Our goal is to create brilliant software and we setup an end-to-end software development life-cycle (SDLC) aiming at that. Therefore, we have created the following manual that describes the life-cycle we have here at joo.la labs.

The development process

This is an internal process and describe steps that are relevant not only to the Joola framework development, but also to commercial implementations. On top of the process described we have our community contributions that act as a key part of the software development path. We believe that sharing information about our internal processes with the community will not only make our process better, but also allow the community to benefit from it.

Assess Needs

The first step in the process is to assess the needs of this iteration. What tasks, defects, milestones, etc... we have to cover during the iteration. This process is key to the success of the iteration, getting a full understanding of a requirement ensures a proper execution during the iteration with a very clear set of expectations. The format of this step is an Assessment Meeting which includes all stake-holders relevant to the iteration. This step can take a few hours/days and the group decides on what is included in the iteration and what's not.

Design Specifications

Once we have the needs clarified we can design the specifications for each of the new features, bug fix, etc... Specifications must be detailed and try to evaluate the entire scope of the task. When designing the specifications we must take into account that a task estimated to take longer than the iteration cannot be accepted, it will need to be broken down. Once we have the specifications for each task, the matching estimations and all required work, we can then approve the final list of issues to be covered in the iteration.

Design/Develop/Test

The actual code work begins, we design, develop and test our work on a regular basis. Git development process follows Vincent Driessen's suggested GitFlow and is managed under GitHub. Code work should be carried out on branches and merged into the develop branch via managed pull-requests. When appropriate, a release process starts which merges develop into master via a release branch.

Code work must adhere to the Coding Guidelines.

Implement Systems

Once the release is ready, according to the Release Management protocol, we can move on to implementation. We have multiple environments for ourselves and managed customers. The Deployment process must be followed by the Release Manager.

Support Operations

Following the release process we can move on to monitor and support. The goal of this phase is to ensure the released artifacts are performing well.

Evaluate Performance

Another key part of the iteration is the evaluation phase. During this phase we will contact customers who received custom features and confirm their satisfaction and acceptance. For our larger customers base (SaaS) we will send targeted surveys and use the Site's Feedback option to collect impressions. Another part of this evaluation will be measuring the operational impact of the release, i.e. does it increase or decrease our operating costs.

And Again...

This concludes a single iteration, this is repeated every time. The more we do it, the better we get at it.

Code testing

Joola puts a lot of emphasis on code testing, we aspire to have 100% of code covered with tests.

We use the Mocha testomg framework for our test management, it's simple and straightforward. Running the test suite on your machine, should ensure that the system is functioning properly and that you have properly installed all required packages, dependencies, etc...

Running Tests

You have a few options to start the tests, the most basic one is npm test, but other methods can come in handy if you wish to run a specific test or write your own.

$ npm test
# or
$ make test
# or
$ mocha

For this and other examples of the testing framework, we'll assume that you've installed Mocha as a global package:

$ npm install -g mocha

Running Specific Tests

Before we get started, we'll assume that you're familiar with Mocha, if that's not the case, make sure you try their website for the basic documentation. Joola is a "service" and requires the Mocha testing framework to init a new Joola server instance before starting the tests, for this purpose, we've created a small helper file named test/unit/starthere.spec.js. This small script starts a new instance and sets the required global variables used throughout the test suite.

When we wish to run a specific test, we can combine the starthere script with the specific test script, for example:

$ mocha test/unit/starthere.spec.js test/unit/1_runtime/grid.spec.js

This will execute the starthere script first which will bring Joola online, and then run the grid set of tests against that instance.

Test Configuration and Caveats

The tests use the basic configuration currently existing on the server, so if we're talking about a newly installed instance, it would be whatever exist in config/baseline.json.

  • When running the tests, it is assumed that no other active Joola node is connected to the redis store used by the tests.
  • The tests are designed to run on a new and non-initialized server, mainly on a Travis CI build.

Writing your own tests

If you wish to contribute and add/modify tests, we welcome the help! Please refer to our [Contribution][contributing] section for additional information. In general:

  • Try to follow the existing code style of existing tests.
  • Make sure you cover all cases as part of your tests.
  • If your test needs to use fixtures, make sure these are available either online (as a gist maybe) or locally in the fixtures folder.

Style guide

Copied from Felix's Node.js Style Guide

This is a guide for writing consistent and aesthetically pleasing node.js code. It is inspired by what is popular within the community, and flavored with some personal opinions.

This guide was created by Felix Geisendörfer and is licensed under the CC BY-SA 3.0 license. You are encouraged to fork this repository and make adjustments according to your preferences.

Creative Commons License

2 Spaces for indention

Use 2 spaces for indenting your code and swear an oath to never mix tabs and spaces - a special kind of hell is awaiting you otherwise.

Newlines

Use UNIX-style newlines (\n), and a newline character as the last character of a file. Windows-style newlines (\r\n) are forbidden inside any repository.

No trailing whitespace

Just like you brush your teeth after every meal, you clean up any trailing whitespace in your JS files before committing. Otherwise the rotten smell of careless neglect will eventually drive away contributors and/or co-workers.

Use Semicolons

According to scientific research, the usage of semicolons is a core values of our community. Consider the points of the opposition, but be a traditionalist when it comes to abusing error correction mechanisms for cheap syntactic pleasures.

80 characters per line

Limit your lines to 80 characters. Yes, screens have gotten much bigger over the last few years, but your brain has not. Use the additional room for split screen, your editor supports that, right?

Use single quotes

Use single quotes, unless you are writing JSON.

Right:

var foo = 'bar';

Wrong:

var foo = "bar";

Opening braces go on the same line

Your opening braces go on the same line as the statement.

