How to run two grunt watch tasks simultaneously

I've found using grunt-concurrent works:

concurrent: {
  options: {
    logConcurrentOutput: true
  },
  prod: {
    tasks: ["watch:A", "watch:C"]
  },
  dev: {
    tasks: ["watch:B", "watch:C"]
  }
}

Then:

grunt.registerTask("prod", ["concurrent:prod"]);
grunt.registerTask("dev", ["concurrent:dev"]);

The best and only working solution is there : https://npmjs.org/package/grunt-focus Add this plugin and then :

focus: {
            sources: {
                include: ['js', 'html', 'css', 'grunt']
            },
            testu: {
                include: ['js', 'html', 'css', 'testu', 'grunt']
            },
            testi: {
                include: ['js', 'html', 'css', 'testu', 'testi', 'grunt']
            }
        },
        watch: {
            js: {
                files: paths.js,
                tasks: ['jshint'],
                options: {
                    livereload: true
                }
            },
            html: {
                files: paths.html,
                options: {
                    livereload: true
                }
            },
            css: {
                files: paths.css,
                tasks: ['csslint'],
                options: {
                    livereload: true
                }
            },
            testu: {
                files: ['test/**/*.js', 'test/**/*.css'],
                tasks: ['mochaTest'],
                options: {}
            },
            testi: {
                files: ['test/**/*.js', 'test/**/*.css'],
                tasks: ['exec:cleanTestDB', 'protractor_webdriver', 'protractor'],
                options: {}
            },
            grunt: {
                files: ['Gruntfile.js', 'server/config/env/*.js'],
                options: {
                    reload: true
                }
            }
        }

Then you use focus:sources or focus:testu as your convenience.

JM.


grunt-concurrent or grunt-focus are both good solutions, but both of them break livereload functionality.

My solution to this is to compose the watch configuration dynamically, with the assumption that you won't be running both configuration at the same time.

You can do something like this

grunt.config.merge({
  watch: {
    options: {
      livereload: true
    },
    C: {
      files: "js/dev/**/*.html",
      tasks: ["copy"]
    }
  }
});

grunt.registerTask('watch-forProd', function () {
  grunt.config.merge({
    watch: {
      A: {
        files: "js/dev/**/*.coffee",
        tasks: ["coffee", "requirejs"]
      }
    }
  });

  grunt.task.run('watch');
});

grunt.registerTask('watch-forDev', function () {
  grunt.config.merge({
    watch: {
      B: {
        files: "js/dev/**/*.coffee",
        tasks: ["coffee"]
      }
    }
  });

  grunt.task.run('watch');
});

grunt.registerTask("prod", ["watch-forProd"]);
grunt.registerTask("dev", ["watch-forDev"]);

EDIT: concurrent now has a logConcurrentOutput option! More info here: https://github.com/sindresorhus/grunt-concurrent#logconcurrentoutput.

Watch is a weirdly concurrent but blocking task, so you have to be creative to get multitask-like functionality working.

Concurrent loses all output from the watch tasks, which isn't ideal.

Try dynamically writing the config object in a custom task:

grunt.registerTask('watch:test', function() {
  // Configuration for watch:test tasks.
  var config = {
    options: {
      interrupt: true
    },
    unit: {
      files: [
        'test/unit/**/*.spec.coffee'
      ],
      tasks: ['karma:unit']
    },
    integration: {
      files: [
        'test/integration/**/*.rb',
        '.tmp/scripts/**/*.js'
      ],
      tasks: ['exec:rspec']
    }
  };

  grunt.config('watch', config);
  grunt.task.run('watch');
});