Test case problem when exception handled

Question

I have the following code with the corresponding test case:

class XXX
    attr_accessor :source
    def check
        begin
            raise ArgumentError, "No source specified." \
                unless @source.empty? != true
            puts "Passed"
        rescue
            print "Error: ",$!, "\n"
        end
    end
end

class TestXXX < Test::Unit::TestCase
    def setup
        @x = XXX.new
    end

    def test_check
        assert_nil(@x.source)
        assert_raise(NoMethodError) { @x.check }
    end
end

Running the test case will produce the following result on assert_raise due to the presence of the 'rescue' routine:

Started
.Error: undefined method `empty?' for nil:NilClass
F
Finished in 0.01 seconds.

  1) Failure:
test_check:18 exception expected but none was thrown.

1 tests, 2 assertions, 1 failures, 0 errors

How do I write a test case for this scenario when I cannot do away with removing the 'rescue' from my code?

Commenting 'rescue' out produces a favorable result for me.

class XXX
    attr_accessor :source
    def check
        begin
            raise ArgumentError, "No source specified." \
                unless @source.empty? != true
            puts "Passed"
        #rescue
        #    print "Error: ",$!, "\n"
        end
    end
end

Started
..
Finished in 0.0 seconds.

1 tests, 2 assertions, 0 failures, 0 errors

Answer 1

Test for the actual behaviour of the method. It doesn't throw an exception it prints, so that's what you need to check for. You probably need a mock object?