minitest provides a complete suite of testing facilities supporting TDD, BDD, mocking, and benchmarking.
"I had a class with Jim Weirich on testing last week and we were allowed to choose our testing frameworks. Kirk Haines and I were paired up and we cracked open the code for a few test frameworks... I MUST say that minitest is *very* readable / understandable compared to the 'other two' options we looked at. Nicely done and thank you for helping us keep our mental sanity." -- Wayne E. Seguin
minitest/test is a small and incredibly fast unit testing framework. It provides a rich set of assertions to make your tests clean and readable.
minitest/spec is a functionally complete spec engine. It hooks onto minitest/test and seamlessly bridges test assertions over to spec expectations.
minitest/benchmark is an awesome way to assert the performance of your algorithms in a repeatable manner. Now you can assert that your newb co-worker doesn't replace your linear algorithm with an exponential one!
minitest/mock by Steven Baker, is a beautifully tiny mock (and stub) object framework.
minitest/pride shows pride in testing and adds coloring to your test output. I guess it is an example of how to write IO pipes too. :P
minitest/test is meant to have a clean implementation for language implementors that need a minimal set of methods to bootstrap a working test suite. For example, there is no magic involved for test-case discovery.
"Again, I can't praise enough the idea of a testing/specing framework that I can actually read in full in one sitting!" -- Piotr Szotkowski
Comparing to rspec:
rspec is a testing DSL. minitest is ruby. -- Adam Hawkins, "Bow Before MiniTest"
minitest doesn't reinvent anything that ruby already provides, like: classes, modules, inheritance, methods. This means you only have to learn ruby to use minitest and all of your regular OO practices like extract-method refactorings still apply.
minitest/autorun - the easy and explicit way to run all your tests.
minitest/test - a very fast, simple, and clean test system.
minitest/spec - a very fast, simple, and clean spec system.
minitest/mock - a simple and clean mock/stub system.
minitest/benchmark - an awesome way to assert your algorithm's performance.
minitest/pride - show your pride in testing!
Incredibly small and fast runner, but no bells and whistles.
Written by squishy human beings. Software can never be perfect. We will all eventually die.
See design_rationale.rb to see how specs and tests work in minitest.
Given that you'd like to test the following class:
class Meme def i_can_has_cheezburger? "OHAI!" end def will_it_blend? "YES!" end end
Define your tests as methods beginning with test_
.
require "minitest/autorun" class TestMeme < Minitest::Test def setup @meme = Meme.new end def test_that_kitty_can_eat assert_equal "OHAI!", @meme.i_can_has_cheezburger? end def test_that_it_will_not_blend refute_match /^no/i, @meme.will_it_blend? end def test_that_will_be_skipped skip "test this later" end end
require "minitest/autorun" describe Meme do before do @meme = Meme.new end describe "when asked about cheeseburgers" do it "must respond positively" do _(@meme.i_can_has_cheezburger?).must_equal "OHAI!" end end describe "when asked about blending possibilities" do it "won't say no" do _(@meme.will_it_blend?).wont_match /^no/i end end end
For matchers support check out:
Add benchmarks to your tests.
# optionally run benchmarks, good for CI-only work! require "minitest/benchmark" if ENV["BENCH"] class TestMeme < Minitest::Benchmark # Override self.bench_range or default range is [1, 10, 100, 1_000, 10_000] def bench_my_algorithm assert_performance_linear 0.9999 do |n| # n is a range value @obj.my_algorithm(n) end end end
Or add them to your specs. If you make benchmarks optional, you'll need to wrap your benchmarks in a conditional since the methods won't be defined. In minitest 5, the describe name needs to match /Bench(mark)?$/
.
describe "Meme Benchmark" do if ENV["BENCH"] then bench_performance_linear "my_algorithm", 0.9999 do |n| 100.times do @obj.my_algorithm(n) end end end end
outputs something like:
# Running benchmarks: TestBlah 100 1000 10000 bench_my_algorithm 0.006167 0.079279 0.786993 bench_other_algorithm 0.061679 0.792797 7.869932
Output is tab-delimited to make it easy to paste into a spreadsheet.
