urllib.parse.parse_qs should not parse a query string containing illegal characters

[declaresub]

Bug report

Bug description:

urllib.parse.parse_qsl will happily parse a query string containing '#'.

from urllib.parse import parse_qsl
parse_qsl('foo=#', strict_parsing=True)

Output is [('foo', '#')] .

But 'foo=#' is an invalid query string according to RFC 3986. Similarly, '[', and ']' are excluded from the set of valid query characters, but parse_qsl parses strings like 'foo=[' and 'foo=]' . In the absence of any allocation of responsibility, this looks like a bug to me.

CPython versions tested on:

3.13

Operating systems tested on:

macOS

[Laxyny]

Interesting edge case — since # is reserved as a fragment delimiter in RFC 3986, I agree it probably shouldn't be parsed as a valid value. Might be worth aligning behavior with urlsplit here.

[terryjreedy]

The module is documented as following rfc3986. Does it say anything about what should raise an error for this function? If so, it might help to quote a bit or refer to a page/header.