Re: [PATCH] docs: freeze some package versions for sphinx virtualenv setup

[Vegard Nossum <vegard.nossum@oracle.com>]

On 27/02/2024 09:24, Lukas Bulwahn wrote:
> On Tue, Feb 27, 2024 at 7:57 AM Akira Yokosawa <akiyks@gmail.com> wrote:
>> On Mon, 26 Feb 2024 10:38:54 +0100, Lukas Bulwahn wrote:
>>> Akira-san Yokosawa reported this already in January 2023 and Jani Nikula
>>> was fast in providing a fix (see Link). This however remained as a thread
>>> on the linux-doc mailing and was never turned into a patch for Jonathan to
>>> pick up.
>>
>> Have you read Jon's message in the thread archived at:
>>
>>      https://lore.kernel.org/linux-doc/874jf4m384.fsf@meer.lwn.net/
>>
>> I think you can see why there had been no submission of a proper patch
>> in your patch's direction nor others.
>>
>> If you still want your patch applied soon, you need to convince Jon.

[...]

> Hence, I proposed this patch here. It works (for me). But I do not
> care if it is picked or not---because I am just continuing the
> whack-a-mole game. As of now, I do not know how much work the better
> solution for "updating to recent versions" needs. I will spend half a
> day and hopefully I can then present a good working patch. But let us
> see.

I now locally have a file requirements-2.4.4.txt that was the result of
running 'pip freeze' in a working Sphinx-2.4.4 virtualenv:

alabaster==0.7.13
Babel==2.14.0
certifi==2024.2.2
charset-normalizer==3.3.2
docutils==0.20.1
idna==3.6
imagesize==1.4.1
Jinja2==3.0.3
MarkupSafe==2.1.5
packaging==23.2
Pygments==2.17.2
PyYAML==6.0.1
requests==2.31.0
snowballstemmer==2.2.0
Sphinx==2.4.4
sphinxcontrib-applehelp==1.0.4
sphinxcontrib-devhelp==1.0.2
sphinxcontrib-htmlhelp==2.0.1
sphinxcontrib-jsmath==1.0.1
sphinxcontrib-qthelp==1.0.3
sphinxcontrib-serializinghtml==1.1.5
urllib3==2.2.0

My preferred way forward would be to put this in
Documentation/sphinx/requirements-2.4.4.txt and then update
Documentation/sphinx/requirements.txt to be completely unconstrained as
much as possible (just specify the bare minimum set of packages with
only lower bounds on the versions).

If you agree with the approach, feel free to submit that as a patch (if
not, I can submit one in the next few days), I just didn't want to snipe
your contribution.


Vegard