Discrepancy between Float32 and Float64

[arlk]

MWE:

julia> f(x) = sin(x)^2
f (generic function with 1 method)

julia> f(-1..1)
[0, 0.708074]

julia> f(-1f0..1f0)
[-0.708074f0, 0.708074f0]

Is this simply a duplicate of #415?

[lbenet]

Thanks for reporting!! The problem, I think, is related to pow function for Interval{Float32}, which is using pow in Base.

julia> (-1 .. 1)^2
[0, 1]

julia> (-1f0 .. 1f0)^2
[-1f0, 1f0]

julia> @which (-1f0 .. 1f0)^2
^(x::Number, p::Integer) in Base at intfuncs.jl:290