Configuring tasks

This guide explains how to configure tasks for your project using a Gruntfile. If you don't know what a Gruntfile is, please read the Getting Started guide and check out a Sample Gruntfile.

Grunt Configuration

Task configuration is specified in your Gruntfile via the grunt.initConfig method. This configuration will mostly be under task-named properties, but may contain any arbitrary data. As long as properties don't conflict with properties your tasks require, they will be otherwise ignored.

Also, because this is JavaScript, you're not limited to JSON; you may use any valid JavaScript here. You may even programmatically generate the configuration if necessary.

grunt.initConfig({
  concat: {
    // concat task configuration goes here.
  },
  uglify: {
    // uglify task configuration goes here.
  },
  // Arbitrary non-task-specific properties.
  my_property: 'whatever',
  my_src_files: ['foo/*.js', 'bar/*.js'],
});

Task Configuration and Targets

When a task is run, Grunt looks for its configuration under a property of the same name. Multi-tasks can have multiple configurations, defined using arbitrarily named "targets." In the example below, the concat task has foo and bar targets, while the uglify task only has a bar target.

grunt.initConfig({
  concat: {
    foo: {
      // concat task "foo" target options and files go here.
    },
    bar: {
      // concat task "bar" target options and files go here.
    },
  },
  uglify: {
    bar: {
      // uglify task "bar" target options and files go here.
    },
  },
});

Specifying both a task and target like grunt concat:foo or grunt concat:bar will process just the specified target's configuration, while running grunt concat will iterate over all targets, processing each in turn. Note that if a task has been renamed with grunt.task.renameTask, Grunt will look for a property with the new task name in the config object.

Options

Inside a task configuration, an options property may be specified to override built-in defaults. In addition, each target may have an options property which is specific to that target. Target-level options will override task-level options.

The options object is optional and may be omitted if not needed.

grunt.initConfig({
  concat: {
    options: {
      // Task-level options may go here, overriding task defaults.
    },
    foo: {
      options: {
        // "foo" target options may go here, overriding task-level options.
      },
    },
    bar: {
      // No options specified; this target will use task-level options.
    },
  },
});

Files

Because most tasks perform file operations, Grunt has powerful abstractions for declaring on which files the task should operate. There are several ways to define src-dest (source-destination) file mappings, offering varying degrees of verbosity and control. Any multi task will understand all the following formats, so choose whichever format best meets your needs.

All files formats support src and dest but the Compact and Files Array formats support a few additional properties:

  • filter Either a valid fs.Stats method name or a function that is passed the matched src filepath and returns true or false. See examples
  • nonull If set to true then the operation will include non-matching patterns. Combined with grunt's --verbose flag, this option can help debug file path issues.
  • dot Allow patterns to match filenames starting with a period, even if the pattern does not explicitly have a period in that spot.
  • matchBase If set, patterns without slashes will be matched against the basename of the path if it contains slashes. For example, a?b would match the path /xyz/123/acb, but not /xyz/acb/123.
  • expand Process a dynamic src-dest file mapping, see "Building the files object dynamically" for more information.
  • Other properties will be passed into the underlying libs as matching options. See the node-glob and minimatch documentation for more options.

Difference Between Grunt and Task Options

Most tasks perform file operations, so Grunt provides a built-in infrastructure to retrieve the files a task should process. The advantage is that this logic doesn't have to be implemented again by tasks authors. To allow a user to specify these files, Grunt provides options such as nonull and filter.

In addition to the files to work on, each task has its own specific needs. A task author may want to allow its user to configure some options to override the default behavior. These task-specific options shall not be confused with the Grunt options described before.

To further clarify this difference, let's see an example that uses grunt-contrib-jshint:

grunt.initConfig({
  jshint: {
    ignore_warning: {
      options: {
        '-W015': true,
      },
      src: 'js/**',
      filter: 'isFile'
    }
  }
});

This configuration employs the Grunt options src and filter to specify the files to process. It also uses the grunt-contrib-jshint task-specific option -W015 to ignore a specific warning (the one having code W015).

