Test failures w/ charset-normalizer 2.0.7+

[mgorny]

On systems w/ charset-normalizer 2.0.7 or newer, the "Österreich" tests fail:

============================================================== FAILURES ===============================================================
_________________________________________________ test_text_decoder[data3-iso-8859-1] _________________________________________________

data = (b'Accented: \xd6sterreich abcdefghijklmnopqrstuvwxyz', b''), encoding = 'iso-8859-1'

    @pytest.mark.parametrize(
        ["data", "encoding"],
        [
            ((b"Hello,", b" world!"), "ascii"),
            ((b"\xe3\x83", b"\x88\xe3\x83\xa9", b"\xe3", b"\x83\x99\xe3\x83\xab"), "utf-8"),
            ((b"Euro character: \x88! abcdefghijklmnopqrstuvwxyz", b""), "cp1252"),
            ((b"Accented: \xd6sterreich abcdefghijklmnopqrstuvwxyz", b""), "iso-8859-1"),
        ],
    )
    @pytest.mark.asyncio
    async def test_text_decoder(data, encoding):
        async def iterator():
            nonlocal data
            for chunk in data:
                yield chunk
    
        # Accessing `.text` on a read response.
        response = httpx.Response(
            200,
            content=iterator(),
        )
        await response.aread()
>       assert response.text == (b"".join(data)).decode(encoding)
E       AssertionError: assert 'Accented: ųs...nopqrstuvwxyz' == 'Accented: Ös...nopqrstuvwxyz'
E         - Accented: Österreich abcdefghijklmnopqrstuvwxyz
E         ?           ^
E         + Accented: ųsterreich abcdefghijklmnopqrstuvwxyz
E         ?           ^

tests/test_decoders.py:199: AssertionError
__________________________________________ test_response_no_charset_with_iso_8859_1_content ___________________________________________

    def test_response_no_charset_with_iso_8859_1_content():
        """
        A response with ISO 8859-1 encoded content should decode correctly,
        even with no charset specified.
        """
        content = "Accented: Österreich abcdefghijklmnopqrstuzwxyz".encode("iso-8859-1")
        headers = {"Content-Type": "text/plain"}
        response = httpx.Response(
            200,
            content=content,
            headers=headers,
        )
>       assert response.text == "Accented: Österreich abcdefghijklmnopqrstuzwxyz"
E       AssertionError: assert 'Accented: ųs...nopqrstuzwxyz' == 'Accented: Ös...nopqrstuzwxyz'
E         - Accented: Österreich abcdefghijklmnopqrstuzwxyz
E         ?           ^
E         + Accented: ųsterreich abcdefghijklmnopqrstuzwxyz
E         ?           ^

tests/models/test_responses.py:240: AssertionError

2.0.7 was released already two months ago; could you please either update the tests to work against this version, or report a bug against charset-normalizer if this is a bug?