SimulationStatus semantics

[molpopgen]

The docstring for this type doesn't do a great job of explaining what the expectations are.

Further, existing tests are invariant to how this type is used in for loop checking existing fixations in the implementation of GlobalFixation. So either the tests aren't covering that block or something is up.

[molpopgen]

Removing this from the 0.18.0 milestone -- I need to think of what needs testing and how to test it.

[molpopgen]

Further, existing tests are invariant to how this type is used in for loop checking existing fixations in the implementation of GlobalFixation. So either the tests aren't covering that block or something is up.

Here, there is some logic that checks if a mutation's key has changed and the mutation is now in the fixations vector. It is either not possible to hit this block or our tests don't cover this case.

[molpopgen]

Here, there is some logic that checks if a mutation's key has changed and the mutation is now in the fixations vector. It is either not possible to hit this block or our tests don't cover this case.

Some exploration leading to #1160 indicates that this code segment is not getting hit at all. I should probably determine coverage for this module.

Edit: the logic is incorrect in the code not getting executed: the mutation index will always/most often yield the same key value because the mutation hasn't had a chance to be recycled yet.

[molpopgen]

One potential way forward:

1. Change the status class to an enum with variants Restart, Continue, Complete.
2. Handle the "simulate to end of model" concept after the while loop handling the conditional part. With demes models, it is now trivial for us to pick up where we left off, etc..