Compact Format

This form allows a single src-dest (source-destination) file mapping per-target. It is most commonly used for read-only tasks, like grunt-contrib-jshint, where a single src property is needed, and no dest key is relevant. This format also supports additional properties per src-dest file mapping.

grunt.initConfig({
  jshint: {
    foo: {
      src: ['src/aa.js', 'src/aaa.js']
    },
  },
  concat: {
    bar: {
      src: ['src/bb.js', 'src/bbb.js'],
      dest: 'dest/b.js',
    },
  },
});

Files Object Format

This form supports multiple src-dest mappings per-target, where the property name is the destination file, and its value is the source file(s). Any number of src-dest file mappings may be specified in this way, but additional properties may not be specified per mapping.

grunt.initConfig({
  concat: {
    foo: {
      files: {
        'dest/a.js': ['src/aa.js', 'src/aaa.js'],
        'dest/a1.js': ['src/aa1.js', 'src/aaa1.js'],
      },
    },
    bar: {
      files: {
        'dest/b.js': ['src/bb.js', 'src/bbb.js'],
        'dest/b1.js': ['src/bb1.js', 'src/bbb1.js'],
      },
    },
  },
});

Files Array Format

This form supports multiple src-dest file mappings per-target, while also allowing additional properties per mapping.

grunt.initConfig({
  concat: {
    foo: {
      files: [
        {src: ['src/aa.js', 'src/aaa.js'], dest: 'dest/a.js'},
        {src: ['src/aa1.js', 'src/aaa1.js'], dest: 'dest/a1.js'},
      ],
    },
    bar: {
      files: [
        {src: ['src/bb.js', 'src/bbb.js'], dest: 'dest/b/', nonull: true},
        {src: ['src/bb1.js', 'src/bbb1.js'], dest: 'dest/b1/', filter: 'isFile'},
      ],
    },
  },
});

Older Formats

The dest-as-target file format is a holdover from before multi tasks and targets existed, where the destination filepath is actually the target name. Unfortunately, because target names are filepaths, running grunt task:target can be awkward. Also, you can't specify target-level options or additional properties per src-dest file mapping.

Consider this format deprecated, and avoid it where possible.

grunt.initConfig({
  concat: {
    'dest/a.js': ['src/aa.js', 'src/aaa.js'],
    'dest/b.js': ['src/bb.js', 'src/bbb.js'],
  },
});

Custom Filter Function

The filter property can help you target files with a greater level of detail. Simply use a valid fs.Stats method name. The following will clean only if the pattern matches an actual file:

grunt.initConfig({
  clean: {
    foo: {
      src: ['tmp/**/*'],
      filter: 'isFile',
    },
  },
});

Or create your own filter function and return true or false whether the file should be matched. For example the following will only clean folders that are empty:

grunt.initConfig({
  clean: {
    foo: {
      src: ['tmp/**/*'],
      filter: function(filepath) {
        return (grunt.file.isDir(filepath) && require('fs').readdirSync(filepath).length === 0);
      },
    },
  },
});

Another example—which utilizes the globbing and expand: true features—allows you to avoid overwriting files which already exist in the destination:

grunt.initConfig({
  copy: {
    templates: {
      files: [{
        expand: true,
        cwd: ['templates/css/'],     // Parent folder of original CSS templates
        src: '**/*.css',             // Collects all `*.css` files within the parent folder (and its subfolders)
        dest: 'src/css/',            // Stores the collected `*.css` files in your `src/css/` folder
        filter: function (dest) {    // `dest`, in this instance, is the filepath of each matched `src`
          var cwd = this.cwd,        // Configures variables (these are documented for your convenience only)
              src = dest.replace(new RegExp('^' + cwd), '');
              dest = grunt.task.current.data.files[0].dest;
          return (!grunt.file.exists(dest + src));    // Copies `src` files ONLY if their destinations are unoccupied
        }
      }]
    }
  }
});

Keep in mind the above technique does not account for the rename property when checking if the destination exists.

