Broken links are iterated over, but raise error when stat()'ed
[x for x in Path() if x.is_dir()]
will fail if the working directory features some broken link (the link will be iterated over, but calling is_dir() method will cause a FileNotFoundError)
This is not strictly a bug, but it disallows a clean way to iterate over the directory (like in the shown example, which was taken from the pathlib docs). Probably broken links should not be iterated over - or they should not respond with error on is_dir() and similar methods.
(sorry for several editions of issue - I hitted something during editing, releasing it.)