Croak

Grunt made easy for large projects. It's like One Ring to rule them all, but for Grunt

Disclaimer: Croak is still an initial version, but it's ready-to-use in relaxed environments

About

Grunt is an awesome task runner and build automation tool largely adopted for node.js

Croak is just a simple, but featured wrapper for Grunt that aims to help you to manage and orchestrate Grunt tasks configuration (aka Gruntfile) in large and distributed projects, helping you to avoid redundancy, saving time and reducing dramatically changes impact during the project life-cycle and development workflow

Features

• Allows you to centralize your Gruntfile in one place and use it from many places
• Easily manages and orchestrates different configuration for multiple projects
• Avoid redundant Grunt configuration over multiple repositories
• Abstracts and provides more control for Grunt tasks configuration
• Support to use and discover global node packages as Grunt configuration package
• Extends or overwrites global Gruntfile from local configuration
• Provides the same Grunt API and very similar CLI (things won’t change too much)
• Customizes Grunt execution options from a config file (future .gruntrc)
• You don’t need to have Grunt NPM task locally installed in order to run tasks

Installation

It is recommended that you install Croak as global package:

$ npm install croak -g

Then, create the configuration file:

$ croak init -g

The above command will create a .croakrc file in your user home directory, as global configuration.

Still confused? Please, take some minutes to read the detailed documentation below

Table of Contents

• Why we created Croak?
  • When do I NOT need it?
  • When do I need it?
• How it works
  • Example use case scenario
  • Can abstraction be dangerous?
• Current stage
  • Grunt support
• Using Croak
  • Configuration file
    • Global config
    • Local config
    • Available config options
      • Global options
      • Croak-specific
      • Grunt-specific
  • Switching to Croak
    • Adapting your existing Gruntfile
    • Use Croak from an existing Gruntfile
  • Croakfile
  • Command-line interface
    • Running tasks
    • Configuration
  • Programmatic API
• FAQ
• Development
• Contributing
• To Do/Wish list
• Contributors
• License

Why we created Croak?

Croak arises from the need to create a specific solution to abstract all the automation configuration stuff. It allows you to delegate responsibilities properly in your project to the developers, without losing the desired level of control and centralization.

Because a design principle, node.js (and consequently Grunt) packages and its dependencies are locally installed, and that is fantastic. However, this is not always the best choice for some scenarios. As you probably already know, dependencies globalization in node.js can be considered an anti-pattern, but globalization can be an ideal choice if you are in a big and distributed project.

When a project have a considerable number of repositories, by inertia you probably tend to use the local packages and configuration across each different repository, but these configurations are usually equal or very similar for each different repository.

The above project scenario is really hard to maintain when you need to apply changes massively in your project, like automation, building or deployment configuration changes that is completely usual during the project evolution

Croak is created to supply this need. It aims to provide a specific solution to easily centralize and orchestrate Grunt configuration, so that you can rapidly apply massive changes in your project.

Croak has basically two main goals:

• Reduces the Grunt configuration changes impact during your project life-cycle
• Gives configuration control and provides abstraction to developers, without hurting their feelings

When do I NOT (probably) need it?

• When you have only one or two repositories in your project
• When you have an ultra specific grunt config on each project repository
• When you or your team don’t want to spend time updating/synchronizing build configuration
• When you don’t need to centralize and (consequently) don’t need to take the control
• When you have a lot of free time and you enjoy doing redundant and not very interesting tasks
• When developers of your project... (at least one):
  • are ultra skilled and responsible guys
  • have a lot of time to maintain their own automation tasks configuration
  • need to take full control of the automation tasks configuration

When do I need it?

• When you want to centralize and take control of the project building and automation configuration
• When you want to reduce the configuration change time and impact in your project
• When your project have a considerable number of repositories (more than 3)
• When your project configuration is equal or very similar in different repositories
• When you want to keep clean each repository of Grunt tasks
• When you do NOT want to spend time doing redundant and not very funny stuff
• When you do not want to spend much time providing support to developers about configuration stuff
• Or whatever is not on the 'When do I not need' list...

How it works

In a few words, Croak allows you to have a multiple Gruntfile in global locations and runs them using a simple project identifier (like a linker). It also allows you to define pre-configured options to pass to Grunt in a simple configuration file

Moreover, Croak allows you to extend or overwrite Gruntfile configurations, so you can provide a way to customize tasks configuration from local configuration (see the Croakfile section)

Example scenario

A representative scenario could be, for example, that your project has a relevant number of repositories. Each repository needs its automation configuration for development, testing or building stuff.

