FUCK SETI. Run specs with your spare CPU cycles.
- Spin up a manager on each machine that can run your specs.
- Start a dispatcher in your project directory.
- Say farewell to your long coffee breaks.
gem install specjour
Give it a try
specjour starts a dispatcher, a manager, and multiple workers - all
of the componenets necessary for distributing your test suite.
(Read the Rails section first if your project uses Rails)
$ cd myproject $ specjour
Start a manager
specjour listen will start a manager which advertises that it's ready
to run specs. By default, the manager runs tests for the project in the
current directory and uses your system cores to determine the number of workers
to start. If your system has two cores, two workers will run tests.
$ specjour listen
Distribute the tests
Dispatch the tests among the managers you started. Specjour checks the 'spec' and 'features' directories for tests.
Distribute a subset of tests
The first parameter to the specjour command is a test directory. It defalts to the current directory and searches for 'spec' and 'features' paths therein.
$ specjour spec # all rspec tests $ specjour spec/models # only model tests $ specjour features # only features $ specjour ~/my_other_project/features
Each worker should run their specs in an isolated database. Modify the test database name in your
config/database.yml to include the following environment variable (Influenced by parallel_tests):
test: database: blog_test<%=ENV['TEST_ENV_NUMBER']%>
specjour prepare will set up the database for each worker.
Specjour contains ActiveRecord hooks that clear database tables before running tests using
DELETE FROM <table_name>;. Additionally, test databases will be created if they don't exist (i.e.
CREATE DATABASE blog_test8 for the 8th worker) and your schema will be loaded when the database is out of date.
Specjour allows you to hook in to the test process on a per-machine and per-worker level through the before_fork and after_fork configuration blocks. If the default ActiveRecord hook doesn't set up the database properly for your test suite, override it with a custom after_fork hook.
# .specjour/hooks.rb # Modify the way you use bundler Specjour::Configuration.before_fork = lambda do system('bundle install --without production') end # Modify your database setup Specjour::Configuration.after_fork = lambda do # custom database setup here end
A preparation hook is run when
specjour prepare is invoked. This hook allows
you to run arbitrary code on all of the listening workers. By default, it drops
and recreates the ActiveRecord database on all workers.
# .specjour/hooks.rb # Modify preparation Specjour::Configuration.prepare = lambda do # custom preparation code end
Customize what gets rsync'd
The standard rsync configuration file may be too broad for your project. If you find you're rsyncing gigs of extraneous data from your public directory, add an exclusion to your projects rsyncd.conf file.
$ vi workbeast/.specjour/rsyncd.conf
Listen for multiple projects
By default, a manager will listen to the project in the current directory. If you want to listen for multiple projects, use the
$ specjour listen --projects bizconf workbeast # run specs for the bizconf and workbeast projects
Give your project an alias
By default, the dispatcher looks for managers matching the project's directory name. If you have multiple teams working on different branches of the same project you may want to isolate each specjour cluster. Give your project an alias and only listen for that alias.
~/bizconf $ specjour listen -p bizconf_08 ~/bizconf $ specjour -a bizconf_08 ~/bizconf $ specjour listen -p bizconf_09 ~/bizconf $ specjour -a bizconf_09
- shayarnett - Cucumber support, pairing and other various patches
- voxdolo - Endless support, alpha testing, various patches
- leshill - Made rsync daemon configurable
- testjour - Ripped off your name
- parallel_tests - Made my test suite twice as fast
Note on Patches/Pull Requests
- Fork the project.
$ source .devto ensure you're using the local specjour binary, not the rubygems version
- Make your feature addition or bug fix.
- Add tests for it. This is important so I don't break it in a future version unintentionally.
- Commit, do not mess with rakefile, version, or history. (if you want to have your own version, that is fine but bump version in a commit by itself I can ignore when I pull)
- Send me a pull request. Bonus points for topic branches.
Copyright (c) 2010 Sandro Turriate. See MIT_LICENSE for details.