Mocks and stubs defined using terminology by Fowler & Meszaros at www.martinfowler.com/bliki/TestDouble.html:
“Mocks are pre-programmed with expectations which form a specification of the calls they are expected to receive. They can throw an exception if they receive a call they don't expect and are checked during verification to ensure they got all the calls they were expecting.”
class MemeAsker def initialize(meme) @meme = meme end def ask(question) method = question.tr(" ", "_") + "?" @meme.__send__(method) end end require "minitest/autorun" describe MemeAsker, :ask do describe "when passed an unpunctuated question" do it "should invoke the appropriate predicate method on the meme" do @meme = Minitest::Mock.new @meme_asker = MemeAsker.new @meme @meme.expect :will_it_blend?, :return_value @meme_asker.ask "will it blend" @meme.verify end end end
Minitest mocks do not support multi-threading. If it works, fine, if it doesn't you can use regular ruby patterns and facilities like local variables. Here's an example of asserting that code inside a thread is run:
def test_called_inside_thread called = false pr = Proc.new { called = true } thread = Thread.new(&pr) thread.join assert called, "proc not called" end
Mocks and stubs are defined using terminology by Fowler & Meszaros at www.martinfowler.com/bliki/TestDouble.html:
“Stubs provide canned answers to calls made during the test”.
Minitest's stub method overrides a single method for the duration of the block.
def test_stale_eh obj_under_test = Something.new refute obj_under_test.stale? Time.stub :now, Time.at(0) do # stub goes away once the block is done assert obj_under_test.stale? end end
A note on stubbing: In order to stub a method, the method must actually exist prior to stubbing. Use a singleton method to create a new non-existing method:
def obj_under_test.fake_method ... end
Ideally, you'll use a rake task to run your tests, either piecemeal or all at once. Both rake and rails ship with rake tasks for running your tests. BUT! You don't have to:
% ruby -Ilib:test test/minitest/test_minitest_test.rb Run options: --seed 37685 # Running: ...................................................................... (etc) Finished in 0.107130s, 1446.8403 runs/s, 2959.0217 assertions/s. 155 runs, 317 assertions, 0 failures, 0 errors, 0 skips
There are runtime options available, both from minitest itself, and also provided via plugins. To see them, simply run with --help
:
% ruby -Ilib:test test/minitest/test_minitest_test.rb --help minitest options: -h, --help Display this help. -s, --seed SEED Sets random seed. Also via env. Eg: SEED=n rake -v, --verbose Verbose. Show progress processing files. -n, --name PATTERN Filter run on /regexp/ or string. -e, --exclude PATTERN Exclude /regexp/ or string from run. Known extensions: pride, autotest -p, --pride Pride. Show your testing pride! -a, --autotest Connect to autotest server.
You can set up a rake task to run all your tests by adding this to your Rakefile:
require "rake/testtask" Rake::TestTask.new(:test) do |t| t.libs << "test" t.libs << "lib" t.test_files = FileList["test/**/test_*.rb"] end task :default => :test
To define a plugin, add a file named minitest/XXX_plugin.rb to your project/gem. That file must be discoverable via ruby's LOAD_PATH (via rubygems or otherwise). Minitest will find and require that file using Gem.find_files. It will then try to call plugin_XXX_init
during startup. The option processor will also try to call plugin_XXX_options
passing the OptionParser instance and the current options hash. This lets you register your own command-line options. Here's a totally bogus example:
# minitest/bogus_plugin.rb: module Minitest def self.plugin_bogus_options(opts, options) opts.on "--myci", "Report results to my CI" do options[:myci] = true options[:myci_addr] = get_myci_addr options[:myci_port] = get_myci_port end end def self.plugin_bogus_init(options) self.reporter << MyCI.new(options) if options[:myci] end end
Minitest uses composite reporter to output test results using multiple reporter instances. You can add new reporters to the composite during the init_plugins phase. As we saw in plugin_bogus_init
above, you simply add your reporter instance to the composite via <<
.
