Fudge is a CI build tool for Ruby/Rails projects.
- Define your CI build process in a Ruby DSL.
- Keep your build definition in your code repository and version control.
- Define different build processes for different branches of your code.
- Run common tasks including code coverage assertion.
- Define custom tasks for your build process.
Add to your project's Gemfile:
gem 'fudge'Run in your project root:
bundle install
To create a blank Fudgefile, run in your project root:
bundle exec fudge init
To run the CI build (this is what you'd put in your CI server):
bundle exec fudge build
This will run the build named 'default'. To run a specific build in the Fudgefile:
bundle exec fudge build the_build_name
To list builds defined in your Fudgefile:
bundle exec fudge list
The list of builds can be filtered to include only those whose name match a given string (case-insensitive):
bundle exec fudge list a_build_name
To define a build with a given name (or for a given branch):
build :some_name do
endAn optional description can be provided by supplying an about string (descriptions, if supplied, are output by the list command):
build :some_name, :about => "Runs Rspec unit tests" do
endTo define some tasks on that build:
build :some_name do
task :rspec, :coverage => 100
endAny options passed to the task method will be passed to the task's initializer.
You can also use one of the alternative method missing syntax:
build :some_name do |b|
rspec :coverage => 100
endSome tasks are composite tasks, and can have tasks added to themselves, for example the each_directory task:
build :some_name do
task :each_directory, '*' do
task :rspec
end
endYou can define task groups to be reused later on in your Fudgefile. For example:
task_group :tests do
rspec
end
task_group :docs do
yard
end
task_group :quality do
cane
end
build :some_name do
task_group :tests
task_group :docs
task_group :quality
end
build :nodoc do
task_group :tests
endTask groups can take arguments, so you can make conditional task groups for sharing between build. For example:
task_group :deploy do |to|
shell "cp -r site/ #{to}"
end
build :default do
task_group :deploy, '/var/www/dev'
end
build :production do
task_group :deploy, '/var/www/live'
endYou can define success and failure callbacks using the following syntax:
build :default do
rspec
on_success do
shell 'deploy.sh'
end
on_failure do
shell 'send_errors.sh'
end
endBuild will by default run without callbacks enabled. To run a build with callbacks, run:
bundle exec fudge build --callbacks
You can mix task groups with callbacks however you like, for example:
task_group :deploy do
shell 'deploy.sh'
end
task_group :error_callbacks do
on_failure do
shell 'send_errors.sh'
end
end
build :default do
on_success do
task_group :deploy
end
task_group :error_callbacks
endFudge supports several tasks by default. Most of them depend on a gem which also must be included in your project's Gemfile (or with add_development_dependency in your gem's .gemspec).
Run the Brakeman Rails security scanner.
task :brakeman
will fail if any security warnings are encountered.
task :brakeman, :max => 2
will allow a maximum of two known security issues to get through.
Checks code style using the cane gem. This can be run over the entire tree or in the enclosing subdirectory (each_directory or in_directory). Options can be set in the .cane file, but for convenience the max_width option can be used to easily override the default line width of 80.
cane :max_width => 120
Ensures that the code block runs in a clean Bundler environment.
Run the resulting block in each directory (see examples above).
each_directory 'meta_*', :exclude => ['meta_useless', 'meta_broken'] do
...
end
Run a rake command, requiring that it return success.
Run a generic shell command, requiring that it return success.
Like shell, but does not spawn a new shell to run the command, and allows more control over the command's process and environment. See examples below.
Code duplication can be detected by Flay. See examples below.
Flog calculates code complexity using an ABC metric and allows for maximum individual values and maximum average values. This can be used to ensure that you are alerted quickly when new complex code is added to your project. See examples below.
Run the resulting block in a specific directory (as with each_directory).
Run rspec, enforcing minimum :coverage as a percent (using simplecov). See examples above.
Run cucumber, enforcing minimum :coverage as a percent (using simplecov), like rspec.
Runs YARD to ensure documentation coverage.
yard 'stats --list-undoc', :coverage => 100
will require 100% coverage, and show all code that is not documented.
A task is a class that responds to two methods:
self.name- A class method that returns a symbol representing the task. This is what will be used to identify your task in the Fudgefile. If not defined, it will be derived from the class name (e.g. in below example, it will be:loud_task).run- An instance method which carries out the contents of the task. Should returntrueorfalsedepending on whether the task succeeded.
