`QRFactorization` throws when given singular matrices

[oscardssmith]

this gives a LAPACKException because A is singular and QRFactorization defaults to NoPivot instead of ColumnNorm.

using LinearSolve
A = vcat(rand(1,2), zeros(2,2)), rand(3))
solve(LinearProblem(A, QRFactorization())

I'm not sure if there's a good way to detect whether the factorization is singular in this case, but we should do so and return a failure rather than throw an error.

[avik-pal]

Fixed in the latest release