Matrix inverse fails for Cholesky factorizations of diagonal matrices

[wzhorton]

The strides function throws a method error when called from inv on Cholesky factorizations of diagonal matrices.

using LinearAlgebra
A = Diagonal([1,2,3])
cholA = cholesky(A)
inv(cholA) # 'no method matching strides(::Diagonal{Float64, Vector{Float64}})'

Issue 44053 might be related, but deals with construction, rather than a specific routine.

Although taking the Cholesky of a diagonal matrix is a trivial process, it makes sense to support it in the context of inverting a covariance matrix within a function, e.g. sympd_invert(Sigma) = inv(cholesky(Sigma)). Receiving a diagonal covariance Sigma is reasonable in several situations.

---

Version Info:
Julia Version 1.7.3
Commit 742b9abb4d (2022-05-06 12:58 UTC)
Platform Info:
OS: macOS (x86_64-apple-darwin21.4.0)
CPU: Intel(R) Core(TM) i5-7360U CPU @ 2.30GHz
WORD_SIZE: 64
LIBM: libopenlibm
LLVM: libLLVM-12.0.1 (ORCJIT, skylake)

[dkarrasch]

This is fixed on v1.8:

julia> inv(cholA)
3×3 Matrix{Float64}:
 1.0  0.0  0.0
 0.0  0.5  0.0
 0.0  0.0  0.333333

The question is whether we would like to specialize that to give a Diagonal result.