Package Details: python-mcp 1.27.1-2

Git Clone URL: https://aur.archlinux.org/python-mcp.git (read-only, click to copy)
Package Base: python-mcp
Description: Model Context Protocol SDK.
Upstream URL: https://github.com/modelcontextprotocol/python-sdk
Licenses: MIT
Submitter: medaminezghal
Maintainer: medaminezghal
Last Packager: medaminezghal
Votes: 5
Popularity: 1.11
First Submitted: 2025-02-04 16:43 (UTC)
Last Updated: 2026-05-15 22:01 (UTC)

Pinned Comments

medaminezghal commented on 2026-05-15 22:58 (UTC)

@everyone I’m working towards making this package automatically updated by GitHub CI, this is my repository used to maintain my packages. Every package before it’s updated should be successfully builded in clean environment (ensured by CI). So if you encounter any errors, just clean the cache and try again. Everyone is welcomed to help improve packaging by making issue or improve the CI by pull request (I’m new to CI, in fact I wrote it using AI, I think it still need improvement, I would be happy if anyone put his experience to help improve the CI).

Latest Comments

« First ‹ Previous 1 2 3 4 5 6 7 8 9 Next › Last »

gwuensch commented on 2026-01-12 15:48 (UTC)

python-pytest-benchmark is a required dependency for python-cuda and python-pydantic

This is simply not true. python-pytest-benchmark is not a runtime dependency for any of these packages. If you are installing pre-built packages from the official repos, you shouldn't require python-pytest-benchmark at all. Even if you are building locally, you can remove the package after the build has finished (makepkg -sr). Or just build in a clean chroot.

Having said all this, I agree that it would be a good solution if the PKGBUILD could just explicitly disable the benchmarks so people can stop complaining about this.

sashank commented on 2026-01-12 03:17 (UTC)

Hey @medaminezghal, python-pytest-benchmark is a required dependency for python-cuda and python-pydantic (which is a requirement for python-google-genai). It's untenable to uninstall it. In fact, even though python-pytest-benchmark ostensibly doesn't play nicely with python-pytest-xdist, it xdist is listed as a required dependency here on the AUR.

So, I think the better bet, given how both pytest extensions are required for running tests in various AI oriented packages is to either patch the pyproject.toml file or modify the pytest command to not run benchmark.

Those options are discussed here by the benchmark devs: https://github.com/ionelmc/pytest-benchmark/issues/286

gwuensch commented on 2025-12-28 15:31 (UTC)

@PhCl please read the previous two comments and you'll have your answer.

PhCl commented on 2025-12-28 15:27 (UTC)

"check" fails for me:

==> Starting check()...
INTERNALERROR> Traceback (most recent call last):
INTERNALERROR>   File "/usr/lib/python3.13/site-packages/_pytest/main.py", line 285, in wrap_session
INTERNALERROR>     config._do_configure()
INTERNALERROR>     ~~~~~~~~~~~~~~~~~~~~^^
INTERNALERROR>   File "/usr/lib/python3.13/site-packages/_pytest/config/__init__.py", line 1119, in _do_configure
INTERNALERROR>     self.hook.pytest_configure.call_historic(kwargs=dict(config=self))
INTERNALERROR>     ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^
INTERNALERROR>   File "/usr/lib/python3.13/site-packages/pluggy/_hooks.py", line 534, in call_historic
INTERNALERROR>     res = self._hookexec(self.name, self._hookimpls.copy(), kwargs, False)
INTERNALERROR>   File "/usr/lib/python3.13/site-packages/pluggy/_manager.py", line 120, in _hookexec
INTERNALERROR>     return self._inner_hookexec(hook_name, methods, kwargs, firstresult)
INTERNALERROR>            ~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
INTERNALERROR>   File "/usr/lib/python3.13/site-packages/pluggy/_callers.py", line 167, in _multicall
INTERNALERROR>     raise exception
INTERNALERROR>   File "/usr/lib/python3.13/site-packages/pluggy/_callers.py", line 121, in _multicall
INTERNALERROR>     res = hook_impl.function(*args)
INTERNALERROR>   File "/usr/lib/python3.13/site-packages/pytest_benchmark/plugin.py", line 448, in pytest_configure
INTERNALERROR>     bs = config._benchmarksession = BenchmarkSession(config)
INTERNALERROR>                                     ~~~~~~~~~~~~~~~~^^^^^^^^
INTERNALERROR>   File "/usr/lib/python3.13/site-packages/pytest_benchmark/session.py", line 65, in __init__
INTERNALERROR>     self.logger.warn(
INTERNALERROR>     ~~~~~~~~~~~~~~~~^
INTERNALERROR>         "Benchmarks are automatically disabled because xdist plugin is active."
INTERNALERROR>         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
INTERNALERROR>         "Benchmarks cannot be performed reliably in a parallelized environment.",
INTERNALERROR>         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
INTERNALERROR>     )
INTERNALERROR>     ^
INTERNALERROR>   File "/usr/lib/python3.13/site-packages/pytest_benchmark/logger.py", line 46, in warn
INTERNALERROR>     warner(PytestBenchmarkWarning(text))
INTERNALERROR>     ~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
INTERNALERROR> pytest_benchmark.logger.PytestBenchmarkWarning: Benchmarks are automatically disabled because xdist plugin is active.Benchmarks cannot be performed reliably in a parallelized environment.
==> ERROR: A failure occurred in check().

medaminezghal commented on 2025-12-03 09:09 (UTC)

@bitgamma python-pytest-benchmark conflicts with python-pytest-xdist. Just remove python-pytest-benchmark and it will work.

bitgamma commented on 2025-12-03 08:41 (UTC)

Build fails here with:

Benchmarks are automatically disabled because xdist plugin is active.Benchmarks cannot be performed reliably in a parallelized environment.

Workaround: setting PYTEST_XDIST_AUTO_NUM_WORKERS=0

medaminezghal commented on 2025-11-04 06:32 (UTC)

@dreieck @gwuensch Thank you for your suggestions.

gwuensch commented on 2025-11-03 21:28 (UTC)

Missing dependency python-pyjwt as of #1247

gwuensch commented on 2025-11-03 19:39 (UTC)

your PKGBUILD needlessly clones the whole upstream repository again at version change.

In case this is changed (I'm neutral on whether this is actually necessary), it would be nice if prepare() could include git clean -fdx to remove stale wheels and other build artifacts that commonly lead to packaging errors.

dreieck commented on 2025-11-03 17:02 (UTC)

@dreieck that's weird, works for me™. Error message says your uv version is too old. Maybe try upgrading, the current version in extra is 0.9.7, not 0.9.3.

It turned out that it is Artix that is lagging behind.

Regards!

« First ‹ Previous 1 2 3 4 5 6 7 8 9 Next › Last »