Globbing patterns

It is often impractical to specify all source filepaths individually, so Grunt supports filename expansion (also known as globbing) via the built-in node-glob and minimatch libraries.

While this isn't a comprehensive tutorial on globbing patterns, know that in a filepath:

  • * matches any number of characters, but not /
  • ? matches a single character, but not /
  • ** matches any number of characters, including /, as long as it's the only thing in a path part
  • {} allows for a comma-separated list of "or" expressions
  • ! at the beginning of a pattern will negate the match

All most people need to know is that foo/*.js will match all files ending with .js in the foo/ subdirectory, but foo/**/*.js will match all files ending with .js in the foo/ subdirectory and all of its subdirectories.

Also, in order to simplify otherwise complicated globbing patterns, Grunt allows arrays of file paths or globbing patterns to be specified. Patterns are processed in-order, with !-prefixed matches excluding matched files from the result set. The result set is uniqued.

For example:

// You can specify single files:
{src: 'foo/this.js', dest: ...}
// Or arrays of files:
{src: ['foo/this.js', 'foo/that.js', 'foo/the-other.js'], dest: ...}
// Or you can generalize with a glob pattern:
{src: 'foo/th*.js', dest: ...}

// This single node-glob pattern:
{src: 'foo/{a,b}*.js', dest: ...}
// Could also be written like this:
{src: ['foo/a*.js', 'foo/b*.js'], dest: ...}

// All .js files, in foo/, in alpha order:
{src: ['foo/*.js'], dest: ...}
// Here, bar.js is first, followed by the remaining files, in alpha order:
{src: ['foo/bar.js', 'foo/*.js'], dest: ...}

// All files except for bar.js, in alpha order:
{src: ['foo/*.js', '!foo/bar.js'], dest: ...}
// All files in alpha order, but with bar.js at the end.
{src: ['foo/*.js', '!foo/bar.js', 'foo/bar.js'], dest: ...}

// Templates may be used in filepaths or glob patterns:
{src: ['src/<%= basename %>.js'], dest: 'build/<%= basename %>.min.js'}
// But they may also reference file lists defined elsewhere in the config:
{src: ['foo/*.js', '<%= jshint.all.src %>'], dest: ...}

For more on glob pattern syntax, see the node-glob and minimatch documentation.

Building the files object dynamically

When you want to process many individual files, a few additional properties may be used to build a files list dynamically. These properties may be specified in both Compact and Files Array mapping formats.

expand Set to true will enable the following properties:

  • cwd All src matches are relative to (but don't include) this path.
  • src Pattern(s) to match, relative to the cwd.
  • dest Destination path prefix.
  • ext Replace any existing extension with this value in generated dest paths.
  • extDot Used to indicate where the period indicating the extension is located. Can take either 'first' (extension begins after the first period in the file name) or 'last' (extension begins after the last period), and is set by default to 'first' [Added in 0.4.3]
  • flatten Remove all path parts from generated dest paths.
  • rename Embeds a customized function, which returns a string containing the new destination and filename. This function is called for each matched src file (after extension renaming and flattening). More information

In the following example, the uglify task will see the same list of src-dest file mappings for both the static_mappings and dynamic_mappings targets, because Grunt will automatically expand the dynamic_mappings files object into 4 individual static src-dest file mappings—assuming 4 files are found—when the task runs.

Any combination of static src-dest and dynamic src-dest file mappings may be specified.

