https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83584
--- Comment #17 from joseph at codesourcery dot com <joseph at codesourcery dot com> --- There are certainly cases where something that is undefined at compile time in ISO C (i.e. the undefinedness is a property of the program, not of a particular execution path through the program) are unconditional hard errors - when GCC does not have any support for defining any semantics for the undefined construct. Since it's OK for such things to be unconditional hard errors, it's also OK for them to be unconditional pedwarns (subject to the general principle of not really wanting a warning without an option to turn it off - but at least, on-by-default pedwarns are OK). And it's also OK for them to be pedwarns-if-pedantic. But the wording should be improved (e.g. "does not permit" or "does not support" instead of "forbids"), and there may be a case for a separate option that could turn this diagnostic off when used with -pedantic / -pedantic-errors.
