How To Test Ruby CLI: Code Generators

One of the most interesting challenges I met during Hanami development is to write isolated integration tests. The framework ships with an extensive Command Line Interface (CLI) to generate (or destroy) code, start the server, open a Ruby (or a database) console, create/drop the database, run the migrations, precompile assets, and print informations.

Code Generators

Code generators is a helpful feature that allows to generate code files. The main purpose is to speedup the development process. Hanami has code generators for projects, apps, actions, models, and migrations. Here’s an example:

➜ hanami generate model book
      create  lib/bookshelf/entities/book.rb
      create  lib/bookshelf/repositories/book_repository.rb
      create  db/migrations/20170131083319_create_books.rb
      create  spec/bookshelf/entities/book_spec.rb
      create  spec/bookshelf/repositories/book_repository_spec.rb

This generator creates an entity, a repository, a migration and the related testing code.

How To Test Code Generators

There are a few things to verify that a code generator is working as expected. The first one is the exit code of the command, then we can verify the output of the command, and finally inspect the files.

it 'generates model' do
  input      = "book"
  model      = "book"
  class_name = "Book"
  table_name = "books"
  project    = "bookshelf"

  with_project(project) do
    output = [
      "create  lib/#{project}/entities/#{model}.rb",
      "create  lib/#{project}/repositories/#{model}_repository.rb",
      /create  db\/migrations\/(\d+)_create_#{table_name}.rb/,
      "create  spec/#{project}/entities/#{model}_spec.rb",
      "create  spec/#{project}/repositories/#{model}_repository_spec.rb"
    ]

    run_command "hanami generate model #{input}", output

    expect("lib/#{project}/entities/#{model}.rb").to have_file_content <<-END
class #{class_name} < Hanami::Entity
end
END

    # ...
  end
end

This is a simplified version of the testing code.

I use the with_project helper, to setup a new Hanami project at the runtime. Within its block the commands are run at the root of this fixture project.

The output variable sets the expected lines of output of the hanami generate command. Each element in the array is a matcher for a line of output. If you look at the example output above, you can spot the similarities.

When the test invokes run_command, it uses a real shell command to actually generate these files. I pass the command as first argument and the expected output as second one. The helper asserts that the exit code of the command is successful.

At this point I want to verify that the files were generated correctly. Aruba, the CLI testing framework used here, offers convenient RSpec matchers like be_an_existing_file.

Using the have_file_content matcher, I can assert that the generated file has the expected contents.

Conclusion

To test Hanami CLI is complex: it requires a reliable CI build (with full isolation of each spec), and a deep knowledge of the Ruby toolchain. Next time we’ll talk about testing the Hanami console.