FWIW, that version ends up being slower, because you're constructing and iterating over a tuple for every iteration, which incurs a similar cost to running `.strip()` twice. The sub-generator example I gave is better because you're only constructing the generator expression once, and requires less overhead for each iteration.