Scripts to Rule Them All
Shell
Switch branches/tags
Nothing to show
Clone or download
MikeMcQuaid Merge pull request #38 from nschonni/patch-1
Add Travis-CI badge to README
Latest commit 504fdf4 Feb 19, 2018

README.md

Scripts To Rule Them All

Build Status

This is a set of boilerplate scripts describing the normalized script pattern that GitHub uses in its projects. While these patterns can work for projects based on any framework or language, these particular examples are for a simple Ruby on Rails application.

The Idea

If your scripts are normalized by name across all of your projects, your contributors only need to know the pattern, not a deep knowledge of the application. This means they can jump into a project and make contributions without first learning how to bootstrap the project or how to get its tests to run.

The intricacies of things like test commands and bootstrapping can be managed by maintainers, who have a rich understanding of the project's domain. Individual contributors need only to know the patterns and can simply run the commands and get what they expect.

The Scripts

Each of these scripts is responsible for a unit of work. This way they can be called from other scripts.

This not only cleans up a lot of duplicated effort, it means contributors can do the things they need to do, without having an extensive fundamental knowledge of how the project works. Lowering friction like this is key to faster and happier contributions.

The following is a list of scripts and their primary responsibilities.

script/bootstrap

script/bootstrap is used solely for fulfilling dependencies of the project.

This can mean RubyGems, npm packages, Homebrew packages, Ruby versions, Git submodules, etc.

The goal is to make sure all required dependencies are installed.

script/setup

script/setup is used to set up a project in an initial state. This is typically run after an initial clone, or, to reset the project back to its initial state.

This is also useful for ensuring that your bootstrapping actually works well.

script/update

script/update is used to update the project after a fresh pull.

If you have not worked on the project for a while, running script/update after a pull will ensure that everything inside the project is up to date and ready to work.

Typically, script/bootstrap is run inside this script. This is also a good opportunity to run database migrations or any other things required to get the state of the app into shape for the current version that is checked out.

script/server

script/server is used to start the application.

For a web application, this might start up any extra processes that the application requires to run in addition to itself.

script/update should be called ahead of any application booting to ensure that the application is up to date and can run appropriately.

script/test

script/test is used to run the test suite of the application.

A good pattern to support is having an optional argument that is a file path. This allows you to support running single tests.

Linting (i.e. rubocop, jshint, pmd, etc.) can also be considered a form of testing. These tend to run faster than tests, so put them towards the beginning of a script/test so it fails faster if there's a linting problem.

script/test should be called from script/cibuild, so it should handle setting up the application appropriately based on the environment. For example, if called in a development environment, it should probably call script/update to always ensure that the application is up to date. If called from script/cibuild, it should probably reset the application to a clean state.

script/cibuild

script/cibuild is used for your continuous integration server. This script is typically only called from your CI server.

You should set up any specific things for your environment here before your tests are run. Your test are run simply by calling script/test.

script/console

script/console is used to open a console for your application.

A good pattern to support is having an optional argument that is an environment name, so you can connect to that environment's console.

You should configure and run anything that needs to happen to open a console for the requested environment.