The following is a recommended setup for Circle CI 2.0 with any rails project. This guide will walk you through the different parts of the setup to get you familiar with the process.
If you just want the script, it's here
CircleCI 2.0 requires a new dot folder and YAML config as follow: .circleci/config.yml.
Each config must contain the following 2 keys: version and jobs.
Initially, your setup will look like this:
---
version: 2
jobs:
build:
working_directory: ~/your-app-name
steps:
- checkout
This setup won't work, but it provides the framework of the configuration file. The remainder of the guide will explain how to flesh is out.
The docker image will help to compose a consistent environment for each build. There are a set of prebuilt images to choose from, which cover different ruby versions and other common requirements like browser and/or JavaScript.
The following config will set Docker to run ruby 2.5.1 with browser in the box:
jobs:
build:
working_directory: ~/your-app-name
docker:
- image: circleci/ruby:2.5.1-node-browsers
environment:
RAILS_ENV: test
As with Docker, we need to select an image to preload the correct version of Postgres. The below config uses postgres:9.6.2-alpine and the environment variables are loaded from config/database.yml.
build:
working_directory: ~/your-app-name
docker:
- image: circleci/ruby:2.5.1-node-browsers
environment:
RAILS_ENV: test
- image: circleci/postgres:9.6.2-alpine
environment:
POSTGRES_USER: postgres
POSTGRES_DB: test_database_name
Circle CI 2.0 uses the steps key to determine the order in which the build is processed. These steps include checkout, run, restore_cache etc.
We almost always use figaro gem for storing and loading environment variables, so firstly we must create a copy of the application.yml file:
- run:
name: Create environment variables file
command: |
cp config/examples/application.yml config/application.yml
Where name will set name of the task. And command will be exact bash cmd to be executed.
The run key gives lots of flexibility to run any command and in any order. Just give them a name, a command to run and off you go.
Rather than running bundle for each build, which can be slow, we can cache the bundle to reduce the time the build takes to run:
- restore_cache:
keys:
- v1-dependencies-{{ checksum "Gemfile.lock" }}
- v1-dependencies-
- run:
name: Install dependencies
command: |
bundle install --jobs=4 --retry=3 --path vendor/bundle
- save_cache:
paths:
- ./vendor/bundle
key: v1-dependencies-{{ checksum "Gemfile.lock" }}
This will try and restore any cached dependencies using a checksum based on your lock file.
We need to make sure that we avoid certain commands from running before their dependencies are ready. For example, we need to make sure Postgres is up before we run our test suite.
To solve this, we can tell dockerize to wait for a given port before it runs the next command:
- run:
name: Wait for DB
command: dockerize -wait tcp://localhost:5432 -timeout 1m
If you try to run your test suite against an empty DB, it's going to fail. Make sure it's got the correct data structure by loading the schema:
- run:
name: Database setup
environment:
DATABASE_URL: 'postgres://postgres@localhost:5432/test_database_name'
command: |
bundle exec rake db:create db:schema:load --trace
bundle exec rake db:migrate
We can execute as many tasks as we want. Typically we would run our lints, rubocop and our tests rspec, so here is the configuration for that:
By creating one run command per task we can give them a descriptive name and keep them as separate processes during the build.
- run:
name: Rubocop
command: bundle exec rubocop
- run:
name: Tests
environment:
DATABASE_URL: 'postgres://postgres@localhost:5432/test_database_name'
command: |
bundle exec rspec spec --format progress
---
version: 2
jobs:
build:
working_directory: ~/your-app-name
docker:
- image: circleci/ruby:2.5.1-node-browsers
environment:
RAILS_ENV: test
RACK_ENV: test
- image: circleci/postgres:9.6.2-alpine
environment:
POSTGRES_USER: postgres
POSTGRES_DB: test_database_name
steps:
- checkout
- run:
name: Create environment variables file
command: |
cp config/examples/application.yml config/application.yml
- restore_cache:
keys:
- v1-dependencies-{{ checksum "Gemfile.lock" }}
- v1-dependencies-
- run:
name: Install dependencies
command: |
bundle install --jobs=4 --retry=3 --path vendor/bundle
- save_cache:
paths:
- ./vendor/bundle
key: v1-dependencies-{{ checksum "Gemfile.lock" }}
- run:
name: Wait for DB
command: dockerize -wait tcp://localhost:5432 -timeout 1m
- run:
name: Database setup
environment:
DATABASE_URL: 'postgres://postgres@localhost:5432/test_database_name'
command: |
bundle exec rake db:create db:schema:load --trace
bundle exec rake db:migrate
- run:
name: Rubocop
command: bundle exec rubocop
- run:
name: run tests
environment:
DATABASE_URL: 'postgres://postgres@localhost:5432/test_database_name'
command: |
mkdir /tmp/test-results
bundle exec rspec spec --format progress