How to reuse steps for different branches

[fuxx]

I've recently started using Pipelines for my project and I solved most of my problems from testing to deployment with Rails. 

When I used Pipelines for my feature and unit tests only I have handled everything in my default step because I wanted to know if my feature branches broke things. Since I want to use Pipelines as continuous deployment too I've found out that my default step is not invoked as soon as you specify something for a specific branch. The question is if there is a way to reuse code within the YAML file as my file has a lot of duplicated lines already. Or is there maybe a better way as what I do?

This is my current pipeline file, I have changed some values for privacy reasons.

image: ruby:2.5.0

pipelines:
  default:
  - step:
      caches:
      - bundler
      script:
      - curl -sL https://deb.nodesource.com/setup_8.x | bash -
      - apt-get update && apt-get install -y nodejs apt-transport-https
      - apt-get install -y build-essential libpq-dev cron vim imagemagick mysql-client
      - export DATABASE_URL=mysql2://test_user:test_user_password@127.0.0.1/example_ci_test
      - cp config/database.ci.yml config/database.yml && cp config/secrets.ci.yml config/secrets.yml
      - bundle install --retry 5 --path vendor
      - RAILS_ENV=test bundle exec rake db:drop db:create db:migrate
      - bundle exec rspec --format documentation
      services:
      - redis
      - mysql
  master:
  - step:
      caches:
      - bundler
      script:
      - curl -sL https://deb.nodesource.com/setup_8.x | bash -
      - apt-get update && apt-get install -y nodejs apt-transport-https
      - apt-get install -y build-essential libpq-dev cron vim imagemagick mysql-client
      - export DATABASE_URL=mysql2://test_user:test_user_password@127.0.0.1/example_ci_test
      - cp config/database.ci.yml config/database.yml && cp config/secrets.ci.yml config/secrets.yml
      - bundle install --retry 5 --path vendor
      - RAILS_ENV=test bundle exec rake db:drop db:create db:migrate
      - bundle exec rspec --format documentation
      services:
      - redis
      - mysql
  - step:
      caches:
      - bundler
      name: Deploy to integration
      deployment: staging
      script:
      - curl -sL https://deb.nodesource.com/setup_8.x | bash -
      - apt-get update && apt-get install -y nodejs apt-transport-https
      - cp config/database.ci.yml config/database.yml && cp config/secrets.ci.yml config/secrets.yml
      - gem install bundler
      - RAILS_ENV=production bundle install --path vendor
      - export DEPLOY_PATH=/home/x/integration.intern.x.com
      - export SERVER_HOST=88.x.x.x
      - export SERVER_PORT=22
      - export SERVER_USER=x
      - RAILS_ENV=production bundle exec rake docker:deploy
  - step:
      caches:
      - bundler
      name: Deploy to production
      deployment: production
      trigger: manual
      script:
      - curl -sL https://deb.nodesource.com/setup_8.x | bash -
      - apt-get update && apt-get install -y nodejs apt-transport-https
      - cp config/database.ci.yml config/database.yml && cp config/secrets.ci.yml config/secrets.yml
      - gem install bundler
      - RAILS_ENV=production bundle install --path vendor
      - export DEPLOY_PATH=/home/x/intern.x.com
      - export SERVER_HOST=88.x.x.x
      - export SERVER_PORT=22
      - export SERVER_USER=x
      - RAILS_ENV=production bundle exec rake docker:deploy

definitions:
  caches:
    bundler: ./vendor
  services:
    redis:
      image: redis
    mysql:
      image: mysql:5.7
      environment:
        MYSQL_DATABASE: example_ci_test
        MYSQL_RANDOM_ROOT_PASSWORD: 'yes'
        MYSQL_USER: test_user
        MYSQL_PASSWORD: test_user_password

Thanks in advance!

[Graham Gatus · Accepted solution]

@fuxx, you can use YAML anchors and references to reduce duplication in your pipelines yaml file, which are part of the YAML specification.

e.g Below I've step up 2 steps, with both steps re-using whats been defined in 'commonStep'.

commonStep: &commonStep 
  step: 
    script: 
      - echo "hello" 
      - echo "world"

pipelines: 
  default: 
    - <<: *commonStep 
    - <<: *commonStep

The first block of yaml steps up a step and assigns an anchor to it. We can then reference the anchor and inline the yaml associated with that anchor using '<<:'.

[Daniel Santos]

I just want to add that we have the following page as a reference on this topic:

• YAML anchors - Atlassian Documentation

[Kevin Coy]

It seems like if I use an anchor to create a Maven settings.xml file via a script in a step that it doesn't get found in the subsequent steps. Is this expected? Is there a workaround?

[raymond Kelly]

For more info see https://community.atlassian.com/t5/Bitbucket-questions/Bitbucket-pipeline-reuse-of-code/qaq-p/1134797

[Rohan Jaiswal]

@Daniel Santos 

Maybe add some more keywords like "reusable steps" or "step as variables"

I was searching for this document for hours, but passed by YAML anchors many times

• YAML anchors - Atlassian Documentation