metpy_1.5.1+ds-2_source.changes ACCEPTED into unstable

[Antonio Valentino]

Dear Bas,

Il 02/08/23 21:24, Sebastiaan Couwenberg ha scritto:
> On 8/2/23 05:29, Sebastiaan Couwenberg wrote:
>> Testing migration is blocked by autopkgtest failure on arm64:
>>
>> 167s =================================== FAILURES ===================================
>> 167s ___________________________ test_moist_lapse_failure ___________________________
>> 167s
>> 167s     @pytest.mark.xfail(sys.platform == 'win32', reason='solve_ivp() does not error on Windows')
>> 167s     @pytest.mark.xfail(packaging.version.parse(scipy.__version__) < packaging.version.parse('1.7'),
>> 167s                        reason='solve_ivp() does not error on Scipy < 1.7')
>> 167s     def test_moist_lapse_failure():
>> 167s         """Test moist_lapse under conditions that cause the ODE solver to fail."""
>> 167s         p = np.logspace(3, -1, 10) * units.hPa
>> 167s >       with pytest.raises(ValueError) as exc:
>> 167s E       Failed: DID NOT RAISE <class 'ValueError'>
>> 167s
>> 167s tests/calc/test_thermo.py:206: Failed
>>
>> https://ci.debian.net/data/autopkgtest/testing/arm64/m/metpy/36313527/log.gz
> 
> I've added a patch to use xfail for this test on arm64 like it does on Windows and with old SciPy.
> 
> Do you think we should upstream that, or dig deeper why it doesn't raise an exception on arm64?

Thanks for taking the time to fix the issue.
This week I'm traveling, but ASAP I will look at it and, id it is the case, I will open an issue/PR upstream.


cheers
-- 
Antonio Valentino