Indeed, strange. There are also other flags for floating point exceptions, but I have never heard about anyone actually using them. I'd like to have a look at language that supports these flags - do you know any?