URL validation fails if URL is missing user field

[ddennerline3]

The test code below fails to parse the URL when the user field is missing, but the password is present.

from marshmallow import Schema, fields, ValidationError

def test_marshmallow_url():
    """Test marshmallow validating a URL without a username"""

    # When tracing through the code, this is the regex that was built to validate the URL 
    validate_url_regex = r"^(?:[a-z0-9\.\-\+]*)://(?:[^:@]+?(:[^:@]*?)?@|)(?:(?:[A-Z0-9](?:[A-Z0-9-]{0,61}[A-Z0-9])?\.)+(?:[A-Z]{2,6}\.?|[A-Z0-9-]{2,}\.?)|localhost|\d{1,3}\.\d{1,3}\.\d{1,3}\.\d{1,3}|\[[A-F0-9]*:[A-F0-9:]+\]|(?:[A-Z0-9](?:[A-Z0-9-]{0,61}[A-Z0-9])?\.?))(?::\d+)?(?:/?|[/?]\S+)\Z"
    url_happy = "rediss://bill:[email protected]:6379"

    class UrlTest(Schema):
        my_url = fields.Url(schemes=["redis", "rediss"], require_tld=False)

    url_test = UrlTest().load({"my_url": url_happy})
    print(url_test)

    url_unhappy = "rediss://:[email protected]:6379"
    try:
        url_test = UrlTest().load({"my_url": url_unhappy})
        print(url_test)
    except ValidationError as err:
        print(url_unhappy, err)

RFC1738 specifies that either field is optional.

https://datatracker.ietf.org/doc/html/rfc1738#section-3.1

[sloria]

Thanks for reporting! I don't have time to dig into this right now, but would def appreciate an investigation and/or a PR from anyone who'd be so kind

[deckar01]

RFC1738 specifies that either field is optional.

A password is only optional after a user name is included.

Some or all of the parts "<user>:<password>@", ":<password>", ":<port>", and "/<url-path>" may be excluded.

password - An optional password. If present, it follows the user name separated from it by a colon.

there is no way to specify a password without specifying a user name

[sloria]

Oh good catch. So the current validation appears to be correct then. I think we can close this

[deckar01]

I started questioning why Redis would go off spec and discovered RFC1738 was the wrong spec to refer to. Redis based their URI scheme off of the NWG's 2005 RFC3986 which supersedes the 1994 RFC1738.

https://www.iana.org/assignments/uri-schemes/prov/redis

It no longer imposes restrictions on userinfo other than character set.

https://www.rfc-editor.org/rfc/rfc3986#section-3.2.1

RFC2396 dropped the "user:password" rules and left it up to the protocol.

https://datatracker.ietf.org/doc/html/rfc2396#appendix-G.2

[sloria]

closed by #2244