MvNormal aliases should have looser type constraints

[nickrobinson251]

Is there a reason for these to be Float64 rather than Real?

Distributions.jl/src/multivariate/mvnormal.jl

Lines 180 to 186 in 432545a

	const IsoNormal = MvNormal{Float64,ScalMat{Float64},Vector{Float64}}
	const DiagNormal = MvNormal{Float64,PDiagMat{Float64,Vector{Float64}},Vector{Float64}}
	const FullNormal = MvNormal{Float64,PDMat{Float64,Matrix{Float64}},Vector{Float64}}
	const ZeroMeanIsoNormal = MvNormal{Float64,ScalMat{Float64},ZeroVector{Float64}}
	const ZeroMeanDiagNormal = MvNormal{Float64,PDiagMat{Float64,Vector{Float64}},ZeroVector{Float64}}
	const ZeroMeanFullNormal = MvNormal{Float64,PDMat{Float64,Matrix{Float64}},ZeroVector{Float64}}

It leads to stuff like this:

julia> z1 = MvNormal([1.0, 1.0])
ZeroMeanDiagNormal(
dim: 2
μ: [0.0, 0.0]
Σ: [1.0 0.0; 0.0 1.0]
)

julia> z2 = MvNormal([1, 1])
MvNormal{Int64,PDiagMat{Int64,Array{Int64,1}},Distributions.ZeroVector{Int64}}(
dim: 2
μ: [0, 0]
Σ: [1 0; 0 1]
)