Four Kitchens

Enforcing coding standards in Meteor

5 Min. ReadDevelopment

This post is the third in a series of posts about the new JavaScript framework, Meteor. If you are unfamiliar with Meteor, check out the first (introduction post) in this series, or go through the demos available on Meteor’s project website.

In this post, I’m going to discuss the importance of enforcing a strict and clear coding standard in your Meteor applications, and then provide you with some tools that will make the process easy.

What are coding standards?

Coding standards are a collection of guidelines for a language that enforce a specific programming style. These guidelines generally apply to naming conventions, comments, indentation, whitespace, and even source tree structure.

For instance, in your JavaScript coding standard you might mandate that all function declarations should have a space after the function name and before the parenthesis that hold your arguments. In addition you might require your function arguments to be separated not only by a comma but also by a space. And you might want all of your functions to have documentation.

In the above example this function would not be correct:

function buildHugeTaco(tortillaType,meatTypes,hasQueso) {
  return new Taco();

And this function would be correct:

 * Builds a huge taco.
 * @param String tortillaType
 *   Type of tortilla in which taco ingredients should be placed.
 *   Allowed values are 'flour', 'corn', and 'wheat'.
 * @param Array meatType
 *   Types of meat that should be placed in this taco. Should be an array
 *   that contains any of the following values: 'beef', 'chicken', 'pork'.
 * @param Boolean hasQueso
 *   Indicates whether or not this taco should be smothered with queso.
 * @return Taco
 *   Object of type taco, ready for consumption by objects of type Mouth.
function buildHugeTaco (tortillaType, meatTypes, hasQueso) {
  return new Taco();

As you can see following a few guidelines in this example made the buildHugeTaco function a lot more readable by improving it’s formatting and providing relevant important context.

Why are coding standards important?

Following a coding standard is critical for successfully developing a codebase that is maintainable and scalable. It ensures that code is formatted consistently and documented properly, which makes the project a lot easier to work with as it grows. In some cases following a standard can even help developers discover and squash bugs, therefore improving the stability of the application. Most importantly, coding standards enable teams of developers to create consistently formatted codebases regardless of the fact that many different people are adding and changing code.

Coding standards in Meteor applications

I’ve read through a large number of the open source Meteor-based applications that are floating around – people have done some really cool work! But one thing that I’ve noticed is that most Meteor-based applications don’t have a defined or enforced coding standard. Often times the formatting and documentation patterns are inconsistent, especially if multiple people have worked on the project. I recently started working on a Meteor-based application and realized why this is the case: there aren’t really any Meteor-specific tools available for defining and enforcing these standards. So as a part of the project I put together a tool and added it to the boilerplate I mantain. Here’s how it works!

JSCS and JSHint

JSCS is a wonderful node module that allows one to define a style guide and then enforce it by programmatically scanning the codebase for code blocks that don’t adhere to the guideline.

JSHint is a tool that detects and reports errors and potential problems in the codebase like syntax errors, implicit type conversions, and leaky variables.

These are the tools that many people use to keep their JavaScript code bases consistent, clean, and error-free. So why not use them within Meteor projects? Here’s how you do it.

Using Gulp to run JSCS and JShint

Gulpjs is an automation tool with which one can define tasks that do useful things with files. In this instance, we’re going to use gulp to run jscs and jshint on certain files within our Meteor application. Here’s an example gulpfile that will do just that:

 * @file
 * Portable Gulp tool that checks a Meteor installation for js syntax errors.
/* globals require */

var gulp = require('gulp'),
    jshint = require('gulp-jshint'),
    jscs = require('gulp-jscs');

 * @task JavaScript lint.
 *   Runs JSCS and JSHint on server, client, lib, and gulp files.
gulp.task('lintjs', function () {
  return gulp.src([

 * @task JavaScript/JSON watch.
 *   Watches changes on relevant js and json files and reports accordingly.
gulp.task('watch', function () {[
  ], ['lintjs']);

The lintjs task runs JSCS and JSHint on the JavaScript files within it’s project but excludes the Meteor core and package JavaScript files so that you don’t get error reports on other people’s code. Jscs and JSHint are configured by .jscsrc and .jshintrc, which should live in the same directory as the gulp file.

The watch task watches the codebase for changes, and when a change is made to a file, it will run the JSCS and JSHint tasks so that as you are developing you will get a real-time report on any errors you might have made.

For a better example see the .jscsrc, .jshintrc, package.json, and gulpfile.js in the Meteor boilerplate. Feel free to use these files in your own projects and if you have improvements to make feel free to fork it, add your updates, and make a pull request!

The future of coding standards in Meteor

If you’ve worked with Meteor before, you might be wondering why I used a gulpfile to scan the codebase instead of writing a package that contained a source handler, and scanned the files during Meteor’s build process. This process would make a lot of sense, however Meteor only allows one source handler to be defined per file type, which is reasonable since you probably shouldn’t have multiple compilers for a single language within a project. There is an issue on the Meteor Github project that defines the lack of a build step that would enable packages to implement code sniffers and proposes a solution that may be built in the near future.

Stay tuned for an update! I will write a follow-up to this post as soon as Meteor core includes an API for scanning files during the build process. In the meantime, like I mentioned before, if you have ideas for improving the gulp-based JSCS/JSHint tools that are in the boilerplate, feel free to fork and contribute!

This post was originally published on my personal blog here.