[Debian-astro-maintainers] Bug#952573: spectral-cube: autopkgtest failure (blocks astropy migration)

[peter green]

Source: spectral-cube
Version: 0.4.5-1
Severity: serious

Astropy was recently updated to support the new version of wcslib. Unfortunately this fix is being blocked from migrating to testing by an autopkgtest failure in spectral-cube, this is happening in both in migration tests and in plain unstable tests.

migration test: https://ci.debian.net/data/autopkgtest/testing/amd64/s/spectral-cube/4375843/log.gz

plain unstable test: https://ci.debian.net/data/autopkgtest/unstable/amd64/s/spectral-cube/4382183/log.gz

> =================================== FAILURES ===================================
> ____________________________ test_stacking_badvels _____________________________
>
>      def test_stacking_badvels():
>          '''
>          Regression test for #493: don't include bad velocities when stacking
>          '''
>      
>          amp = 1.
>          v0 = 0. * u.km / u.s
>          sigma = 8.
>          noise = None
>          shape = (100, 25, 25)
>      
>          test_cube, test_vels = \
>              generate_gaussian_cube(amp=amp, sigma=sigma, noise=noise,
>                                     shape=shape)
>      
>          true_spectrum = gaussian(test_cube.spectral_axis.value,
>                                   amp, v0.value, sigma)
>      
>          test_vels[12,11] = 500*u.km/u.s
>      
>          with warnings.catch_warnings(record=True) as wrn:
>              # Stack the spectra in the cube
>              stacked = \
>                  stack_spectra(test_cube, test_vels, v0=v0,
>                                stack_function=np.nanmean,
>                                xy_posns=None, num_cores=1,
>                                chunk_size=-1,
>                                progressbar=False, pad_edges=False)
> >           assert 'Some velocities are outside the allowed range and will be' in str(wrn[-1].message)
> E           assert 'Some velocities are outside the allowed range and will be' in "'datfix' made the change 'Set DATE-REF to '1858-11-17' from MJD-REF'."
> E            +  where "'datfix' made the change 'Set DATE-REF to '1858-11-17' from MJD-REF'." = str(FITSFixedWarning("'datfix' made the change 'Set DATE-REF to '1858-11-17' from MJD-REF'."))
> E            +    where FITSFixedWarning("'datfix' made the change 'Set DATE-REF to '1858-11-17' from MJD-REF'.") = <warnings.WarningMessage object at 0x7f3989229c50>.message
This seems similar to a testsuite failure that happened in astropy when the inital attempt was made to rebuild it for 7.1 ( https://buildd.debian.org/status/fetch.php?pkg=astropy&arch=amd64&ver=4.0-3%2Bb1&stamp=1582194992&raw=0 ) , said testsuite failure was apparently fixed by https://salsa.debian.org/debian-astro-team/astropy/-/commit/a7a71b323d66984dfc482ea95779e26018fc2cec . Perhaps a similar fix is needed in the spectral-cube tests?


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://alioth-lists.debian.net/pipermail/debian-astro-maintainers/attachments/20200226/ae8288c6/attachment-0001.html>