simulation hits max iters when using `TimeChoiceIntegrator`

[kylebeggs]

Sometime I experience an issue where the end of a simulation hangs for a while and then produces a warning that it hits maximum iterations. If I simply run the TimeChoiceIterator to one less timestep than the tspan, this warning never happens:

for (u_, t) in OS.TimeChoiceIterator(integrator, tspan[1]:dtvis:(tspan[2] - dt₀))
    ...
end

I have not investigated this further at this time but just wanted to report it at this time in case somebody knows the solution immediately.

[termi-official]

Yea, the integrators are a duck-tape solution which I do not want to spent too much time on. You can hotfix this by setting the maxiters for each subintegrator manually (as you would do with the normal DiffEq integrator).

[kylebeggs]

but why doesn't the subintegrator hit the max iters for any other step? i think the adaptive subintegrator may be running into some kind of issue with the end here. Either way, given the interface should be fixed up in the new package, I agree I'm just hotfixing this for now.

[termi-official]

This is essentially SciML/OrdinaryDiffEq.jl#2668