https://gcc.gnu.org/bugzilla/show_bug.cgi?id=124052

--- Comment #3 from Richard Biener <rguenth at gcc dot gnu.org> ---
(In reply to Andrew Pinski from comment #2)
> TREE_CODE (type) != INTEGER_TYPE
> 
> Could also break for enum types?
> And _BitInt types?

Yes, sure.  I'd rather not enable things without at least one functional
testcase.  SCEV & friends are beasts.

Reply via email to