Rebuild for python 3.14

[h-vetinari]

Bot is stuck on missing dependencies; numba (needed for sparse) will take a while yet to support 3.14.

[conda-forge-admin]

Hi! This is the friendly automated conda-forge-linting service.

I just wanted to let you know that I linted all conda-recipes in your PR (recipe/meta.yaml) and found it was in an excellent condition.

I do have some suggestions for making it better though...

For recipe/meta.yaml:

• ℹ️ The recipe is not parsable by parser conda-souschef (grayskull). This parser is not currently used by conda-forge, but may be in the future. We are collecting information to see which recipes are compatible with grayskull.

_{This message was generated by GitHub Actions workflow run https://github.com/conda-forge/conda-forge-webservices/actions/runs/18185256803. Examine the logs at this URL for more detail.}

[h-vetinari]

Let's wait for pandas at least

[EwoutH]

Python 3.14 is releasing in two (!) weeks. Is Pandas blocking?

[h-vetinari]

Is Pandas blocking?

Pandas has an optional dependency with an uncapped constraint on pyarrow. I want this interaction tested in advance.

Python 3.14 is releasing in two (!) weeks.

Aside from the fact that this is simply a reality due to the situation with pandas, the correct frame of reference is IMO that we're already months faster than the old days of starting to migrate (and make packages compatible) after 3.N.0 GA.

If anyone has an expectation that everything will (or should) be ready when python 3.N releases, they're mistaken, and we need to disabuse them of that notion, because all this is based merely on a coincidence of wants¹ in the ecosystem, with zero assurances on anyone's part except a voluntary best effort with no specific delivery date.

Footnotes

1. in the sense that they want the same thing (the ecosystem becoming compatible with a new python version), not that it's accidental. ↩

[weiji14]

@conda-forge-admin, please rerender

Let's see if this works after conda-forge/pandas-feedstock#233 is merged.

[weiji14]

Interesting, Python 3.14 tests run fine on CUDA-enabled builds, but fail on non-CUDA builds. E.g. at https://dev.azure.com/conda-forge/feedstock-builds/_build/results?buildId=1355697&view=logs&j=bc49ae96-d8c4-56c3-76c1-860716735ea0&t=0253630d-55a2-5a8f-6f1e-0f6dde5e58c0&l=7641, there is this one test failure:

___________________________ test_to_pandas_zero_copy ___________________________

    @pytest.mark.pandas
    def test_to_pandas_zero_copy():
        import gc
    
        arr = pa.array(range(10))
    
        for i in range(10):
            series = arr.to_pandas()
>           assert sys.getrefcount(series) == 2
E           assert 1 == 2
E            +  where 1 = <built-in function getrefcount>(0    0\n1    1\n2    2\n3    3\n4    4\n5    5\n6    6\n7    7\n8    8\n9    9\ndtype: int64)
E            +    where <built-in function getrefcount> = sys.getrefcount

pyarrow/tests/test_array.py:295: AssertionError

I see at apache/arrow@f58ba4c (in apache/arrow#47616) that a refcount of 1 or 2 should be valid (makes sense since this should be zero copy). Maybe apply that commit as a patch?

[h-vetinari]

Python 3.14 tests run fine on CUDA-enabled builds, but fail on non-CUDA builds.

That's because we don't run the tests for the CUDA builds, because we only build/publish them once.

pyarrow-feedstock/recipe/meta.yaml

Lines 214 to 219 in 1e54db0

	- name: pyarrow-tests
	script: build-pyarrow.sh # [unix]
	script: build-pyarrow.bat # [win]
	version: {{ version }}
	build:
	skip: true # [cuda_compiler_version != "None"]

[conda-forge-admin]

Hi! This is the friendly conda-forge automerge bot!

I considered the following status checks when analyzing this PR:

• linter: passed
• azure: passed

Thus the PR was passing and merged! Have a great day!