Ambiguity errors from new matrix multiplication methods in LinearAlgebra

[andreasnoack]

The methods in

LeastSquaresOptim.jl/src/utils/lsmr.jl

Lines 16 to 21 in c0047cb

	function mul!(C::Vector{Float64}, A::Matrix{Float64}, B::Vector{Float64}, α::Number, β::Number)
	gemm!('N', 'N', convert(Float64, α), A, B, convert(Float64, β), C)
	end
	function mul!(C::Vector{Float64}, A::Adjoint{Float64, Matrix{Float64}}, B::Vector{Float64}, α::Number, β::Number)
	gemm!('C', 'N', convert(Float64, α), A', B, convert(Float64, β), C)
	end

need to be behind a conditions similarly to https://github.com/JuliaLinearAlgebra/GenericLinearAlgebra.jl/blob/1ac7a0f83121849efd76715ee85a115047d6468d/src/juliaBLAS.jl#L87 to avoid breakage in the upcoming Julia version 1.3.

[matthieugomez]

Solved & tagged a new version. Thanks!