https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106654
Jason Merrill <jason at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Ever confirmed|0 |1
Last reconfirmed| |2022-09-17
Status|UNCONFIRMED |NEW
--- Comment #8 from Jason Merrill <jason at gcc dot gnu.org> ---
(In reply to Aldy Hernandez from comment #7)
> Silly question, why can't you expand the [[assume]] construct into:
>
> if (x > 5)
> __builtin_unreachable ();
>
> ...like we always have. Then no changes are needed to ranger :). Or does
> this have to do with the whole side-effect thing?
Exactly. For expressions with no side-effects, we can do that. For, say, a
call to a non-const function, we need to avoid actually emitting the call.
