Question

I’m using Riot (https://github.com/thumblemonks/riot) for testing though we could use something else.

I get test output like this:

> ruby my_test_file.rb

Running a test
  + something works
  + something else works
  - something failed

I’d like to programmatically access this information. Something like:

test = TestClass.load("my_test_file.rb")
result = test.run
result.errors # some array

Answer 1

Answering my own question. I can post a more detailed answer if anyone’s interested.

The gist of it is evaluating the test files then getting the context from the evaluated test:

Riot.alone! # so Riot doesn't automatically run the tests

["/path/to/test/file"].each do |test_file|
  eval(IO.read(path), binding, test_file) # load test file

  Riot.root_contexts.each do |context|
    reporter = MyReporter.new
    context.run reporter
    # do something cool with reporter results   
  end

  Riot.root_contexts.clear # clean out root_contexts so it's clean for the next run
end

Where MyReporter is your implementation that handles pass/fail test results. See https://github.com/thumblemonks/riot/blob/master/lib/riot/reporter.rb