grunt.initConfig({
  uglify: {
    static_mappings: {
      // Because these src-dest file mappings are manually specified, every
      // time a new file is added or removed, the Gruntfile has to be updated.
      files: [
        {src: 'lib/a.js', dest: 'build/a.min.js'},
        {src: 'lib/b.js', dest: 'build/b.min.js'},
        {src: 'lib/subdir/c.js', dest: 'build/subdir/c.min.js'},
        {src: 'lib/subdir/d.js', dest: 'build/subdir/d.min.js'},
      ],
    },
    dynamic_mappings: {
      // Grunt will search for "**/*.js" under "lib/" when the "uglify" task
      // runs and build the appropriate src-dest file mappings then, so you
      // don't need to update the Gruntfile when files are added or removed.
      files: [
        {
          expand: true,     // Enable dynamic expansion.
          cwd: 'lib/',      // Src matches are relative to this path.
          src: ['**/*.js'], // Actual pattern(s) to match.
          dest: 'build/',   // Destination path prefix.
          ext: '.min.js',   // Dest filepaths will have this extension.
          extDot: 'first'   // Extensions in filenames begin after the first dot
        },
      ],
    },
  },
});

The rename Property

The rename property is unique, as the only valid value for it is a JavaScript function. Although the function returns a string, you cannot simply use a string as a value for rename (doing so results in an error: Property 'rename' of object # is not a function). In the following example, the copy task will create a backup of README.md.

grunt.initConfig({
  copy: {
    backup: {
      files: [{
        expand: true,
        src: ['docs/README.md'],    // The README.md file has been specified for backup
        rename: function () {       // The value for rename must be a function
          return 'docs/BACKUP.txt'; // The function must return a string with the complete destination
        }
      }]
    }
  }
});

When the function is called, the dest and matched src path are passed in and can be used for returning the output string. In the following example, files are copied from the dev folder to the dist folder, and renamed to have the word "beta" removed .

grunt.initConfig({
  copy: {
    production: {
      files: [{
        expand: true,
        cwd: 'dev/',
        src: ['*'],
        dest: 'dist/',
        rename: function (dest, src) {          // The `dest` and `src` values can be passed into the function
          return dest + src.replace('beta',''); // The `src` is being renamed; the `dest` remains the same
        }
      }]
    }
  }
});

If multiple matched src paths are renamed to an identical destination (i.e. if two different files get renamed to the same file), each output will be added to an array of sources for it.

Templates

Templates specified using <% %> delimiters will be automatically expanded when tasks read them from the config. Templates are expanded recursively until no more remain.

The entire config object is the context in which properties are resolved. Additionally, grunt and its methods are available inside templates, eg. <%= grunt.template.today('yyyy-mm-dd') %>.

  • <%= prop.subprop %> Expand to the value of prop.subprop in the config, regardless of type. Templates like this can be used to reference not only string values, but also arrays or other objects.
  • <% %> Execute arbitrary inline JavaScript code. This is useful with control flow or looping.

Given the sample concat task configuration below, running grunt concat:sample will generate a file named build/abcde.js by concatenating the banner /* abcde */ with all files matching foo/*.js + bar/*.js + baz/*.js.

grunt.initConfig({
  concat: {
    sample: {
      options: {
        banner: '/* <%= baz %> */\n',   // '/* abcde */\n'
      },
      src: ['<%= qux %>', 'baz/*.js'],  // [['foo/*.js', 'bar/*.js'], 'baz/*.js']
      dest: 'build/<%= baz %>.js',      // 'build/abcde.js'
    },
  },
  // Arbitrary properties used in task configuration templates.
  foo: 'c',
  bar: 'b<%= foo %>d', // 'bcd'
  baz: 'a<%= bar %>e', // 'abcde'
  qux: ['foo/*.js', 'bar/*.js'],
});

Importing External Data

In the following Gruntfile, project metadata is imported into the Grunt config from a package.json file, and the grunt-contrib-uglify plugin uglify task is configured to minify a source file and generate a banner comment dynamically using that metadata.

Grunt has grunt.file.readJSON and grunt.file.readYAML methods for importing JSON and YAML data.

grunt.initConfig({
  pkg: grunt.file.readJSON('package.json'),
  uglify: {
    options: {
      banner: '/*! <%= pkg.name %> <%= grunt.template.today("yyyy-mm-dd") %> */\n'
    },
    dist: {
      src: 'src/<%= pkg.name %>.js',
      dest: 'dist/<%= pkg.name %>.min.js'
    }
  }
});