Test regression in test_response_decode_text_using_autodetect with chardet 6.0.0

[musicinmybrain]

While httpx doesn’t depend on chardet anymore, it still uses it for some tests. The chardet package has just released version 6.0.0, a major update with a new “encoding era” concept and a number of other significant changes.

This leads to a regression in the tests. The test skips here are the same ones we currently use downstream in Fedora; they may now be overly expansive, but they allow me to demonstrate the regression I’m talking about without having to contend with pre-existing issues. Certainly, many tests hang if I skip nothing at all.

$ git clone https://github.com/encode/httpx.git
$ cd httpx
$ uv venv
$ . .venv/bin/activate
(httpx) $ uv pip install -r requirements.txt
(httpx) $ pytest -v -k 'not client and not test_api and not test_config and not test_exceptions and not test_main and not test_timeouts and not test_utils'
[…]
=========================================================================== 1003 passed, 415 deselected in 0.96s ============================================================================
(httpx) $ uv pip install --upgrade chardet
[…]
 - chardet==5.2.0
 + chardet==6.0.0
(httpx) $ pytest -v -k 'not client and not test_api and not test_config and not test_exceptions and not test_main and not test_timeouts and not test_utils'
[…]
========================================================================================= FAILURES ==========================================================================================
________________________________________________________________________ test_response_decode_text_using_autodetect _________________________________________________________________________

    def test_response_decode_text_using_autodetect():
        # Ensure that a 'default_encoding="autodetect"' on the response allows for
        # encoding autodetection to be used when no "Content-Type: text/plain; charset=..."
        # info is present.
        #
        # Here we have some french text encoded with ISO-8859-1, rather than UTF-8.
        text = (
            "Non-seulement Despréaux ne se trompait pas, mais de tous les écrivains "
            "que la France a produits, sans excepter Voltaire lui-même, imprégné de "
            "l'esprit anglais par son séjour à Londres, c'est incontestablement "
            "Molière ou Poquelin qui reproduit avec l'exactitude la plus vive et la "
            "plus complète le fond du génie français."
        )
        content = text.encode("ISO-8859-1")
        response = httpx.Response(200, content=content, default_encoding=autodetect)
    
        assert response.status_code == 200
        assert response.reason_phrase == "OK"
>       assert response.encoding == "ISO-8859-1"
E       AssertionError: assert 'WINDOWS-1252' == 'ISO-8859-1'
E         
E         - ISO-8859-1
E         + WINDOWS-1252

tests/models/test_responses.py:1014: AssertionError
====================================================================== 1 failed, 1002 passed, 415 deselected in 1.76s =======================================================================

For this particular case, the byte string is consistent with both encoding names,

>>> content.decode('WINDOWS-1252') == content.decode('ISO-8859-1')
True

so one claimed encoding isn’t necessarily more correct than the other.

I plan to open a PR that will adjust the test to accept either detected encoding.

[musicinmybrain]

Addressed via #3773.