Mateusz Baran

> * group 1: the signature `solver(M, f, p0::Number)` to write a wrapper for `f` creates an ambiguity with `solver(M, objective, p0)` – where we also do not want to...

> Ad 1) The function is used when updating the Hessian at I see, thanks. > Ad 3) Sure, I am not 100% sure how to realise that, but when...

Nice, so that PR would completely resolve this :+1:

This sounds like a good plan :slightly_smiling_face: .

Going back to the original issue, one small problem with these `cat` methods is that constant propagation of keyword arguments is needed to make this type stable and it's a...

> Hi @mateuszbaran, thank you for your kind response. This means that in Julia 1.6 you are going to choose returning a `SArray` (and being type stable)? I think it...

Hm, I will try to fix this soon. Apart from SciML sometimes doing breaking changes in non-breaking releases, ManifoldDiffEq.jl hooks into some of their internals so they can justifiably feel...

> What breaking change? Please help me understand what you are asking for here and I can help out. Thank you for offering help but everything except this issue has...

> I would highly recommend sticking to public APIs if you want stability. There are no guarantees to the outside that internal functions will not change any arguments or behavior....

The issue was fixed by #33 .