[Bug c/122572] [15/16 regression] Segmentation fault on aarch64 building with -O2, no issue with -O2 -fschedule-insns

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=122572

--- Comment #12 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
(In reply to Sam James from comment #9)
> Can you get a testcase with __builtin_types_compatible_p on x86?

No because the types are correctly compatible, it is the internal aliasing that
goes wrong.