Reject repeated keys in strict urlencoded_params_matcher

[Atishyy27]

urlencoded_params_matcher parses the body with dict(parse_qsl(...)), which collapses repeated keys (last value wins). Under the default strict_match=True a body carrying a duplicate key then compares equal to a params dict that cannot represent it:

>>> from urllib.parse import parse_qsl
>>> dict(parse_qsl("a=1&a=2"))
{'a': '2'}          # a=1 is gone, so this equals the strict spec {"a": "2"}

So a strict matcher for {"a": "2"} wrongly matches a body of a=1&a=2 (for example data=[("a", "1"), ("a", "2")]), even though it sent an a=1 the spec never allowed. Same lossy-parse_qsl family as #804; query_string_matcher avoids it by comparing the parsed pair lists rather than dicts.

Fix: in the strict path, reject the body when the parsed pairs outnumber the collapsed dict. The non-strict path (meant to allow extra params) is untouched. Added test_urlencoded_params_matcher_strict_repeated_keys (fails on master, passes here) and a CHANGES entry. Prepared with AI assistance under my direction; I reviewed and tested it myself.

[Atishyy27]

All six workflow runs on this branch sit at conclusion: action_required, so no CI has run and there is nothing for a reviewer to look at yet. Could someone approve the run?