urllib: urlretrieve() on Windows, given a local file url and no target filename method returns an invalid filename

Bug report

Bug description:

On windows, if the request.urlretrieve method receives an url pointing to a local file and no filename is provided, then the returned filename will be invalid:

from urllib import request

request.urlretrieve("file:///C:/testdir/testfile.txt")
# ('\\\\\\C:\\testdir\\testfile.txt', <...>)
# Expected 'C:\\testdir\\testfile.txt'

# Other variants of urlretrieve calls work as expected
request.urlretrieve("file:///C:/testdir/testfile.txt", "C:\\testdir\\testcopy.txt")
# ('C:\\testdir\\testcopy.txt', <...>)

request.urlretrieve("https://google.com/")
# ('C:\\Users\\...\\Temp\\tmpwp4pu1bw', <...>)

I would expect that if no copy is necessary then the filename returned by request.urlretrieve would be equal to request.url2pathname(parse.urlparse("file:///C:/testdir/testfile.txt").path)

CPython versions tested on:

3.14

Operating systems tested on:

Windows

Linked PRs

• gh-156550

Start with urllib.request.urlretrieve and the urllib.request.url2pathname(parse.urlparse(...).path) behavior described in the report. Verify the no-filename local-file case on Windows while preserving the working explicit-filename and remote-URL variants; done means the returned filename matches the expected Windows path.