AbstractReporter
defines the API for reporters. You may subclass it and override any method you want to achieve your desired behavior.
Called when the run has started.
Called for each result, passed or otherwise.
Called at the end of the run.
Called to see if you detected any problems.
Using our example above, here is how we might implement MyCI:
# minitest/bogus_plugin.rb module Minitest class MyCI < AbstractReporter attr_accessor :results, :addr, :port def initialize options self.results = [] self.addr = options[:myci_addr] self.port = options[:myci_port] end def record result self.results << result end def report CI.connect(addr, port).send_results self.results end end # code from above... end
The following implementation and test:
class Worker < SimpleDelegator def work end end describe Worker do before do @worker = Worker.new(Object.new) end it "must respond to work" do _(@worker).must_respond_to :work end end
outputs a failure:
1) Failure: Worker#test_0001_must respond to work [bug11.rb:16]: Expected #<Object:0x007f9e7184f0a0> (Object) to respond to #work.
Worker is a SimpleDelegate which in 1.9+ is a subclass of BasicObject. Expectations are put on Object
(one level down) so the Worker (SimpleDelegate) hits method_missing
and delegates down to the Object.new
instance. That object doesn't respond to work so the test fails.
You can bypass SimpleDelegate#method_missing
by extending the worker with Minitest::Expectations
. You can either do that in your setup at the instance level, like:
before do @worker = Worker.new(Object.new) @worker.extend Minitest::Expectations end
or you can extend the Worker class (within the test file!), like:
class Worker include ::Minitest::Expectations end
Use a module. That's exactly what they're for:
module UsefulStuff def useful_method # ... end end describe Blah do include UsefulStuff def test_whatever # useful_method available here end end
Remember, describe
simply creates test classes. It's just ruby at the end of the day and all your normal Good Ruby Rules (tm) apply. If you want to extend your test using setup/teardown via a module, just make sure you ALWAYS call super. before/after automatically call super for you, so make sure you don't do it twice.
Use a constant with begin…end like this:
describe Blah do SETUP = begin # ... this runs once when describe Blah starts end # ... end
This can be useful for expensive initializations or sharing state. Remember, this is just ruby code, so you need to make sure this technique and sharing state doesn't interfere with your tests.
uninitialized constant MiniTest::Test (NameError)
?¶ ↑Are you running the test with Bundler (e.g. via bundle exec
)? If so, in order to require minitest, you must first add the gem 'minitest'
to your Gemfile and run bundle
. Once it's installed, you should be able to require minitest and run your tests.
arel
journey
mime-types
nokogiri
rails (active_support et al)
rake
rdoc
…and of course, everything from seattle.rb…
Minitest requires Hoe.
This is a common problem in Windows, where the default external Encoding is often CP850, but can affect any platform. Minitest can run test suites using any Encoding, but to run Minitest's own tests you must have a default external Encoding of UTF-8.
If your encoding is wrong, you'll see errors like:
--- expected +++ actual @@ -1,2 +1,3 @@ # encoding: UTF-8 -"Expected /\\w+/ to not match \"blah blah blah\"." +"Expected /\\w+/ to not match # encoding: UTF-8 +\"blah blah blah\"."
To check your current encoding, run:
ruby -e 'puts Encoding.default_external'
If your output is something other than UTF-8, you can set the RUBYOPTS env variable to a value of '-Eutf-8'. Something like:
RUBYOPT='-Eutf-8' ruby -e 'puts Encoding.default_external'
Check your OS/shell documentation for the precise syntax (the above will not work on a basic Windows CMD prompt, look for the SET command). Once you've got it successfully outputing UTF-8, use the same setting when running rake in Minitest.
