[student-shivamm]

What type of PR is this?

• Bug Fix

Description

Registering a Unicode hostname with an explicit port can raise
urllib3.exceptions.LocationParseError instead of matching the same URL passed
to Requests. _clean_unicode() currently treats the full authority as DNS
labels, so example.рф:8443 becomes example.xn--:8443-uye2a. Userinfo can
likewise become part of an encoded hostname label.

Separate userinfo and the port before applying the existing Punycode conversion
to hostname labels. This also fixes Unicode add_passthru() prefixes while
preserving the significance of credentials, ports and paths.

Minimal reproduction on master:

import requests
import responses

with responses.RequestsMock(assert_all_requests_are_fired=False) as mock:
    url = "https://example.рф:8443/path"
    mock.get(url, body="matched")
    assert requests.get(url).text == "matched"

No new dependency or public API is added.

Related Issues

No existing matching issue found.

PR checklist

• Read the contributing guidelines.
• Ran complete tox and all-files pre-commit checks locally.
  Applicable configured pre-commit hooks pass for all three changed files.
  Full Python 3.12 test runs pass with Requests 2.34.2 / urllib3 2.8.0 and with
  Requests 2.30.0 / urllib3 1.26.20: 278 passed in each environment.
  Other Python versions/full tox matrix were not run.
  Mypy 2.4.0 reports the same three existing errors on pinned master and this
  patch, in unchanged matchers.py and test_recorder.py; it is not claimed to pass.
• Added the change to CHANGES.

Added/updated tests?

• Yes: 17 parameterized cases cover matching, stubbed passthrough and
  significant differences in authority/path components. The identical suite
  on master yields 15 failed / 263 passed; the patch yields 278 passed.
  All changed executable encoding lines are covered.