Add tests/fixes for inplace rules

[kshyatt]

Added more tests to test_pullbacks_match to make sure the state of the arguments is restored, and the final argument derivatives match between inplace and non in place methods.

Unfortunately, the Mooncake FD tester doesn't work well for our functions, because A becomes a scratch space, and the inputs are also the outputs (so get incremented twice under the FD scheme).

[kshyatt]

I'd like to do a bit of tidying up with this one but it can be merged if tests pass and people feel like it, I can just make a separate PR.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

Files with missing lines	Coverage Δ
...gebraKitMooncakeExt/MatrixAlgebraKitMooncakeExt.jl	96.78% <100.00%> (+1.56%)	⬆️

🚀 New features to boost your workflow:

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

[lkdvos]

Overall looks good to me!

I have two remaining questions:

1. This also bumps the Mooncake version, should this come before or after #164 ?
2. I might be getting this wrong, but I am kind of confused now about why we are zero!ing the dargs. Aren't we supposed to accumulate the contributions of the gradient into that, not setting them to? I.e., should this not just be darg1 += zero(darg1), or effectively we don't have to do that?

[EDIT] ignore this last part, I'm stupid.