Implement schur_full/schur_vals for Diagonal inputs

[lkdvos]

Fixes #276: Diagonal input selects DiagonalAlgorithm, which schur never implemented, so schur_full/schur_vals threw a MethodError.

A diagonal matrix is already in Schur form, so T = A, Z = I and vals = diagview(A). Two choices worth flagging:

• No reordering, unlike the Diagonal path of eig, which sorts. So schur_vals(::Diagonal) != eig_vals(::Diagonal) in ordering; the testsuite's valsc ≈ eig_vals(A) is relaxed accordingly.
• vals keeps eltype(A) (no complexification), and Z stays Diagonal, as qr/lq keep Q and R diagonal.

Enables the Diagonal schur tests that were commented out as "not supported yet", CPU and CUDA. Note this alone does not fix schur_full(::DiagonalTensorMap): TensorKit omits schur from its default_algorithm forwarding, so selection fails there for every tensor, diagonal or not. That needs a separate PR.

🤖 Generated with Claude Code

[codecov]

Codecov Report

❌ Patch coverage is 92.30769% with 3 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
src/implementations/schur.jl	92.30%	3 Missing ⚠️

Files with missing lines	Coverage Δ
src/implementations/schur.jl	88.65% <92.30%> (+2.45%)	⬆️

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[Jutho]

What is the rational for schur_vals of a real diagonal not changing to complex? eig_vals does do this, and schur_vals on the equivalent full matrix would also return complex results.

[lkdvos]

No rationale... I forgot we had this whole discussion with the eigenvalues as well, I'll change it