Yeah, I'm of the opinion that compilers that treat an empty identifier list in a function definition placed before another function definition as meaning "no information about the number or types of the parameters is supplied" when the definition is taken as serving as a forward declaration are likely non-conforming, but in theory this is academic because identifier lists are obsolescent, which leads me to my rant:
They didn't remove identifier lists in C17 even though there probably hasn't been new code written with non-empty identifier lists in over two decades, and my guess is that's because there's plenty of new code that uses empty identifer lists in function definitions or mistakenly in forward declarations to mean "no parameters". At some point they should admit that this is a mess that could be cleaned up by simply removing identifier lists and re-introducing empty identifier lists as a special case having semantics identical to void parameter type lists.
It'll upset those who have continued to use empty identifier lists in forward declarations to mean "no information about the number or types of the parameters is supplied", but that hasn't been good practice for awhile, and you get no sympathy for having the rug pulled from under you when you insist on using a feature that has been obsolescent for over two decades.
Originally Posted by Bjarne Stroustrup (2000-10-14)
