Related to my previous question, I now run into an issue with the step! interface.
Sorry that I didn’t manage to create a MWE, if needed I will try harder 
At the moment, I am in a situation where for exactly the same SDE/ODE problem solve(prob) works, but step!(prob, tspan[end]) does not work, due to dt becoming too small in the very first time-step. (Notably, the solution produced by solve uses time steps that are larger than admin.)
Is there a well known trick to make the integrator interface treat a step like an initial step?
(Anyway, I get that the question is ill-posed. Let me know if I need to add more details.)
An MWE might be needed.