which used by `fstream-ignore`) does not follow the gitignore spec. To filter filenames according to a .gitignore file, I recommend this npm package, `ignore`. To parse an `.npmignore` file, you should use `minimatch`, because an `.npmignore` file is parsed by npm using `minimatch` and it does not work in the .gitignore way. ### Tested on `ignore` is fully tested, and has more than **five hundreds** of unit tests. - Linux + Node: `0.8` - `7.x` - Windows + Node: `0.10` - `7.x`, node < `0.10` is not tested due to the lack of support of appveyor. Actually, `ignore` does not rely on any versions of node specially. Since `4.0.0`, ignore will no longer support `node < 6` by default, to use in node < 6, `require('ignore/legacy')`. For details, see [CHANGELOG](https://github.com/kaelzhang/node-ignore/blob/master/CHANGELOG.md). ## Table Of Main Contents - [Usage](#usage) - [`Pathname` Conventions](#pathname-conventions) - See Also: - [`glob-gitignore`](https://www.npmjs.com/package/glob-gitignore) matches files using patterns and filters them according to gitignore rules. - [Upgrade Guide](#upgrade-guide) ## Install ```sh npm i ignore ``` ## Usage ```js import ignore from 'ignore' const ig = ignore().add(['.abc/*', '!.abc/d/']) ``` ### Filter the given paths ```js const paths = [ '.abc/a.js', // filtered out '.abc/d/e.js' // included ] ig.filter(paths) // ['.abc/d/e.js'] ig.ignores('.abc/a.js') // true ``` ### As the filter function ```js paths.filter(ig.createFilter()); // ['.abc/d/e.js'] ``` ### Win32 paths will be handled ```js ig.filter(['.abc\\a.js', '.abc\\d\\e.js']) // if the code above runs on windows, the result will be // ['.abc\\d\\e.js'] ``` ## Why another ignore? - `ignore` is a standalone module, and is much simpler so that it could easy work with other programs, unlike [isaacs](https://npmjs.org/~isaacs)'s [fstream-ignore](https://npmjs.org/package/fstream-ignore) which must work with the modules of the fstream family. - `ignore` only contains utility methods to filter paths according to the specified ignore rules, so - `ignore` never try to find out ignore rules by traversing directories or fetching from git configurations. - `ignore` don't cares about sub-modules of git projects. - Exactly according to [gitignore man page](http://git-scm.com/docs/gitignore), fixes some known matching issues of fstream-ignore, such as: - '`/*.js`' should only match '`a.js`', but not '`abc/a.js`'. - '`**/foo`' should match '`foo`' anywhere. - Prevent re-including a file if a parent directory of that file is excluded. - Handle trailing whitespaces: - `'a '`(one space) should not match `'a '`(two spaces). - `'a \ '` matches `'a '` - All test cases are verified with the result of `git check-ignore`. # Methods ## .add(pattern: string | Ignore): this ## .add(patterns: Array): this - **pattern** `String | Ignore` An ignore pattern string, or the `Ignore` instance - **patterns** `Array` Array of ignore patterns. Adds a rule or several rules to the current manager. Returns `this` Notice that a line starting with `'#'`(hash) is treated as a comment. Put a backslash (`'\'`) in front of the first hash for patterns that begin with a hash, if you want to ignore a file with a hash at the beginning of the filename. ```js ignore().add('#abc').ignores('#abc') // false ignore().add('\#abc').ignores('#abc') // true ``` `pattern` could either be a line of ignore pattern or a string of multiple ignore patterns, which means we could just `ignore().add()` the content of a ignore file: ```js ignore() .add(fs.readFileSync(filenameOfGitignore).toString()) .filter(filenames) ``` `pattern` could also be an `ignore` instance, so that we could easily inherit the rules of another `Ignore` instance. ## .addIgnoreFile(path) REMOVED in `3.x` for now. To upgrade `ignore@2.x` up to `3.x`, use ```js import fs from 'fs' if (fs.existsSync(filename)) { ignore().add(fs.readFileSync(filename).toString()) } ``` instead. ## .filter(paths: Array<Pathname>): Array<Pathname> ```ts type Pathname = string ``` Filters the given array of pathnames, and returns the filtered array. - **paths** `Array.` The array of `pathname`s to be filtered. ### `Pathname` Conventions: #### 1. `Pathname` should be a `path.relative()`d pathname `Pathname` should be a string that have been `path.join()`ed, or the return value of `path.relative()` to the current directory, ```js // WRONG, an error will be thrown ig.ignores('./abc') // WRONG, for it will never happen, and an error will be thrown // If the gitignore rule locates at the root directory, // `'/abc'` should be changed to `'abc'`. // ``` // path.relative('/', '/abc') -> 'abc' // ``` ig.ignores('/abc') // WRONG, that it is an absolute path on Windows, an error will be thrown ig.ignores('C:\\abc') // Right ig.ignores('abc') // Right ig.ignores(path.join('./abc')) // path.join('./abc') -> 'abc' ``` In other words, each `Pathname` here should be a relative path to the directory of the gitignore rules. Suppose the dir structure