Thank you for the review, @merwok!

Pardon the bump, but I figured this PR might be a good place to ask. @barneygale is there a specific reason why e.g. absolute(), is_mount(), etc use os.path with this patch instead of self.pathmod (or self._pathmod with the revert)? Especially since before this change they did use self._flavour. Using self._pathmod there would make it easier for custom subclasses to provide a different implementation.

os.path is the only valid value for pathmod in subclasses of Path, because Path is specifically concerned with the local filesystem. Many Path methods call os functions which can't be replaced by swapping out pathmod, and so it seemed silly to support customizing a small handful of other methods via pathmod. Instead, I think you should override Path methods as you need, and leave pathmod alone. Does that make sense?

Alternatively, if your subclass is not concerned with the local filesystem, you could subclass from pathlib_abc.PathBase, which doesn't use os.path: https://pypi.org/project/pathlib-abc/

Thanks for the explanation, that makes sense. Curiosity satisfied :) I must admit that our use-case and restrictions are quite unique, overriding the methods as we need is indeed what we do now.

Thanks for linking pathlib_abc, I didn't know about that and definitely looks useful to look into for the future!