Instead of incorrect PR#2328, I made this patch instead. This should cover all the cases of possible branching (forking first or branching in the orig repo), be faster and less memory-demanding. The algorithm tries to walk through minimum number of nodes while looking for intersection between compare branch and current branch.
Instead of incorrect PR#2328, I made this patch instead. This should cover all the cases of possible branching (forking first or branching in the orig repo), be faster and less memory-demanding. The algorithm tries to walk through minimum number of nodes while looking for intersection between compare branch and current branch.