The results from running 'fmusim' on ModelExchange CoupledClutches FMUs (FMI2 and FMI3) do NOT match the (expected) results from FMPy and the old FMUChecker. The issue is reproducible with both MWorks and MapleSim FMUs. Based on logging information, the main difference seems to be that the 'fmusim' tool calls
fmi2SetContinuousStates(x={0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0}, nx=18) -> OK
upon entering ContinuousTimeMode for the first time but not all of the states have start=0.
Should it not call fmi2GetContinuousStates first?