Don't assume transpose makes a copy

[lkdvos]

Here I address a range of subtle bugs that were already there and some new ones that got introduced when the TensorKit update happened.
The primary problem is that we shouldn't assume that transpose or any of its variants produces independent data. Obvious offenders were e.g. qr_compact!(repartition(t, 2, 2)), which isn't guaranteed to not alter t.

The problem wasn't caught before because in most cases, a copy has to be made anyways (e.g. when there are symmetries involved, or for most array sizes), but after a bug report by @20akshay00 the issue shows itself when working with product states, since there 1xdx1 arrays can always trivially be permuted without copies.

I hope that I caught all of the cases here, but it's a bit hard to test this in general since some of the algorithms seem to even be somewhat robust against these kinds of issues.

[codecov]

Codecov Report

❌ Patch coverage is 82.35294% with 3 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
src/utility/utility.jl	75.00%	2 Missing ⚠️
src/states/orthoview.jl	50.00%	1 Missing ⚠️

Files with missing lines	Coverage Δ
src/algorithms/changebonds/optimalexpand.jl	100.00% <100.00%> (ø)
src/algorithms/changebonds/randexpand.jl	100.00% <100.00%> (ø)
src/algorithms/groundstate/idmrg.jl	98.55% <100.00%> (ø)
src/states/quasiparticle_state.jl	86.25% <100.00%> (ø)
src/states/orthoview.jl	89.68% <50.00%> (ø)
src/utility/utility.jl	74.22% <75.00%> (ø)

... and 2 files with indirect coverage changes

🚀 New features to boost your workflow:

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