"Unfriendly" behavior in notebooks when bugs occur.

The firsts symptoms of the error in #1966 were a Jupyter notebook kernel repeatedly dying. This behavior is due to the use of tsk_bug_assert , which calls abort, preventing any communication back to the user. At the very least, we should document that mystery kernel deaths mean that you should copy your code to a script and run it in the terminal.

I'm not sure there's much we can do here @molpopgen - tsk_bug_assert is only called in situations that we think are bugs, so should happen very very rarely (#1966 is a symptom of the relative immaturity of the sweeps code). We should never be calling tsk_bug_assert in a normal error-handling way.

I guess we could add a note to the "developers" page saying that this is a caveat of doing stuff in a Jupyter notebook, and then hope that someone googline "msprime Jupyter kernel dies" or something will hit that text?