Re: [PATCH v2 0/7] Python: Drop support for Python 3.6

[Markus Armbruster <armbru@redhat.com>]

Peter Maydell <peter.maydell@linaro.org> writes:

> On Wed, 15 Feb 2023 at 19:05, Markus Armbruster <armbru@redhat.com> wrote:
>>
>> The discussion under PATCH 6 makes me think there's a bit of confusion
>> about the actual impact of dropping support for Python 3.6.  Possibly
>> because it's spelled out in the commit message of PATCH 7.  Let me
>> summarize it in one sentence:
>>
>>     *** All supported host systems continue to work ***
>>
>> Evidence: CI remains green.
>>
>> On some supported host systems, different packages need to be installed.
>> On CentOS 8, for instance, we need to install Python 3.8.13 or 3.9.16
>> instead of 3.6.8.  Let me stress again: same repository, different
>> package.  No downsides I can see.
>
> Yes; I never had any issues with this part of it. If there was
> a "Sphinx that also used that Python" in that repo, the answer
> would be easy.
>
>> The *one* exception is Sphinx on CentOS 8.  CentOS 8 does not ship a
>> version of Sphinx that works with Python 3.7 or newer.  This series
>> proposes to simply stop building the docs there, unless the user
>> provides a suitable version of Sphinx (which is easy enough with pip).
>> That's PATCH 7.
>
> Yes; this brings CentOS 8 down from "fully supported" to "kinda
> supported but not for everything", which is less than ideal.

I acknowledge there's a difference between "you need to dnf install
python-sphinx to be able to build docs" and "you need to pip install
sphinx to be able to build docs", and that the difference is negligible
in some scenarios, and a show stopper in others.  I wasn't fully aware
of the latter.

> I think the level of not-idealness of that side of the scales
> is probably clear enough. The difficulty I think for those who
> haven't had their arms deep in QEMU's Python code is not having
> the background info to be able to weigh up how heavy the other side
> of the tradeoff scales is (since the naive "well, just keep writing
> Python 3.6 for the moment" take is clearly wrong).

Fair point.  I hope the situation is more clear now.