Check the correct logic of setting `ts = FALSE` in a `slim()` runs

[bodkan]

I switched to tree sequence as a storage of simulation state for users who wish to re-start simulations (for situations like "an allele of interest is lost", etc.).

But there's a weird interaction of this with the ts = argument of slim(). I added this for situations in which users would like to switch off tree-sequence recording in "SLiM customization snippets" of slendr models... but these two things obviously don't interact well.

Easy fix, just have to look at this more carefully.

("Premature optimization" strikes again.)