Right:

if (true) {
  console.log('winning');
}

Wrong:

if (true)
{
  console.log('losing');
}

Also, notice the use of whitespace before and after the condition statement.

Declare one variable per var statement

Declare one variable per var statement, it makes it easier to re-order the lines. Ignore Crockford on this, and put those declarations wherever they make sense.

Right:

var keys   = ['foo', 'bar'];
var values = [23, 42];

var object = {};
while (items.length) {
  var key = keys.pop();
  object[key] = values.pop();
}

Wrong:

var keys = ['foo', 'bar'],
    values = [23, 42],
    object = {},
    key;

while (items.length) {
  key = keys.pop();
  object[key] = values.pop();
}

Use lowerCamelCase for variables, properties and function names

Variables, properties and function names should use lowerCamelCase. They should also be descriptive. Single character variables and uncommon abbreviations should generally be avoided.

Right:

var adminUser = db.query('SELECT * FROM users ...');

Wrong:

var admin_user = db.query('SELECT * FROM users ...');

Use UpperCamelCase for class names

Class names should be capitalized using UpperCamelCase.

Right:

function BankAccount() {
}

Wrong:

function bank_Account() {
}

Use UPPERCASE for Constants

Constants should be declared as regular variables or static class properties, using all uppercase letters.

Node.js / V8 actually supports mozilla's const extension, but unfortunately that cannot be applied to class members, nor is it part of any ECMA standard.

Right:

var SECOND = 1 * 1000;

function File() {
}
File.FULL_PERMISSIONS = 0777;

Wrong:

const SECOND = 1 * 1000;

function File() {
}
File.fullPermissions = 0777;

Object / Array creation

Use trailing commas and put short declarations on a single line. Only quote keys when your interpreter complains:

Right:

var a = ['hello', 'world'];
var b = {
  good: 'code',
  'is generally': 'pretty',
};

Wrong:

var a = [
  'hello', 'world'
];
var b = {"good": 'code'
        , is generally: 'pretty'
        };

Use the === operator

Programming is not about remembering stupid rules. Use the triple equality operator as it will work just as expected.

Right:

var a = 0;
if (a === '') {
  console.log('winning');
}

Wrong:

var a = 0;
if (a == '') {
  console.log('losing');
}

Use multi-line ternary operator

The ternary operator should not be used on a single line. Split it up into multiple lines instead.

Right:

var foo = (a === b)
  ? 1
  : 2;

Wrong:

var foo = (a === b) ? 1 : 2;

Do not extend built-in prototypes

Do not extend the prototype of native JavaScript objects. Your future self will be forever grateful.

Right:

var a = [];
if (!a.length) {
  console.log('winning');
}

Wrong:

Array.prototype.empty = function() {
  return !this.length;
}

var a = [];
if (a.empty()) {
  console.log('losing');
}

Use descriptive conditions

Any non-trivial conditions should be assigned to a descriptively named variable or function:

Right:

var isValidPassword = password.length >= 4 && /^(?=.*\d).{4,}$/.test(password);

if (isValidPassword) {
  console.log('winning');
}

Wrong:

if (password.length >= 4 && /^(?=.*\d).{4,}$/.test(password)) {
  console.log('losing');
}

Write small functions

Keep your functions short. A good function fits on a slide that the people in the last row of a big room can comfortably read. So don't count on them having perfect vision and limit yourself to ~15 lines of code per function.

Return early from functions

To avoid deep nesting of if-statements, always return a function's value as early as possible.

Right:

function isPercentage(val) {
  if (val < 0) {
    return false;
  }

  if (val > 100) {
    return false;
  }

  return true;
}

Wrong:

function isPercentage(val) {
  if (val >= 0) {
    if (val < 100) {
      return true;
    } else {
      return false;
    }
  } else {
    return false;
  }
}

Or for this particular example it may also be fine to shorten things even further:

function isPercentage(val) {
  var isInRange = (val >= 0 && val <= 100);
  return isInRange;
}

Name your closures

Feel free to give your closures a name. It shows that you care about them, and will produce better stack traces, heap and cpu profiles.

Right:

req.on('end', function onEnd() {
  console.log('winning');
});

Wrong:

req.on('end', function() {
  console.log('losing');
});

No nested closures

Use closures, but don't nest them. Otherwise your code will become a mess.

Right:

setTimeout(function() {
  client.connect(afterConnect);
}, 1000);

function afterConnect() {
  console.log('winning');
}

Wrong:

setTimeout(function() {
  client.connect(function() {
    console.log('losing');
  });
}, 1000);

Use slashes for comments

Use slashes for both single line and multi line comments. Try to write comments that explain higher level mechanisms or clarify difficult segments of your code. Don't use comments to restate trivial things.

Right:

// 'ID_SOMETHING=VALUE' -> ['ID_SOMETHING=VALUE'', 'SOMETHING', 'VALUE']
var matches = item.match(/ID_([^\n]+)=([^\n]+)/));

// This function has a nasty side effect where a failure to increment a
// redis counter used for statistics will cause an exception. This needs
// to be fixed in a later iteration.
function loadUser(id, cb) {
  // ...
}

var isSessionValid = (session.expires < Date.now());
if (isSessionValid) {
  // ...
}

Wrong:

// Execute a regex
var matches = item.match(/ID_([^\n]+)=([^\n]+)/));

// Usage: loadUser(5, function() { ... })
function loadUser(id, cb) {
  // ...
}

// Check if the session is valid
var isSessionValid = (session.expires < Date.now());
// If the session is valid
if (isSessionValid) {
  // ...
}

Object.freeze, Object.preventExtensions, Object.seal, with, eval

Crazy shit that you will probably never need. Stay away from it.

Getters and setters

Do not use setters, they cause more problems for people who try to use your software than they can solve.

Feel free to use getters that are free from side effects, like providing a length property for a collection class.