Degree bounds and LCs

[UlrichHaboeck75]

The current way of taking into account degree bounds in the context of LCs is wrong.
It combines the shifted polynomial/commitments according to the LC coefficients, which are not random. If we want to serve degree bound proofs when opening an LC correctly, we need to do set up the query-point dependent degree bound identities for each of the polynomials in question separately.

However, as degree bound proofs are somehow specific purpose proofs, and we won't make use of such proofs in Coboundary Marlin anyway, it might we worth to completely remove it from repository.