Usually, this configuration tends to grow when your repository has a certain amount of different tasks and it also tends to be very similar or redundant in different repositories in your project

For example you could need the same tasks in your project to run tests, generate documentation, statically analyze the code, generate a coverage report or measure the code complexibility

Continuous changes and improvements are a permanent feature in any software project, and Croak just helps with that

You can see a complete example project structure using Croak here

Can abstraction be dangerous?

Abstraction is not always the best choice and some people can hate it. Basically, an abstraction in software development tries to reduce notably the complexity, underlines details and gives more control

With Croak you can provide an ideal level of abstraction to developers without losing control and, additionally, give them freedom if they want to customize or extend Grunt configuration allowing them to adapt it to their needs

If you want to make less impact to the developers, take a look to the Croak Grunt faker

Current stage

Croak is an initial ready-to-use version, though it’s under active development process and important changes can be done in a near future version

Grunt support

Croak supports Grunt ~0.4.0

Using Croak

Configuration file

Croak uses disk files to store persistently the configuration. The Croak configuration file must be called .croakrc

Croak supports two types of configuration files

Global config

A global file can store your project configuration and this configuration will be used in each Croak execution

It will be located in the user home directories by default.

If you want to define a global cross-user shared configuration, you can define CROAKRC environment variable in order to specify a custom global file location, instead of the current user home directory

Local config

Local file can be located in any path, but it’s recommended that you create it in your project or repository root folder

Usually, a local file should have only one project configured. Any local configuration has priority and will overwrite global configuration (for example, if you configure the same project in both files)

Croak implements a similar file discovery algorithm like the one Grunt uses to discover the Gruntfile

Available configuration options

Global config

Per-project config

Croak-specific

Grunt-specific

The following options will be (probably) available in Grunt in future versions, inside a file called .gruntrc, however, Croak already supports it

You can use any of the config options also as a command line flag. See the Grunt CLI documentation

Example file configuration

A multi-project .croakrc configuration file

$default = my-project
 
[super-project]
extend = true
gruntfile = ${HOME}/projects/super-project/build/Gruntfile.coffee
debug = true
stack = true
 
[my-project]
extend = true
overwrite = true
register_tasks = true
; resolve an node package in the system locally and globally installed 
; then Croak will try to discover the Gruntfile inside the package directory 
package = my-project-builder
; grunt-specific 
base = ${HOME}/projects/my-project/my-package/
no_color = false
no_write = false
debug = false
verbose = false
force = true
stack = true
tasks = ../custom-tasks/
npm = mytask

You can use any existing environment variables in config values, using the ${VARIABLE_NAME} notation

Notes about cross-OS variables

In order to use the same config values across diferent OS, Croak will transparently translate common environment variables to the specific running OS

For example, if your use ${HOME}, Croak will translate it into %USERPROFILE% under Windows and viceversa. The same case is applied for ${PWD}, ${TMPDIR}, %HOMEDRIVE%, translating this last one into / under Unix-like OS

Built-in Croak variables

Additionally, Croak introduces support for an easy way to use relative paths from Croak-specific files locations, like the .croakrc config file path

Switching to Croak

Adapting your existing Gruntfile

Croak will automatically expose the croak object in Gruntfile, so you can use this configuration like template values in your tasks config

This is really useful because, in much cases, you need to use absolute paths in your Gruntfiles

Croak grunt object

The following properties will be available as Croak task in Grunt. You can use it as Grunt templating in tasks configuration. If you don't use the base option, you should use one of the following options to configure your tasks in order to use absolute paths

The above properties will be also available from grunt.croak object

Example Gruntfile

module.exports = function (grunt) {
  if (grunt.croak) {
    grunt.log.writeln('Running Grunt from Croak (v.' + grunt.croak.version + ')');
  }
 
  grunt.initConfig({
    clean: {
      options: {
        force: true
      },
      views: '<%= croak.cwd %>/tmpl/'
      tmp: '<%= croak.root %>/.tmp/'
    },
    jshint: {
      options: {
        jshintrc: '.jshintrc'
      },
      all: [
        '<%= croak.base %>/src/scripts/{,*/}*.js'
        '<%= croak.config %>/demo/{,*/}*.js'
        '!<%= croak.root %>/src/scripts/vendor/*.js'
      ]
    },
    connect: {
      server: {
        options: {
          port: 9001,
          base: '<%= croak.root %>'
        }
      }
    }
  });
};

Use Croak from an existing Gruntfile

The Croak Grunt task is still in progress

Croakfile

Like Grunt, Croak has its own specific tasks configuration file.