For example, here is a simple task which will print some output and always pass:
class LoudTask < Fudge::Tasks::Task
def self.name
:loud
end
def run
puts "I WAS RUN"
true
end
endTo make your task available to Fudge, you simply register it in Fudge::Tasks:
require 'fudge'
Fudge::Tasks.register(LoudTask)This will make the LoudTask task available in your Fudgefile's like so:
build :some_name do
task :loud
endMany tasks simply run a shell command and may accept some extra configuration options. To define a task of this kind, you can sublcass Shell and simply define the cmd method:
class LsTask < Fudge::Tasks::Shell
def cmd
"ls #{arguments}"
end
endThe arguments method is provided by the Shell base class and will be a string of all other positional arguments passed to the task. For example:
build :default do
task :ls, '-l', '-a'
endwould run the command ls -l -a.
You can take hash-like options, which will automatically be set if you have an attribute with the same name. For example:
class LsTask < Fudge::Tasks::Shell
attr_accessor :all
def cmd
arguments << ' -a' if all
"ls #{arguments}"
end
endNow this task can be used like so:
build :default do
task :ls, :all => true
endYou can define that some output from a command is required by responding to check_for with a regexp. For example:
class LsTask < Fudge::Tasks::Shell
def cmd
"ls #{arguments}"
end
def check_for
/4 files found/
end
endThe above task will only pass if the output contains "4 files found".
If you want to do some further processing on the contents matched by the regexp, you can provide an array with the second element being a lambda, which wil be called to process the output:
class LsTask < Fudge::Tasks::Shell
def cmd
"ls #{arguments}"
end
def check_for
[/(\d+) files found/, lambda { |n| n.to_i >= 4 }]
end
endThe above task will only pass if the output contains "n files found", where n is a number, and also n is at least 4.
This task is useful if you want to set an environment variable for a shell command, but the command won't allow the variable to be supplied at the end of the command line. That is, if something like this doesn't work because the command treats the variable assignment as a parameter:
task :shell, 'awkward_command SOME_VAR=true'
# This won't work either because shell tries to run a command called SOME_VAR=:
task :shell, 'SOME_VAR=true awkward_command'SubProcess allows you to set the variable this way:
task :sub_process, 'awkward_command', :environment => { 'SOME_VAR' => 'true' }SubProcess is also useful if you need to manipulate the process's execution environment, for example, by clearing environment variables, or redirecting IO. For example, this invocatin will unset all environment variables, except SOME_VAR which is explicitly supplied, before running command:
task :sub_process, 'command', :environment => { 'SOME_VAR' => 'true' },
:spawn_options => { :unsetenv_others => true }See the Ruby doc for Process::spawn for details of the options that can be passed in spawn_options.
This task should otherwise act like the Shell task.
Some tasks may require you to run a number of commands one after the other. You can hook into other fudge tasks by including the Fudge DSL into your composite task:
class DeployTask < Fudge::Tasks::CompositeTask
include Fudge::TaskDSL
def self.name
:deploy
end
def initialize(*args)
super
task :shell, 'build_docs.sh'
task :shell, 'cp -r docs/ /var/ww/deploy/docs'
end
end
Fudge::Tasks.register(DeployTask)The above will run the given tasks in the order defined, and only pass if both tasks pass. It can then be used in a FudgeFile like so:
build :default do
task :deploy
endSometimes you'll want different options to be used for specific subdirectories. This is especially useful with code metric tools.
Instead of having all of these values listed explicitly in your Fudgefile you can instead place them in a fudge_settings.yml file in each subdirectory.
So instead of this in your Fudgefile...
in_directory 'meta_addresses' do
task :flay, :exclude => '^\.\/(db|factories|spec)\/'
task :flog, :exclude => '^\.\/(db|factories|spec)\/', :max => 20, :average => 5, :methods => true
end
in_directory 'meta_banks' do
task :flay, :exclude => '^\.\/(db|factories|spec)\/', :max => 172
task :flog, :exclude => '^\.\/(db|factories|spec)\/', :max => 74.9, :average => 9.1, :methods => true
endyou can just have this:
each_directory 'meta_*' do
task :flay, :exclude => '^\.\/(db|factories|spec)\/'
task :flog, :exclude => '^\.\/(db|factories|spec)\/', :methods => true
endand this in your meta_addresses/fudge_settings.yml:
flog:
max: 20
average: 5
and this in your meta_banks/fudge_settings.yml:
flay:
max: 172
flog:
max: 74.9
average: 9.1
You can set the default values in your Fudgefile and override them only as necessary in specific subdiretories.
This gem is available as open source under the terms of the MIT licence.
Copyright (c) 2018 Sage Group Plc. All rights reserved.
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent