Cleaning up the code, I made the change_set tracking code less efficient than is possible, and worse than it was.
Perhaps we can add a more efficient algorithm without making a gigantic mess this time.
Though there should probably be a reason to do so, which would probably need someone running it over a large number of files.
Otherwise really, parsing should dominate i imagine. But if it is an issue with it, it can currently be improved.
Cleaning up the code, I made the change_set tracking code less efficient than is possible, and worse than it was. Perhaps we can add a more efficient algorithm without making a gigantic mess this time.
Though there should probably be a reason to do so, which would probably need someone running it over a large number of files. Otherwise really, parsing should dominate i imagine. But if it is an issue with it, it can currently be improved.