is: ``` /path/to/your/repo |-- a | |-- a.js | |-- .b | |-- .c |-- .DS_store ``` Then the `paths` might be like this: ```js [ 'a/a.js' '.b', '.c/.DS_store' ] ``` #### 2. filenames and dirnames `node-ignore` does NO `fs.stat` during path matching, so for the example below: ```js // First, we add a ignore pattern to ignore a directory ig.add('config/') // `ig` does NOT know if 'config', in the real world, // is a normal file, directory or something. ig.ignores('config') // `ig` treats `config` as a file, so it returns `false` ig.ignores('config/') // returns `true` ``` Specially for people who develop some library based on `node-ignore`, it is important to understand that. Usually, you could use [`glob`](http://npmjs.org/package/glob) with `option.mark = true` to fetch the structure of the current directory: ```js import glob from 'glob' glob('**', { // Adds a / character to directory matches. mark: true }, (err, files) => { if (err) { return console.error(err) } let filtered = ignore().add(patterns).filter(files) console.log(filtered) }) ``` ## .ignores(pathname: Pathname): boolean > new in 3.2.0 Returns `Boolean` whether `pathname` should be ignored. ```js ig.ignores('.abc/a.js') // true ``` ## .createFilter() Creates a filter function which could filter an array of paths with `Array.prototype.filter`. Returns `function(path)` the filter function. ## .test(pathname: Pathname) since 5.0.0 Returns `TestResult` ```ts interface TestResult { ignored: boolean // true if the `pathname` is finally unignored by some negative pattern unignored: boolean } ``` - `{ignored: true, unignored: false}`: the `pathname` is ignored - `{ignored: false, unignored: true}`: the `pathname` is unignored - `{ignored: false, unignored: false}`: the `pathname` is never matched by any ignore rules. ## static `ignore.isPathValid(pathname): boolean` since 5.0.0 Check whether the `pathname` is an valid `path.relative()`d path according to the [convention](#1-pathname-should-be-a-pathrelatived-pathname). This method is **NOT** used to check if an ignore pattern is valid. ```js ignore.isPathValid('./foo') // false ``` ## ignore(options) ### `options.ignorecase` since 4.0.0 Similar as the `core.ignorecase` option of [git-config](https://git-scm.com/docs/git-config), `node-ignore` will be case insensitive if `options.ignorecase` is set to `true` (the default value), otherwise case sensitive. ```js const ig = ignore({ ignorecase: false }) ig.add('*.png') ig.ignores('*.PNG') // false ``` ### `options.ignoreCase?: boolean` since 5.2.0 Which is alternative to `options.ignoreCase` ### `options.allowRelativePaths?: boolean` since 5.2.0 This option brings backward compatibility with projects which based on `ignore@4.x`. If `options.allowRelativePaths` is `true`, `ignore` will not check whether the given path to be tested is [`path.relative()`d](#pathname-conventions). However, passing a relative path, such as `'./foo'` or `'../foo'`, to test if it is ignored or not is not a good practise, which might lead to unexpected behavior ```js ignore({ allowRelativePaths: true }).ignores('../foo/bar.js') // And it will not throw ``` **** # Upgrade Guide ## Upgrade 4.x -> 5.x Since `5.0.0`, if an invalid `Pathname` passed into `ig.ignores()`, an error will be thrown, unless `options.allowRelative = true` is passed to the `Ignore` factory. While `ignore < 5.0.0` did not make sure what the return value was, as well as ```ts .ignores(pathname: Pathname): boolean .filter(pathnames: Array): Array .createFilter(): (pathname: Pathname) => boolean .test(pathname: Pathname): {ignored: boolean, unignored: boolean} ``` See the convention [here](#1-pathname-should-be-a-pathrelatived-pathname) for details. If there are invalid pathnames, the conversion and filtration should be done by users. ```js import {isPathValid} from 'ignore' // introduced in 5.0.0 const paths = [ // invalid ////////////////// '', false, '../foo', '.', ////////////////// // valid 'foo' ] .filter(isValidPath) ig.filter(paths) ``` ## Upgrade 3.x -> 4.x Since `4.0.0`, `ignore` will no longer support node < 6, to use `ignore` in node < 6: ```js var ignore = require('ignore/legacy') ``` ## Upgrade 2.x -> 3.x - All `options` of 2.x are unnecessary and removed, so just remove them. - `ignore()` instance is no longer an [`EventEmitter`](nodejs.org/api/events.html), and all events are unnecessary and removed. - `.addIgnoreFile()` is removed, see the [.addIgnoreFile](#addignorefilepath) section for details. **** # Collaborators - [@whitecolor](https://github.com/whitecolor) *Alex* - [@SamyPesse](https://github.com/SamyPesse) *Samy Pessé* - [@azproduction](https://github.com/azproduction) *Mikhail Davydov* - [@TrySound](https://github.com/TrySound) *Bogdan Chadkin* - [@JanMattner](https://github.com/JanMattner) *Jan Mattner* - [@ntwb](https://github.com/ntwb) *Stephen Edgar* - [@kasperisager](https://github.com/kasperisager) *Kasper Isager* - [@sandersn](https://github.com/sandersn) *Nathan Shively-Sanders*