`urllib.parse.parse_qsl` is accepting illegal characters

[shloktech]

Bug report

Bug description:

urllib.parse.parse_qsl parses query strings containing the ^ and ` characters, even though these are not valid query characters under RFC 3986.

Observed behaviour:

parse_qsl('foo=^', strict_parsing=True)
# [('foo', '^')]

parse_qsl('bar=`', strict_parsing=True)
# [('bar', '`')]

Expected behaviour:
According to RFC 3986, both ^ and ` must be percent-encoded if used in a URI. However, parse_qsl accepts them as-is without raising an error or warning. This could lead to applications treating invalid URLs as valid.

Detailed Code:

import sys
import platform
from urllib.parse import parse_qsl


def test_parse_qsl(query):
    try:
        result = parse_qsl(query, strict_parsing=True)
        print(f"Query: {query!r} -> Parsed: {result}")
    except ValueError as e:
        print(f"Query: {query!r} -> Error: {e}")


# Test invalid query strings
test_parse_qsl("foo=^")
test_parse_qsl("bar=`")

# System information
os_name = platform.system()
os_release = platform.release()
python_impl = platform.python_implementation() 
python_version = sys.version.split()[0]
python_compiler = platform.python_compiler() 

print("\n--- System Information ---")
print(f"Python Implementation : {python_impl}")
print(f"Python Version : {python_version}")
print(f"Python Compiler       : {python_compiler}")
print(f"Operating System : {os_name} {os_release}")
print(f"Machine         : {platform.machine()}")

# Output:
# Query: 'foo=^' -> Parsed: [('foo', '^')]
# Query: 'bar=`' -> Parsed: [('bar', '`')]

# --- System Information ---
# Python Implementation : CPython
# Python Version : 3.13.1
# Python Compiler       : GCC 14.2.0
# Operating System : Linux 4.14.174
# Machine         : x86_64

---

References:

1. Section 2.2 Reserved Characters: https://datatracker.ietf.org/doc/html/rfc3986#section-2.2

2. Section 2.3 Unreserved Characters: https://datatracker.ietf.org/doc/html/rfc3986#section-2.3

CPython versions tested on:

3.13

Operating systems tested on:

Linux

Linked PRs

• gh-138284 : urllib.parse.parse_qsl now raises ValueError if illegal characters is passed, according to RFC 3986 #138291

[shloktech]

How did I discover it?
I was working on a new version of https://pypi.org/project/md2docx-python/
This utility converts markdown to docx and vice versa. In this version, there was a planned feature in the new version which revolves around parse_qsl. While doing a sanity check around my changes, I discovered this issue. For now, I have addressed the issue by adding a sanity check before using parse_qsl