[Bug tree-optimization/126703] copysign(x, -abs(t)) -> copysign(x, -1.0)

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=126703

--- Comment #3 from Drea Pinski <pinskia at gcc dot gnu.org> ---
(In reply to Richard Biener from comment #2)
> I wonder if we can have value_range::sign_bit () and thus can fold
> copysign (x, t) with known signbit of 't' to copysign (x, -1) or copysign
> (x, 1)
> and for same sign_bit of x and t elide it?

Yes that definitely would work.