documentation/content/en/development/overview.md

title	description	menu
Overview	How to setup a development environment for Mastodon	docs parent weight development 1

Mastodon is a Ruby on Rails application with a React.js front-end. It follows standard practices of those frameworks, so if you are already familiar with Rails or React.js, you will not find any surprises here.

The best way of working with Mastodon in a development environment is installing all the dependencies on your system, rather than using Docker or Vagrant. You need Ruby, Node.js, PostgreSQL and Redis, which is a pretty standard set of dependencies for Rails applications.

Environments

An "environment" is a set of configuration values intended for a specific use case. Some environments could be: development, in which you intend to change the code; test, in which you intend to run the automated test suite; staging, which is meant to preview the code to end-users; and production, which is intended to face end-users. Mastodon comes with configurations for development, test and production.

The default value of RAILS_ENV is development, so you don't need to set anything extra to run Mastodon in development mode. In fact, all of Mastodon's configuration has correct defaults for the development environment, so you do not need an .env file unless you need to customize something. Here are some of the different behaviours between the development environment and the production environment:

• Ruby code reloads itself when you change it, which means you don't need to restart the Rails server process to see changes
• All errors you encounter show stack traces in the browser, rather than being hidden behind a generic error page
• Webpack runs continuously and re-compiles JS and CSS assets when you change any of the front-end files, and the pages automatically reload
• Caching is disabled by default
• An admin account with the e-mail admin@localhost:3000 and password mastodonadmin is created automatically during db:seed

It should be noted that the Docker configuration distributed with Mastodon is optimized for the production environment, and so is an extremely bad fit for development. The Vagrant configuration, on the other hand, is meant specifically for development and not production use.

Setup

Run following commands in the project directory bundle install, yarn install.

In the development environment, Mastodon will use PostgreSQL as the currently signed-in Linux user using the ident method, which usually works out of the box. The one command you need to run is rails db:setup which will create the databases mastodon_development and mastodon_test, load the schema into them, and then create seed data defined in db/seed.rb in mastodon_development. The only seed data is an admin account with the credentials admin@localhost:3000 / mastodonadmin.

Please keep in mind, by default Mastodon will run on port 3000. If you configure a different port for it, the generated admin account will use that number.

Running

There are multiple processes that need to be run for the full set of Mastodon's functionality, although they can be selectively omitted. To run all of them with just one command, you can install Foreman with gem install foreman --no-document and then use:

foreman start

In the Mastodon directory. This will start processes defined in Procfile.dev, which will give you: A Rails server, a Webpack server, a streaming API server, and Sidekiq. Of course, you can run any of those things stand-alone depending on your needs.

Testing

Command	Description
rspec	Run the Ruby test suite
yarn run test	Run the JavaScript test suite
rubocop	Check the Ruby code for conformance with our code style

Most notable libraries used

Knowledge and understanding of these libraries will simplify work with the Mastodon code.

Ruby

• haml, a templating language
• devise, for authentication
• doorkeeper, for acting as an OAuth 2 provider
• paperclip, for file uploads and attachments
• sidekiq, for background processing

JavaScript

• immutable, for immutable data structures
• react, for rendering the dynamic web application
• react-redux, for managing React state
• react-router-dom, for navigation within React
• react-intl, for localizations within React

Code structure

The following overview should not be seen as complete or authoritative, but as a rough guidance to help you find your way in the application.

Ruby

Path	Description
app/controllers	Code that binds business logic to templates
app/helpers	Code that can be used from views, i.e. common operations
app/lib	Code that doesn't fit in the other categories
app/models	Code that represents data entities
app/serializers	Code that generates JSON from models
app/services	Complex logical operations involving multiple models
app/views	Templates for generating HTML or other output
app/workers	Code that executes outside the request-response cycle
spec	Automated test suite

JavaScript

Path	Description
app/javascript/mastodon	Code for the multi-column React.js application
app/javascript/packs	Code for non-React.js pages

CSS and other assets

Path	Description
app/javascript/images	Images
app/javascript/styles	Code that turns into CSS via Sass

Localizations

Path	Description
config/locales	Server-side localizations in the YML format
app/javascript/mastodon/locales	Client-side localizations in the JSON format

Localization maintenance

All locale files are normalized to ensure consistent formatting and key order, which minimizes changesets in version control.

Command	Description
i18n-tasks normalize	Normalize server-side translations
yarn run manage:translations	Normalize client-side translations