URLs with query params should pass URL validator

[laurihy]

Fixes #523

Previously URL validator didn't accept URLs with query parameters if there wasn't also a path specified. This was, because in that case the old regex would include query-params to the host, and attempt to call HostnameValidation with the incorrect host.

Old regex:

regex = re.compile(r"^[a-z]+://(?P<host>[^/:]+)(?P<port>:[0-9]+)?(?P<path>\/.*)?$")
matches = regex.match('https://foo.io?no=path')
print(matches.group('host'))
# -> foo.io?no=path

New regex:

regex = (
    r"^[a-z]+://"
    r"(?P<host>[^\/\?:]+)"
    r"(?P<port>:[0-9]+)?"
    r"(?P<path>\/.*?)?"
    r"(?P<query>\?.*)?$"
)
matches = regex.match('https://foo.io?no=path')
print(matches.group('host'))
# -> foo.io
print(matches.group('query'))
# -> ?no=path

Added a bunch of tests to verify this behavior.

[whb07]

looks good in general just missing the tweaks for the typo and an extra test case for the tld test set

[laurihy]

Thanks for the good comments! Added a few more test cases, and fixed the typo in the comment.

[whb07]

Great, I saw the changes and it looks good to me. I also played with it locally and manually to verify some behavior and the fixes work 🙃