The documentation of Unicode::UCD says the functions return undef for errors. Actually, mostly it does this for scalar context, and an empty list for list context, as recommended by Damian. Aaron disagrees with that recommendation http://aaroncrane.co.uk/talks/calamitous_context/paper.html Basically Aaron's argument comes down to, correct me if I misstate this, that you can't get a return that works right for all circumstances, and Damian's solution favors a rarer case at the expense of a more common case. My position is that the goal should be that Unicode::UCD functions should all behave consistently with each other in this regard. If the documentation and behavior matched, we would just leave it at whatever that was, saying the churn wasn't worth it. But since the documentation and behavior disagree, what should we do? Short term for 5.22, and long term, if different?Thread Next