As part of removing "flavour" classes in #31691, I changed pathlib's glob() implementation: previously it used re.IGNORECASE to implement case-insensitive matches, whereas after it called os.path.normcase() on the pattern and the paths. The new behaviour is a little slower, and I think we should restore the previous implementation.

Linked PRs

• GH-104104: Optimize pathlib.Path.glob() by avoiding repeated calls to os.path.normcase() #104105