Please diagnose on _Noreturn calls within [[gnu::const]], [[gnu::pure]], [[reproducible]], [[unsequenced]]

[alejandro-colomar]

Please diagnose on _Noreturn calls within the body of a function with any of these attributes: [[gnu::const]], [[gnu::pure]], [[reproducible]], [[unsequenced]].

This is UB, and could be easily diagnosed. I'd put it in -Wextra for the moment, to be careful.

Cc: @AaronBallman

[llvmbot]

@llvm/issue-subscribers-clang-frontend

Author: Alejandro Colomar (alejandro-colomar)

Please diagnose on _Noreturn calls within the body of a function with any of these attributes: [[gnu::const]], [[gnu::pure]], [[reproducible]], [[unsequenced]].

This is UB, and could be easily diagnosed. I'd put it in -Wextra for the moment, to be careful.

Cc: @AaronBallman

[alejandro-colomar]

GCC bug: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=119041

[efriedma-quic]

The C standard says the following is well-defined, as far as I can tell:

_Noreturn void f() { while (true); }
void g() [[unsequenced]] { f(); }
int main() { g(); }

We could assume nobody is going to write code like this in practice, and warn anyway, I guess.

[alejandro-colomar]

The C standard says the following is well-defined, as far as I can tell:

_Noreturn void f() { while (true); }
void g() [[unsequenced]] { f(); }
int main() { g(); }

The C Committee discussed yesterday that this is a defect in the standard, and that we should change the standard.

Which is what triggered me to propose this.

We could assume nobody is going to write code like this in practice, and warn anyway, I guess.