And yes, GCC extensions are often going to be adopted in clang, but generally not the broader world of C and C++ compilers. Everything in my article conforms to the standard.

▲

fuhsnn 5 hours ago | parent [-]

I played with a lot of preprocessor implementations and did my own (redesigned chibicc's expansion algorithm), not many of them even have paint-blue behavior exactly right (the standard text is vague, to me it was more "matching GCC's" than "conforming to standard").

	▲	viega 5 hours ago \| parent [-]
		That's interesting. I agree with you that the standards text is pretty vague. I think that's why other attempts to show how to do this kind of thing don't get deep enough on the semantics, and why I adopted a "try it and see" strategy. I do try to avoid this kind of thing unless necessary, so I don't have experience as to where the different compilers will fall down on different corner cases. I'd find it very interesting though, so please do share if you kept any record or have any memory!