check whether checking mutation parents on ts init breaks something

[petrelharp]

Over in tskit-dev/tskit#3212 we are considering adding a check for mutation parent correctness to the tree sequence init procedure. I have a vague recollection that this might mess up something in our workflow, but looking at the code I don't see what that might be. We should verify.

[bhaller]

OK. I'll assign this one to you, but will be happy to help if you tell me what I should do. :->

[benjeffery]

The possible breakage is for those working with tree sequences where mutation parents are either all tskit.NULL or wrong in someway (don't agree with topology). Currently tskit allows these tree sequences to be used, the new PR enforces that the parents are correct.