Bug 2537327 - python-astropy: FTBFS in Fedora Rawhide/F46 and Branched/F45
Summary: python-astropy: FTBFS in Fedora Rawhide/F46 and Branched/F45
Keywords:
Status: NEW
Alias: None
Product: Fedora
Classification: Fedora
Component: python-astropy
Version: rawhide
Hardware: Unspecified
OS: Unspecified
unspecified
unspecified
Target Milestone: ---
Assignee: Sergio Pascual
QA Contact: Fedora Extras Quality Assurance
URL: https://koschei.fedoraproject.org/pac...
Whiteboard:
Depends On:
Blocks: PPCTracker F45FTBFS F46FTBFS, RAWHIDEFTBFS
TreeView+ depends on / blocked
 
Reported: 2026-09-21 08:42 UTC by Ben Beasley
Modified: 2026-09-21 08:47 UTC (History)
6 users (show)

Fixed In Version:
Clone Of:
Environment:
Last Closed:
Type: Bug
Embargoed:


Attachments (Terms of Use)

Description Ben Beasley 2026-09-21 08:42:50 UTC
Description of problem:
Package python-astropy fails to build from source in Fedora Rawhide.

Version-Release number of selected component (if applicable):
8.0.1-3.fc45

Steps to Reproduce:
koji build --scratch f46 python-astropy-8.0.1-3.fc45.src.rpm

Additional info:
This package is tracked by Koschei. See:
https://koschei.fedoraproject.org/package/python-astropy

The failure is on ppc64le only; the first failing Rawhide build in Koschei was https://koschei.fedoraproject.org/build/23576949, which helps narrow down the possible causes. It seems like the update of scipy from 1.16 to 1.18 should be the prime suspect here.

=================================== FAILURES ===================================
_______________ test_polynomial_poorly_conditioned[fixed2-False] _______________
fixed = {'c0_1': True}, warns = False
    @pytest.mark.parametrize(
        "fixed, warns",
        [
            ({}, True),  # tests fitting non-fixed parameters models produces warnings
            (
                {"c1_0": True},
                True,
            ),  # tests fitting fixed par models produces warnings - #14037
            (
                {"c0_1": True},
                False,
            ),  # https://github.com/astropy/astropy/pull/14037#pullrequestreview-1191726872
        ],
    )
    def test_polynomial_poorly_conditioned(fixed, warns):
        p0 = models.Polynomial2D(degree=1, c0_0=3, c1_0=5, c0_1=0, fixed=fixed)

        fitter = LinearLSQFitter()

        x = [1, 2, 3, 4, 5]
        y = [1, 1, 1, 1, 1]
        values = p0(x, y)

        if warns:
            with pytest.warns(
                AstropyUserWarning, match="The fit may be poorly conditioned"
            ):
                p = fitter(p0, x, y, values)
        else:
            p = fitter(p0, x, y, values)
>           assert np.allclose(p0.parameters, p.parameters, rtol=0, atol=1e-14)
E           assert False
E            +  where False = <function allclose at 0x7ffef24936b0>(array([3., 5., 0.]), array([3., 5., 0.]), rtol=0, atol=1e-14)
E            +    where <function allclose at 0x7ffef24936b0> = np.allclose
E            +    and   array([3., 5., 0.]) = <Polynomial2D(1, c0_0=3., c1_0=5., c0_1=0.)>.parameters
E            +    and   array([3., 5., 0.]) = <Polynomial2D(1, c0_0=3., c1_0=5., c0_1=0.)>.parameters
../../../BUILDROOT/usr/lib64/python3.15/site-packages/astropy/modeling/tests/test_fitters.py:1108: AssertionError
=========================== short test summary info ============================
FAILED ../../modeling/tests/test_fitters.py::test_polynomial_poorly_conditioned[fixed2-False]
= 1 failed, 31026 passed, 979 skipped, 4 deselected, 263 xfailed in 494.84s (0:08:14) =


Note You need to log in before you can comment on or make changes to this bug.