Updates for OrdinaryDiffEq split

[ChrisRackauckas]

A few changes will need to be made:

• Change to depending on OrdinaryDiffEqCore
• Add get_fsalfirstlastoverloads for the cache that give what the fsalfirst/fsallast pointers and remove them from the initialize! function (see Refactor ODEIntegrator to not allow undef fsal states SciML/OrdinaryDiffEq.jl#2390)

[nathanaelbosch]

#325 implements get_fsalfirstlast, but in the minimal way to just fix the current installation errors by overloading

OrdinaryDiffEq.get_fsalfirstlast(cache::EKCache, rate_prototype) = nothing, nothing

I also had OrdinaryDiffEq.isfsal(::AbstractEK) = false in my code for a while, so I'm a bit suprised to even run into this issue?

Other then that I will make sure to move over to OrdinaryDiffEqCore as soon as I find the time, and I will update this thread whenever I run into some issues. Thanks again for opening this!

[ChrisRackauckas]

I also had OrdinaryDiffEq.isfsal(::AbstractEK) = false in my code for a while, so I'm a bit suprised to even run into this issue?

We should create a generic fallback for this case.

[nathanaelbosch]

Thanks a lot for the PR #326!