[BUG] Fixed step may cause integrator to fail silently

[jon-wurtz]

Describe the bug
When the time step of an integration is fixed, eg by choosing some small number of time steps for some total evolution time, the integration appears to fail by returning the wrong wavefunction at long times. Crucially, it does so silently without throwing any errors. A fix would be to check the integrator status (eg status(integrator)==True) or to adaptively choose a smaller step if needed.

To Reproduce
In https://github.com/Happy-Diode/Bloqade.jl/blob/master/examples/ring/main.jl reduce nsteps by factors of two until integration obviously fails

[Roger-luo]

I don't think we are able to handle this case, the solver actually returns as :Successful so no error is thrown, it might be something that should be dealt with upstream. I think for now we should just warn user to check convergence on different step size.

On other wrong step size the solver always throws warning