This is also a problem primarily affecting Windows developers. PowerShell has a command called diff, but it is not suitable for use with Minitest.
If you see failures like either of these, you are probably missing diff tool:
4) Failure: TestMinitestUnitTestCase#test_assert_equal_different_long [D:/ruby/seattlerb/minitest/test/minitest/test_minitest_test.rb:936]: Expected: "--- expected\n+++ actual\n@@ -1 +1 @@\n-\"hahahahahahahahahahahahahahahahahahahaha\"\n+\"blahblahblahblahblahblahblahblahblahblah\"\n" Actual: "Expected: \"hahahahahahahahahahahahahahahahahahahaha\"\n Actual: \"blahblahblahblahblahblahblahblahblahblah\"" 5) Failure: TestMinitestUnitTestCase#test_assert_equal_different_collection_hash_hex_invisible [D:/ruby/seattlerb/minitest/test/minitest/test_minitest_test.rb:845]: Expected: "No visible difference in the Hash#inspect output.\nYou should look at the implementation of #== on Hash or its members.\n {1=>#<Object:0xXXXXXX>}" Actual: "Expected: {1=>#<Object:0x00000003ba0470>}\n Actual: {1=>#<Object:0x00000003ba0448>}"
If you use Cygwin or MSYS2 or similar there are packages that include a GNU diff for Windows. If you don't, you can download GNU diffutils from gnuwin32.sourceforge.net/packages/diffutils.htm (make sure to add it to your PATH).
You can make sure it's installed and path is configured properly with:
diff.exe -v
There are multiple lines of output, the first should be something like:
diff (GNU diffutils) 2.8.1
If you are using PowerShell make sure you run diff.exe, not just diff, which will invoke the PowerShell built in function.
Bridge between Capybara RSpec matchers and Minitest::Spec
expectations (e.g. page.must_have_content("Title")
).
Test names print Ruby Object
types in color with your Minitest Spec style tests.
Metadata for describe/it blocks & CLI tag filter. E.g. it "requires JS driver", js: true do
& ruby test.rb --tag js
runs tests tagged :js.
Minimal support to use Spec style in Rails 5+.
for swagger based automated API testing.
Around block for minitest. An alternative to setup/teardown dance.
Adds Minitest assertions to test for errors raised or not raised by Minitest itself.
autotest is a continuous testing facility meant to be used during development.
minitest-bacon extends minitest with bacon-like functionality.
Adds support for RSpec-style let! to immediately invoke let statements before each test.
Helps you isolate and debug random test failures.
Display test results with a Blink1.
Assertions and expectations for testing Capistrano recipes.
Capybara matchers support for minitest unit and spec.
Run Minitest suites as Chef report handlers
CI reporter plugin for Minitest.
Defines contexts for code reuse in Minitest specs that share common expectations.
Wraps assert so failed assertions drop into the ruby debugger.
Patches Minitest to allow for an easily configurable output.
Minimal documentation format inspired by rspec's.
Detailed output inspired by rspec's documentation format.
Print out emoji for your test passes, fails, and skips.
Semantically symmetric aliases for assertions and expectations.
Clean API for excluding certain tests you don't want to run under certain conditions.
Reimplements RSpec's “fail fast” feature
Support unit tests with expectation results in files. Differing results will be stored again in files.
Adds assertion and expectation to help testing filesystem contents.
Makes your Minitest mocks more resilient.
Focus on one test at a time.
A minitest plugin that adds a report of the top tests by number of objects allocated.
Support minitest expectation methods for all objects
Generally useful additions to minitest's assertions and expectations.
Test notifier for minitest via growl.
GLOBALLY ACTIVATE MINITEST PRIDE! RAWR!
Adds Minitest assertions to test for the existence of HTML tags, including contents, within a provided string.
