Provide floattype methods for non-Real types

Author: timholy

src/FixedPointNumbers.jl

@@ -156,6 +156,12 @@ floattype(::Type{X}) where {T <: Integer, X <: FixedPoint{T}} = Float64
 floattype(::Type{X}) where {T <: Integer, X <: Rational{T}} = typeof(zero(T)/oneunit(T))
 floattype(::Type{X}) where {T <: LongInts, X <: FixedPoint{T}} = BigFloat
 
+# Non-Real types
+floattype(::Type{Complex{T}}) where T = Complex{floattype(T)}
+floattype(::Type{<:Irrational}) = Float64
+floattype(::Type{Base.TwicePrecision{Float64}}) = Float64    # wider would be nice, but hardware support is paramount
+floattype(::Type{Base.TwicePrecision{T}}) where T<:Union{Float16,Float32} = widen(T)
+
 float(x::FixedPoint) = convert(floattype(x), x)
 
 function minmax(x::X, y::X) where {X <: FixedPoint}

test/traits.jl

@@ -21,6 +21,12 @@ struct MyReal <: Real end
         @test typemax(T) <= maxintfloat(floattype(T))
     end
     @test floattype(Rational{Int}) === Float64
+    @test floattype(Complex{Int16})   === Complex{Float32}
+    @test floattype(Complex{Float32}) === Complex{Float32}
+    @test floattype(Base.TwicePrecision{Float16}) === Float32
+    @test floattype(Base.TwicePrecision{Float32}) === Float64
+    @test floattype(Base.TwicePrecision{Float64}) === Float64
+    @test floattype(typeof(π))                    === Float64
 
     @test_skip(@test_throws MethodError floattype(MyReal))   # TODO: eliminate `@test_skipped` when depwarn is eliminated. See #177.
 end