[Bug tree-optimization/122197] [16/17 Regression] predictively devirtualization vs middle-end warnings since r16-4000

[pobrn at protonmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=122197

--- Comment #21 from Barnabás Pőcze <pobrn at protonmail dot com> ---
(In reply to Barnabás Pőcze from comment #19)
> [...]
> Would it not be sufficient to copy/move the `-Warray-bounds` suppression
> from `~_Sp_counted_ptr_inplace()` into `_M_dispose()` of the same type?
> Compiles for me after that change.

Any comments about this? Would this be an acceptable workaround?