The problem seemed to be that the orbit positions of the satellites were computed while the parent planet sill holds its ecliptic position without light speed adjustment. We could make sure that the solar system bodies are always iterated in order from parent to children, or (as I just committed), call parent->computePositionWithoutOrbits first thing in Planet::computePosition.
The problem seemed to be that the orbit positions of the satellites were computed while the parent planet sill holds its ecliptic position without light speed adjustment. We could make sure that the solar system bodies are always iterated in order from parent to children, or (as I just committed), call parent- >computePositio nWithoutOrbits first thing in Planet: :computePositio n.
Let me know if that induces any regression.