Rspec::Bash
Run your shell script in a mocked environment to test its behavior using RSpec.
Features
- Test bash functions, entire scripts and inline scripts
- Stub shell commands and their exitstatus and outputs
- Partial mocks of functions
- Control multiple outputs (through STDOUT, STDERR or files)
- Verify STDIN, STDOUT, STDERR
- Verify command was called with specific argument sequence
- Verify command was called correct number of times
- Supports RSpec matchers
Installation
Add this line to your application's Gemfile:
gem 'rspec-bash'And then execute:
$ bundle
Or install it yourself as:
$ gem install rspec-bash
You can setup rspec-bash globally for your spec suite:
in spec_helper.rb:
require 'rspec/bash'
RSpec.configure do |c|
c.include Rspec::Bash
endUsage
see specs in spec/integration folder:
Running script through stubbed env:
require 'rspec/bash'
describe 'my shell script' do
include Rspec::Bash
let(:stubbed_env) { create_stubbed_env }
it 'runs the script' do
stdout, stderr, status = stubbed_env.execute(
'my-shell-script.sh',
{ 'OPTIONAL_ENV' => 'env vars' }
)
expect(status.exitstatus).to eq 0
end
endChoosing a stub type
The stubbed functions/commands for the test runner were traditionally built with ruby. A faster alternative stub, written in bash, has recently been introduced, and is the current default. This can be configured, however, as shown. For ruby:
let(:stubbed_env) { create_stubbed_env(StubbedEnv::RUBY_STUB) }For bash (default, if not provided):
let(:stubbed_env) { create_stubbed_env(StubbedEnv::BASH_STUB) }Via environment variable:
export RSPEC_BASH_STUB_TYPE=ruby_stub
# <run your tests here>
...Stubbing commands:
let(:stubbed_env) { create_stubbed_env }
let!(:bundle) { stubbed_env.stub_command('bundle') }
let!(:absolute_command) { stubbed_env.stub_command('/path/to/bundle') }
let!(:relative_command) { stubbed_env.stub_command('./path/to/bundle') }
it 'is stubbed' do
stubbed_env.execute 'my-script.sh'
expect(bundle).to be_called_with_arguments('install')
expect(absolute_command).to be_called_with_arguments('hello')
expect(relative_command).to be_called_with_arguments('world')
endChanging exitstatus of stubs:
stubbed_env.stub_command('rake').returns_exitstatus(5) stubbed_env.stub_command('rake').with_args('spec').returns_exitstatus(3)Stubbing output:
let(:rake_stub) { stubbed_env.stub_command('rake') }
rake_stub.outputs('informative message', to: :stdout)
.outputs('error message', to: :stderr)
.outputs('log contents', to: 'logfile.log')
# creates 'prefix-foo.txt' when called as 'rake convert foo'
.outputs('converted result', to: ['prefix-', :arg2, '.txt'])Verifying stdin:
let(:stubbed_env) { create_stubbed_env }
it 'verifies stdin with no args' do
cat_stub = stubbed_env.stub_command('cat')
expect(cat_stub.stdin).to eql 'hello'
end
it 'verifies stdin with args' do
mail_stub = stubbed_env.stub_command('mail')
expect(mail_stub.with_args('-s', 'hello').stdin).to eql 'world'
endTest entire script, specific function or inline script
stubbed_env.execute('./path/to/script.sh')stubbed_env.execute_function(
'./path/to/script.sh',
'overridden_function'
)stubbed_env.execute_inline(<<-multiline_script
stubbed_command first_argument second_argument
multiline_script
)Check that mock was called with specific arguments
stubbed_env.execute_inline(<<-multiline_script
stubbed_command first_argument second_argument
multiline_script
)
it 'correctly identifies the called arguments' do
expect(@command).to be_called_with_arguments('first_argument', 'second_argument')
endCheck that mock was not called with any arguments
expect(@command).to be_called_with_no_argumentsCheck that mock was called a certain number of times
@actual_stdout, @actual_stderr, @actual_status = stubbed_env.execute_inline(<<-multiline_script
stubbed_command duplicated_argument
stubbed_command duplicated_argument
stubbed_command once_called_argument
multiline_script
expect(@command).to be_called_with_arguments('duplicated_argument').times(2)
expect(@command).to be_called_with_arguments('once_called_argument').times(1)
endSupports RSpec matchers
it 'stub call with a wildcard used for an argument' do
grep_mock = stubbed_env.stub_command('grep')
grep_mock.with_args('-r', anything).outputs('output from grep')
expect(command).to be_called_with_arguments('output from grep')
endit 'correctly matches when wildcard is used for arguments' do
expect(@command).to be_called_with_arguments(anything, 'second_argument', anything)
endit 'matches any arguments' do
expect(@command).to be_called_with_arguments
endit 'matches all arguments' do
expect(@command).to be_called_with_arguments(any_args)
endit 'matches any String argument' do
expect(@command).to be_called_with_arguments(instance_of(String))
endit 'matches using regexp' do
expect(@command).to be_called_with_arguments(/s..arg/)
endPitfalls and known issues
-
Use
$BASH_SOURCE[0]instead of$0in your Bash scripts when trying to get the directory that your called script is in. This is a good habit to use when writing scripts as$0should rarely be used.$0also has some ramifications when using this gem; it will always bebashand will not be the name of the script. Please see https://www.gnu.org/software/bash/manual/bashref.html#Positional-Parameters for more information on$0 -
The
execute_function()method is recommended to be used only when testing Bash libraries. This is because it needs to source the entire file to run the function under test, so any executable code in the script will be run even if it is outside of the function being tested -
The current form of stub injection does not allow for stubs to be picked up by other commands. Ex.
xargs stubbed_commandwill result in thestubbed_commandnot being found. There is a pending issue for this.
More examples
see the spec/integration folder
Supported ruby versions
Ruby 2+, no JRuby, due to issues with Open3.capture3
Contributing
- Fork it (https://github.com/mdurban/rspec-bash)
- Create your feature branch (
git checkout -b my-new-feature) - Commit your changes (
git commit -am 'Add some feature') - Push to the branch (
git push origin my-new-feature) - Create a new Pull Request