Benoît Legat

Yes, the overload would have been easier if LinearAlgebra was implementing their default dispatch methods like we do in MOI and MA with `foo_fallback` so that`foo` does not have too...

IIRC, I tried debugging it and it didn't seem to be a bug, the problem is just very ill-conditioned but I might be confusing with another issue

I agree it could be moved to MOI at some point. It could allow adding a `dualize` keyword argument to JuMP (even if we could also add Dualization as a...

I haven't checked yet the new version of Protobuf but I can take a look

@ulfworsoe It is now giving `Mosek.MosekError(1050, "")`, is that expected ?

The type of the returned variable may depend on the type of the domain. For instance with SumOfSquares, it is `MatPolynomial{JuMP.VariableRef}` for a domain `FullSpace` while with a domain is...

I should just be rather straightforward, the status is https://github.com/MOSEK/Mosek.jl/blob/ea17cd2ac2f5dae43e2f9a29948ded26c3a57b7b/src/msk_enums.jl#L12375-L12390

@oxinabox this is the issue I mentioned to you at JuliaCon

Oops, indeed this `ordering` and `AbstractOrdering` are not in sync with the rest, we should make it match

We actually have these implemented in DiffOpt: https://github.com/jump-dev/DiffOpt.jl/blob/master/src/utils.jl It's used to communicate the backward differentiation results