Make EulerEq2d run with Kelvin Helmholtz test

• fixed dg weak form computation in 2d for special case when one component of the flux is zero
• fixed/implemented flux and ldg flux equations for 2d
• converted all max calls to absmax calls in equations.jl
• extracted numerical utilities from utils.jl into numutils.jl

The IntegrationTests/refs/euler2d_kelvin_helmholtz_entropyav.toml test now runs with polynomial orders n=1 (up to n=4; but not fully tested). Unfortunately, atm this runs pretty slow, so it is not yet added as ref test, but goal is to make it one in the future.