Lightweight, non-opinionated Continuous Integration
Ruby HTML
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
bin
docs
lib
test
.gitignore
.rubocop.yml
CHANGELOG.md
FUTURE.md
Gemfile
Gemfile.lock
LICENSE
README.md
Rakefile
config.defaults.yaml
robot_sweatshop.gemspec

README.md

Gem Version Freenode IRC

Robot Sweatshop

Jenkins is horrible to maintain and is problematic when automating installation and configuration. Drone assumes that you use Docker. Travis-CI is difficult to self-host. All of these frameworks are highly opinionated in one way or another, forcing you to do things their way.

Robot Sweatshop is a single-purpose CI server that runs collections of arbitrary scripts when it needs to, usually when new code is pushed. There's no assumptions about what you want to report, what front-end you need, or even what repositories you want to clone because you can do that better than I can. It's just you, your code, and the scripts that test and deploy it.

Quick start

  • install ZMQ as described in the EZMQ gem
  • gem install robot_sweatshop
  • sweatshop start
  • sweatshop plan example --auto
  • curl -d '' localhost:10555/run/example
  • cat .robot_sweatshop/log/worker.log

Usage

Drop the --auto flag to create the job interactively. You can specify which branches will trigger the job, which commands will be run, and any environment variables you might need. See the wiki for more details.

Robot Sweatshop uses Eye to handle its processes so you can use its commandline tool to monitor their status.

Oh yeah, and there's a front end in case you want to run and watch jobs without using a console. By default it runs on port 10554.

The implicit job

Just like Travis-CI will look for .travis.yml, Robot Sweatshop will look for .robot_sweatshop.yaml at the end of a job and run the job within immediately after. See the wiki for more details on how you might utilize this.

This is useful because it allows you to have a more centralized source of truth for building and testing your code. You can set up your canonical testing definition in your repository and have it pulled down for any server that runs it.

Configuration

By default, Robot Sweatshop looks in your current working path to configure and run. You can supply a custom configuration with sweatshop config [local|user|system]. See the wiki for more information.

Supported payload formats

  • Github (application/json format only)
  • Bitbucket
  • JSON
  • Empty

Security

You probably don't want to run Robot Sweatshop as a sudo user. Create a sandboxed user and group and run sweatshop start as them.

Extending functionality

For extension scripts and processes, check out Sweatshop Gears with sweatshop-gears --help. This is essentially helper package management that comes bundled with Robot Sweatshop.