Sorry Chris, I apologise but I couldn't resist it. That comparatively short thread and in particular your later comment is indeed required reading for commenting on this review. I personally am still in favour of this pragmatic, incremental change as there is a use case for conditional test inclusion and I think a more full blown evaluation of moving Swift's preprocessor into the lexer is unlikely for the foreseeable (even though I'd fully support that and would even look at it in terms of the compiler if I thought there was backing - proceeding with this proposal would not preclude looking at the lexer later).