This guide will help you get started with a Ruby project on Semaphore. If you’re new to Semaphore please read our Guided tour first.
# .semaphore/semaphore.yml version: v1.0 name: Ruby example agent: machine: type: e1-standard-2 os_image: ubuntu1804 blocks: - name: Hello world task: jobs: - name: Run some code commands: - ruby -e 'puts "evol".reverse'
Ruby on Rails example project#
Semaphore provides a tutorial and demo Rails application with a working CI pipeline that you can use to get started quickly:
Supported Ruby versions#
Semaphore supports all versions of Ruby. You have the following options:
- Linux: Ruby is available out-of-the-box in the Ubuntu 18.04 VM image.
- macOS: Ruby is available out-of-the-box in the macOS VM image.
- Docker: use semaphoreci/ruby or your own Docker image with the version of Ruby and other packages that you need.
Follow the links above for details on currently available language versions and additional tools.
Selecting a Ruby version on Linux#
On Linux, Semaphore uses rbenv to manage the supported
Ruby versions. Any Ruby version listed in
Ubuntu image reference
is supported and can be used on Semaphore jobs. The default version is set by the
.ruby-version file in your repository. If the
.ruby-version file is present,
then you cannot force any other Ruby version for that directory. In that case
you will need to modify or delete
.ruby-version in order to choose a different
You can also change the active Ruby version by calling
followed by the desired Ruby version. Using
sem-version ruby [ruby-version] -f the
desired ruby version can be forced. In this case
.ruby-version file is not taken into account.
Here's an example:
blocks: - name: Tests task: prologue: commands: - sem-version ruby 2.5 -f jobs: - name: Tests commands: - ruby --version
If the version of Ruby that you need is not currently available in the Linux VM, we recommend running your jobs in a custom Docker image.
Feature in beta
Beta features are subject to change.
General test summary guidelines are available here ↗.
Generating JUnit XML report#
In this guide we will focus on
RSpec test runner. If you use TestUnit you can see look for runners that supports junit file generation ↗
We will start with adding rspec_junit_formatter ↗ to your
gem "rspec_junit_formatter", :group => [:test]
Then installing the dependencies:
Now we have to make sure that our formatter is properly configured and used by rspec. There are two ways we can do this:
- Extend your
--format RspecJunitFormatter --out rspec.xml --format documentation
bundle exec rspec --format RspecJunitFormatter --out junit.xml --format documentation
Running your tests with this setup will also generate
junit.xml summary report.
Publishing results to Semaphore#
To make Semaphore aware of your test results you can publish them using test results CLI ↗:
test-results publish junit.xml
We advise to include this call in your epilogue:
epilogue: always: commands: - test-results publish junit.xml
This way even if your job fails(due to the test failures) results will still be published for inspection.
Your CI configuration should look similiar to this:
- name: Tests task: prologue: commands: - checkout - bundle install job: name: "Tests" commands: # Or bundle exec rspec if using .rspec configuration file - bundle exec rspec --format RspecJunitFormatter --out junit.xml --format documentation epilogue: always: commands: - test-results publish junit.xml
You can see how test results are setup in one of our demo projects ↗.
You can use Semaphores
cache command to store and load a gem bundle
and configuration. In the following configuration example, we install
dependencies and warm the cache in the first block, then use the cache
in subsequent blocks.
#.semaphore/semaphore.yml version: v1.0 name: Example Ruby pipeline agent: machine: type: e1-standard-2 os_image: ubuntu1804 blocks: - name: Install dependencies task: jobs: - name: cache bundle commands: - checkout - cache restore - bundle install --deployment --path vendor/bundle - cache store - name: Tests task: prologue: commands: - checkout - cache restore - bundle install --deployment --path vendor/bundle jobs: - name: Test all the things commands: - bundle exec rake test
If you need to clear cache for your project, launch a
cache clear or
cache delete <key>.
Semaphore doesn't set language specific environment variables like
RACK_ENV. You can set these at the task level.
blocks: - name: Tests task: env_vars: - name: RACK_ENV value: test jobs: - name: Everything commands: - bundle exec rake test
Rails 4.1+ supports using the
DATABASE_URL to configure a
database. Older versions of Rails need
to generate a
4.1+ - Using DATABASE_URL#
Rails will read database configuration from
DATABASE_URL if set and
config/database.yml is empty. So, start a database with
DATABASE_URL, and clear the
config/database.yml. Here's an example block:
blocks: - name: Tests task: env_vars: # Matches the configuration used in sem-service - name: DATABASE_URL value: postgresql://postgres@localhost/test?encoding=utf8 - name: RAILS_ENV value: test jobs: - name: Test all the things commands: - checkout - sem-service start postgres # Clear existing config/database.yml - cat /dev/null > config/database.yml - bin/rails db:create - bin/rails test
Older Rails Versions#
If your Rails version does not support
DATABASE_URL, then you can
use a separate
config/database.yml for CI.
.semaphore along side your pipeline
configurations. The configured
password match the
connection info from sem-service.
# .semaphore/database.yml test: adapter: postgresql database: test pool: 5 username: postgres password: ~
RAILS_ENV and replace the existing
before running tests.
blocks: - name: Tests task: env_vars: - name: RAILS_ENV value: test jobs: - name: Test all the things commands: - checkout - sem-service start postgres # Use CI's database.yml - cp .semaphore/database.yml config/database.yml - bin/rails db:create - bin/rails test
C-Extensions & system dependencies#
Projects may need system packages to install gems like
pg. Semaphore provides
sudo access so you may install all required packages. Here's an
example of installing the
blocks: - name: Tests task: prologue: commands: - sudo apt-get update && sudo apt-get install -y libpq-dev - gem install pg jobs: - name: Everything commands: - bundle exec rake test
Capybara is the recommended solution for browser tests in Ruby. The Firefox, Chrome, and Chrome Headless drivers work out of the box.
Refer to the Ubuntu image reference for details on preinstalled browsers and testing tools on Semaphore.
Running RSpec and Cucumber in parallel#
To run Cucumber or RSpec suites in parallel across multiple jobs you can use
To set it up use the following snippet for your RSpec block:
jobs: - name: RSpec parallelism: 5 # Number of jobs to run in parallel commands: - gem install semaphore_test_boosters - rspec_booster --job $SEMAPHORE_JOB_INDEX/$SEMAPHORE_JOB_COUNT # Use environment variable to run portion of a spec suite
The similar setup is also used for Cucumber block:
jobs: - name: Cucumber parallelism: 5 # Number of jobs to run in parallel commands: - gem install semaphore_test_boosters - cucumber_booster --job $SEMAPHORE_JOB_INDEX/$SEMAPHORE_JOB_COUNT # Use environment variable to run portion of a spec suite