There's no way to definitively run a single test using MiniTest as it currently stands.

minitest.rb

require "minitest/unit"
require "minitest/autorun"

class MyTest < MiniTest::Unit::TestCase
  def test_something
    assert true
  end
end

class AnotherTest < MiniTest::Unit::TestCase
  def test_something
    flunk
  end
end

Output.md

(Here's my proposed fix for this)

First, print out the test names:

$ ruby minitest.rb -v                 
Run options: -v --seed 58440

# Running tests:

AnotherTest#test_something = 0.00 s = F
MyTest#test_something = 0.00 s = .


Finished tests in 0.000664s, 3012.0482 tests/s, 3012.0482 assertions/s.

  1) Failure:
test_something(AnotherTest) [minitest.rb:12]:
Epic Fail!

2 tests, 2 assertions, 1 failures, 0 errors, 0 skips

Next, try running a specific test:

$ ruby minitest.rb -n "AnotherTest#test_something"
Run options: -n AnotherTest#test_something --seed 8523

# Running tests:



Finished tests in 0.000480s, 0.0000 tests/s, 0.0000 assertions/s.

0 tests, 0 assertions, 0 failures, 0 errors, 0 skips

What? Nothing ran? Indeed. It's because as far as MiniTest cares, the suite name isn't part of the test name. So you can only use parts of the test name in your filters:

However, there's no guarantee (nor should there be) that test names will be unique across suites:

$ ruby minitest.rb -n "test_something"            
Run options: -n test_something --seed 30485

# Running tests:

F.

Finished tests in 0.000645s, 3100.7752 tests/s, 3100.7752 assertions/s.

  1) Failure:
test_something(AnotherTest) [minitest.rb:12]:
Epic Fail!

2 tests, 2 assertions, 1 failures, 0 errors, 0 skips

So when we try to target a single test, both tests run. Bums.

(Here's my proposed fix for this).

I've implemented my attempt a fix here: lazyatom/minitest@2b12aaa

Any thoughts?