gitftw - Git For The Workflows

Basic git commands wrapped in node for using in your workflows. This includes grunt, gulp, your custom scripts, CI, etc...

http://jmendiara.github.io/gitftw/

Installation

npm install gitftw --save-dev

Basic Usage

var git = require('gitftw');

//Executes locally installed git with the parameters specified as an array
// > /usr/local/bin/git tag 
git(['tag'], function(err, output) {
  if (err) {
    console.error('Command failed with code %d: %s', err.code, err.message);
    return;
  }
  console.log(output); // "'v1.0.0\nv1.0.1\nv1.0.2'" 
});

Commands

There are some basic commands implemented for you, in a workflow developer friendly way.

var git = require('gitftw');

git.pull({
    remote: 'origin',
    branch: 'master',
    rebase: true
  }, cb);

git.add({
    files: ['README.md', 'package.json']
  }, cb);

//removing all local tags
git.getTags(function (err, tags) {
  if (err) {
    return cb(err);
  }
  console.log(tags);  //['v1.0.0', 'v1.0.1', 'v1.0.2']
  git.removeLocalTags({
      tags: tags
  }, cb);
})

All the commands can take 2 parameters, an optional options literal object, and the optional callback. A literal object helps you having configurations in a json, a grunt config, etc... even in a function! (read below)

More information on implemented commands, please refer to the documentation

Knowing what's happening under the hoods

Yes, it's important to see wat is doing an automated workflow. We have some events for you.

var git = require('gitftw');

//Add a listener to the issued git command. Output it
git.events.on('command', console.info);

//Add a listener to the result of the git command. Output it with > 
git.events.on('result', function(res) {
  console.log('> ' + res.split('\n').join('\n> '))
});

Sugar: Promises

git and its commands have a dual API, both the node callback style you have seen in the examples, and the promise style. Bluebird is used internally as the promises library

//removing all local tags
git.getTags()
  .tap(console.log) //['v1.0.0', 'v1.0.1', 'v1.0.2']
  .then(function (tags) {
    return git.removeLocalTags({
      tags: tags
    })
  });

Having promises internally and optionally outside simplifies the development of commands to manage concurrency when issuing git commands with a more functional style.

Why concurrency matters? Not having concurrency accessing the filesystem, where some commands can change its status, is very important for a predictable command sequence. Think on how you use git from the command line. We wanna make this tool for workflows as much predictable as possible. And the resolvable concept helps on both serialization and functional style development

DISCLAIMER for node fan-boys: frontenders are also developers. And are bored of your if(err) return cb(err); verbose style. And I've always wanted to code something dual :)

Advanced usage: Resolvables

You can use resolvables things when calling this library methods. Resolvable is something that have invariant primitives, including all its properties, now or in the future.

Which things are resolvables?

1. A String, Number, Boolean primitives: string, 4, true

2. A array of resolvables: ['string', resolvable]

3. A literal object with resolvables properties: { foo: 'string', bar: resolvable }

4. An object with a toString method

5. An A+ promise (Q, bluebird, native...) that resolves to a resolvable: Promise.when('string')

6. A function that returns resolvables function() { return 'string'; }

resolvables are resolved in serie. Once something is resolved, goes to resolve the next one. Never in parallel. This avoids race conditions in your workflows while maintaining your code clean.

It's easier than you think. The above remove all tags example with resolvables:

//removing all local tags
git.removeLocalTags({
  tags: git.getTags //Command that returns a promise for an array of strings
});

git.getTags, as it's used as a function parameter, will be called, resolved and assigned before git.removeLocalTags gets called.

Take a look to the tests.

License

MIT