[Bug c++/120502] [15/16 Regression] ICE in verify_ctor_sanity with -std=c++20 -O1 and no_unique_address since r15-9189

[ppalka at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=120502

Patrick Palka <ppalka at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|needs-bisection             |
            Summary|[15/16 Regression] ICE in   |[15/16 Regression] ICE in
                   |verify_ctor_sanity with     |verify_ctor_sanity with
                   |-std=c++20 -O1 and          |-std=c++20 -O1 and
                   |no_unique_address           |no_unique_address since
                   |                            |r15-9189
           See Also|                            |https://gcc.gnu.org/bugzill
                   |                            |a/show_bug.cgi?id=119387
                 CC|                            |ppalka at gcc dot gnu.org

--- Comment #2 from Patrick Palka <ppalka at gcc dot gnu.org> ---
Started with r15-9189 "c++: P2280R4 and speculative constexpr folding" which
was backported as r14-11528, though I can't reproduce the ICE on the 14 branch
likely because it also depends on the constexpr prvalue optimization r15-6052.