On Mar 4, 2008, at 2:58 PM, Eli Friedman wrote: > Does always_inline imply that it's illegal to take the address of the > function? The gcc docs don't seem to say that. I'd argue that it should. I don't think people have double checked that with code in the wild yet.