lq for AbstractFloat

[RJDennis]

Most of the matrix decompositions in Julia can be applied to matrices with element types that are subtypes of AbstractFloat. For example

A = BigFloat.(rand(3,3))

qr(A)
lu(A)
svd(A)
schur(A)

all work fine. However,

lq(A)

gives a method error because lq() seems to require eltype(A) to be either Float16, Float32, or Float64.

Can the lq() function be generalized so that it can be applied to matrices whose elements are a subtype of AbstractFloat?

[stevengj]

Can't we just use qr(A')?

[RJDennis]

Yes we can. I had thought that there might be a small change to the existing lq() function that would allow it to accept other subtypes of AbstractFloat.