Replacing Duplicates when there are already duplicates: How are YOU doing it?

Alameth

I'm writing the logic for the Python API Library to manage duplicate filenames. An ambiguity arises when the uploader is told to replace duplicates, but the gallery already has multiple duplicate filenames from previous uploads. (For my test case, I used a gallery that had 4 images all with the same filename.)

The Web Uploader replaces the first image that has the same name. (It also increments the URI, a nice touch.) How have other implementations done this, e.g., PHP? I'm less concerned about whether the behavior of the Python library is optimal as whether it's consistent. (Replacing the last one is actually easiest to code in Python, but also strikes me as most surprising.)