The Croakfile has the same API as Grunt and it can also be a JavaScript or CoffeeScript file. The module should export a function, like in Grunt

The idea behind the Croakfile is to provide an isolated configuration file without conflicting with Gruntfile for specifically extend or overwrite Grunt tasks configuration from local config to global config

If the Croak project configuration allows extend or overwrite Grunt config, you can use the Croakfile to customize and overwrite globally configured tasks to adapt it to your needs

module.exports = function (croak) {
 
  croak.log.ok('Version:', croak.version)
  croak.log.ok('Grunt version:', croak.grunt.version)
 
  croak.extendConfig({
    log: {
      foo: {},
      bar: 'hello croak'
    },
    read: {
      files: [
        '<%= croak.root %>/file.json'
        __dirname + '/another-file.json'
      ]
    }
  })
 
  // you can also register new tasks if 'register_tasks' option is enabled
  croak.registerMultiTask('log', 'Log stuff.', function() {
    grunt.log.writeln(this.target + ': ' + this.data)
  })
 
  croak.registerMultiTask('read', 'Read file.', function() {
    grunt.log.writeln(this.target + ': ' + this.data)
  })
 
  croak.registerTask('default', ['log'])
 
}

Croakfile API

The Croakfile API inherits from Grunt API. You can do the same thing in a Gruntfile. Please, see the Gruntfile documentation for more information

Additionally, Croak provides the next useful and Grunt missing API

croak.extendConfig(config)

Alias to grunt.initConfig(), but just for a more semantic usage from Croakfiles

croak.task.exists()

Check if a tasks were already registered and exist

croak.grunt

Exposes the native Grunt API module object. If you use directly the Grunt API, for example, to register new tasks or remove config, Croak will not control if you can do it. Use it under your own reponsibility

Command-line interface

Croak provides a straightforward CLI that allows you to do almost everything (including all that you can do with Grunt CLI).

Aditionally, if you are convervatory and you want to use Croak keeping the same native Grunt CLI interface, take a look to the Croak Grunt faker

  Usage: croak [options] [command]

  Commands:

    help
      Output the usage information
    version
      Output the version information
    init [options] [name]
      Create new projects in .croakrc
    config [options] <action> [key] [value]
      Read/write/update/remove croak config
    run [options] [task]
      Run Grunt tasks
    add [options] [name]
      Add new projects in .croakrc

  Options:

    -h, --help     output usage information
    -V, --version  output the version number

  Usage examples:

    $ croak init [name]
    $ croak add [name]
    $ croak config [create|list|get|set|remove]
    $ croak run task
    $ croak grunt task

  Command specific help:

    $ croak <command> --help

  Grunt help:

    $ croak grunt --help

Running Grunt tasks

$ croak run task -p my-project

You need to pass the -p flag if there’s not a .croakrc local config file

You can configure the default project to use from a local config file, like this:

$default = my-project

Then you can simply run:

$ croak run task

You can also run subtasks:

$ croak run task:subtask

or multiple tasks:

$ croak run task anothertask

Configuration

Show the current existing config

$ croak config list

Create a config file (add -g flag to create it globally)

$ croak config create

You can CRUD config values easily from CLI

$ croak config [get|set|remove] <key> [value] [-g]

Programmatic API

You can use Croak as node.js module to integrate it in your application

Disclaimer: Croak API can change radically in a near future. Retrocompatibility is not guaranteed for minor 0.x.x beta releases

Installation

Install Croak as local package

$ npm install croak [--save] [--save-dev]

Then require the module

var croak = require('croak')

Croak API

The croak module exposes the following members

version

Current Croak module version

gruntVersion

The current Grunt version (which is using Croak)

grunt

Expose the Grunt module object. See Grunt API for more information

config

Exposes the Croak config module API

load([ configPath ])

Discover configuration files, then it will read, parse and load projects configuration

loadProject([ configPath ])

Load config and returns the default configured project, it it exists

get([ key ])

Alias to config#get

set(key [, value, isLocalConfig])

Alias to config#set

init([ options, projectObj ])

This method will configure Croak with the passed options config object and aditionally you can pass the project config object (that you should get it previously via croak.config.get())

