Re: [OE-core] [PATCH v2 4/5] python3-websockets: enable ptest; use git fetcher

["Tim Orling" <tim.orling@konsulko.com>]

On Thu Nov 5, 2026 at 6:41 AM PST, Ross Burton wrote:
> On 3 Nov 2026, at 19:38, Tim Orling via openembedded-core <tim.orling=konsulko.com@lists.openembedded.org> wrote:
>> The PyPI sdist prunes tests/, so fetch the 17.1 tag from git. This
>> re-enables ptest, reverted in a0488f2e1d when the sdist dropped the tests.
>> 
>> Upstream advises packagers against running the suite because timeout
>> tests can flake on slow build farms; under ptest it runs on the target,
>> and it passed repeatedly on qemux86-64.
>
> Whilst I approve of ptests in general, I think this one will be a cause of pain.
>
> Tests are not shipped in the sdists on purpose[1], and upstream is very clear that the tests are not for packagers[2]:
>
> “”"
> If you’re packaging websockets for a distribution, you must use releases published on PyPI as input. You may check SLSA attestations on GitHub.
> You mustn’t rely on the git repository as input. Specifically, you mustn’t attempt to run the main test suite. It isn’t treated as a deliverable of the project. It doesn’t do what you think it does. It’s designed for the needs of developers, not packagers.
> On a typical build farm for a distribution, tests that exercise timeouts will fail randomly. Indeed, the test suite is optimized for running very fast, with a tolerable level of flakiness, on a high-end laptop without noisy neighbors. This isn’t your context.
> “””
>
> This feels like it will be too flaky for the autobuilder, even if we remove tests which have tight timeouts.

As much as I want ptest to feel comfortable with updates, I cannot deny your conclusions.

I am abandoning this series.

--Tim

>
> Ross
>
> [1] https://github.com/python-websockets/websockets/issues/1739#issuecomment-5016084799
> [2] https://websockets.readthedocs.io/en/stable/project/contributing.html#packaging