Gitective counts commits that are in multiple branches multiple times

[caiusb]

As I was working for #29, this problem appeared in one of the tests. Theoretically, this should not affect the results at all (hopefully the results are replicable 😟 ).

But because the AST difference is very time consuming, this might increase the time it takes to run the analysis to unacceptable levels (it's almost there anyway, so we can't really afford that).

There are 2 solutions to this:

1. Keeping track of all the commits we see and write a filter to exclude them if it sees them once more (easy).
2. Playing with the walk algorithm so that it works as we expect it (hopefully other things will not break as a consequence) (not so easy).

I will try implementing the 1st option, since it's the fastest and should work just fine for our case. But if anybody has some ideas I am more than happy to listen to them.