I was reading makeunicodedata.py and found a small issue that leads to a bit of wasted space in the unicodedata module. This code:

                try:
                    i = decomp_data.index(decomp)
                except ValueError:
                    i = len(decomp_data)
                    decomp_data.extend(decomp)
                    decomp_size = decomp_size + len(decomp) * 2

will always take the exception path because both decomp_data and decomp are lists of ints, so the index call always raises. I think it was copy-pasted from some similar code a few lines up which however is about a lists of strings and a string, so it works there.

I will prepare a PR. The space saving isn't gigantic (about 4.5kib) but as it stands the code is confusing.