The float->int conversion fails for me with as expected with msvc and

current gcc versions. Of course the error message should be a lot better.

When you say "fail" do you mean at compile time or runtime?

Rob