Around and before_all/after_all/around_all hooks
Pretty, single-page HTML reports for your Minitest runs
Implicit declaration of the test subject.
Instrument ActiveSupport::Notifications when test method is executed.
Store information about speed of test execution provided by minitest-instrument in database.
JUnit-style XML reporter for minitest.
Use Minitest assertions with keyword arguments.
Test notifier for minitest via libnotify.
Run test at line number.
Define assert_log and enable minitest to test log messages. Supports Logger and Log4r::Logger.
Provides extensions to minitest for macruby UI testing.
Adds support for RSpec-style matchers to minitest.
Adds assertions that adhere to the matcher spec, but without any expectation infections.
Annotate tests with metadata (key-value).
Provides method call assertions for minitest.
Mongoid assertion matchers for Minitest.
Provides must_not as an alias for wont in Minitest.
Automatically retry failed test to help with flakiness.
Reporter for the Mac OS X notification center.
Fork-based parallelization
Run tests in parallel with a single database.
PowerAssert for Minitest.
Adds support for .predicate? methods.
List the 10 slowest tests in your suite.
Minitest integration for Rails 3.x.
Capybara integration for Minitest::Rails.
Create customizable Minitest output formats.
Colored red/green output for Minitest.
Use RSpec Mocks with Minitest.
minitest-server provides a client/server setup with your minitest process, allowing your test run to send its results directly to a handler.
Minitest assertions to speed-up development and testing of Ruby Sequel database setups.
Support for shared specs and shared spec subclasses
RSpec-style x.should == y
assertions for Minitest.
Adding all manner of shoulds to Minitest (bad idea)
Print a list of tests that take too long
Provides rspec-ish context method to Minitest::Spec
.
Expect syntax for Minitest::Spec
(e.g. expect(sequences).to_include :celery_man).
Minitest::Spec
extensions for Rails and beyond.
Drop in Minitest::Spec
superclass for ActiveSupport::TestCase.
Runs (Get it? It's fast!) your tests and makes it easier to rerun individual failures.
Find leaking state between tests
Stub any instance of a method on the given class for the duration of a block.
Stub constants for the duration of a block.
Add tags for minitest.
Adds a new assertion to minitest for checking the contents of a collection, ignoring element order.
Automatic cassette managment with Minitest::Spec
and VCR.
Get tests results as a TestResult object.
Shoulda style syntax for minitest test::unit.
Bridges between test/unit and minitest.
Minitest matchers for Mongoid.
Minitest integration for mutant.
A pry plugin w/ minitest support. See pry-rescue/minitest.rb.
Easily translate any RSpec matchers to Minitest assertions and expectations.
Authors… Please send me a pull request with a description of your minitest extension.
assay-minitest
detroit-minitest
em-minitest-spec
flexmock-minitest
guard-minitest
guard-minitest-decisiv
minitest-activemodel
minitest-ar-assertions
minitest-capybara-unit
minitest-colorer
minitest-deluxe
minitest-extra-assertions
minitest-rails-shoulda
minitest-spec
minitest-spec-should
minitest-sugar
spork-minitest
minitest/pride fabric: www.spoonflower.com/fabric/3928730-again-by-katie_allen
Ruby 2.3+. No magic is involved. I hope.
sudo gem install minitest
On 1.9, you already have it. To get newer candy you can still install the gem, and then requiring “minitest/autorun” should automatically pull it in. If not, you'll need to do it yourself:
gem "minitest" # ensures you"re using the gem, and not the built-in MT require "minitest/autorun" # ... usual testing stuffs ...
DO NOTE: There is a serious problem with the way that ruby 1.9/2.0 packages their own gems. They install a gem specification file, but don't install the gem contents in the gem path. This messes up Gem.find_files and many other things (gem which, gem contents, etc).
Just install minitest as a gem for real and you'll be happier.
(The MIT License)
Copyright © Ryan Davis, seattle.rb
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the 'Software'), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED 'AS IS', WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.