Re^2: RFC: How did I do writing my first test?

[jcb]

Testing the final newline presence/absence is still beneficial to ensure that the code stays agnostic to it. Further, if the code handles those cases differently, testing that processing is needed for code coverage.

[Arunbear]

I'm not sure ... (on closer inspection of what it does) that seems like basically retesting Perl's readline function, which is already tested elsewhere.

[jcb]

I would offer that such testing is still reasonable, to ensure that the behavior does not change if/when some future version implements more complex wrapping around readline.

If the high-level specification says that the code behaves "like so" with/without a final newline, the testsuite should verify that, even if it is effectively verifying part of perl itself in the current implementation. If nothing else, this will blow up if some future perl changes that behavior, alerting to the need to either add code to maintain the previous behavior or revise the specification.