After it, Croak will try to discover Croakfiles, and then configure and run Grunt

initGrunt([ options ])

Configure Grunt with the passed options and run it

Croak Config API

The Croak config object is exposed via the croak module

var config = require('croak').config

config.load([ filePath ])

This method tries to discover configuration files, then it will read, parse and load them It will throw an Error exception if cannot read or parse configuration files

config.save([ filePath ])

Saves existing loaded configuration in disk

It will throw an Error exception if cannot write data

config.raw()

Returns an object with both global and local properties with the config data as raw string (ini format)

config.get([ key ])

Retrieves an existing config value querying by its key. If not argument is passed, it will return the whole config

croakConfig.get('my-project') // { gruntfile: './path/to/Gruntfile.js' ... }
croakConfig.get('my-project.gruntfile') // './path/to/Gruntfile.js'

config.getDefault()

Returns the default configured project config object. If it is not configured, it will return the first existing project, first by looking into local config, and then in the global one

config.set(key [, value, isLocalConfig])

Setter method for config values. You can set a new project, or a specific project config value with a primitive type

croakConfig.set('my-project', { gruntfile: '../Gruntile.js' })
croakConfig.set('my-project.gruntfile', '../new/path/to/Gruntile.js')

config.setLocal(key [, value])

config.remove(key)

Removes a config value

croakConfig.remove('my-project') // removes the whole project
croakConfig.remove('my-project.gruntfile') // removes a specific key

config.value(key [, value, isLocalConfig])

Alias accessor method for set() and get()

croakConfig.value('my-project') // I'm getting a value
croakConfig.value('my-project.gruntfile', '../Gruntfile.js') // I'm setting a value

config.globalFile()

Returns the absolute path to the global configuration file in the system

config.localFile([ filePath ])

Returns the absolute path to the local configuration file

Croak implements a simple algorithm like Grunt or NPM in order to discover the configuration file in the current and in higher directories.

Optionally, you can explicitly define a configuration path file to use passing the filePath argument

config.clean()

Clean loaded config from cache. To apply it persistently in the disk, you must call it config#save()

config.hasData()

Returns true if the configuration was loaded and exists

config.exists([ key ])

Check if a given property key exists in the current config

config.exists('project') // true
config.exists('nonexisting.gruntfile') // false

config.path()

Returns an object with absolute paths to both global and local config files

config.path()
// { local: '/home/user/.croakrc', local: '/home/user/projects/awesome/.croakrc' }

config.dirname()

The same as config#path(), but this one only will return the absolute path directory where the config file is located

FAQ

Do I need to have Grunt already installed?

No, Croak will do it for you. And also grunt-cli will be replaced by Croak CLI

Can I use the Grunt CLI?

Of course, it’s available using the grunt command

$ croak grunt --help

Do I need to have a Gruntfile in my repository?

No. An already existing Gruntfile is not required.

You only need to specify the global Gruntfile you want to use and, optionally, you can use a Croakfile to overwrite or extend global configuration

Can I use both Croak and Grunt at the same time?

Yes. You must specify a global Gruntfile and also have your own repository Gruntfile.

You can run both like this:

$ croak run task -p project

$ croak grunt localtask

Development

Only node.js is required for development

1. Clone/fork this repository

$ git clone git@github.com:adesisnetlife/croak.git && cd croak

1. Install package dependencies

$ npm install

1. Run tests

$ npm test

1. Coding zen mode

$ grunt dev [--force]

Contributing

Croak is completely written in LiveScript language. Take a look to the language documentation if you are new with it. Please, follow the LiveScript language conventions defined in the coding style guide

You must add new test cases for any feature or refactor you do, also keep in mind to follow the same design/code patterns that already exist

To Do/Wish list

• Create an example use case project
• Switch from promptly to inquirer (more featured and pretty CLI, like yeoman)
• More test cases scenarios and destructive/evil testing
• Croak API improvements: add callbacks and event listeners for a better Grunt execution control
• Error CLI test cases
• Better messages for the CLI and verbose mode
• Better usage of the native grunt API (use grunt.option())
• Simplified and better documentation (create a wiki?)
• Support for multiple Gruntfile in the same project, or distributed Grunt tasks?
• Support for Gruntfile instead of Croakfile? Rationale: easier to understand

Do you miss something? Please, feel free to open an issue or make a PR!

Contributors

• Tomás Aparicio

License

Copyright (c) 2013 Adesis Netlife S.L and contributors

Released under the MIT license