This PR is a WIP to see if the ideas in #163 are helpful.
There is a logic error in the issue as described. The number of indexes of mutations that are "simplified out" is not the same as the length of a recycling queue. The correct procedure is something like:
generate a list of mutation indexes that are preserved after simplification
generate a list of all possible mutation indexes
The recycling queue is the set difference of these two lists.
This PR is a WIP to see if the ideas in #163 are helpful.
There is a logic error in the issue as described. The number of indexes of mutations that are "simplified out" is not the same as the length of a recycling queue. The correct procedure is something like: