Bug#949274: pyresample autopkg tests regressed in unstable

[Antonio Valentino]

Dear Matthias,
thank you for reporting.

On Sun, 19 Jan 2020 08:44:48 +0100 Matthias Klose <doko at debian.org> wrote:
> Package: src:pyresample
> Version: 1.14.0-2
> Severity: serious
> Tags: sid bullseye
> 
> https://ci.debian.net/data/autopkgtest/unstable/amd64/p/pyresample/4019144/log.gz
> 
> [...]
> ======================================================================
> FAIL: test_get_area_def_from_raster_extracts_proj_id
> (pyresample.test.test_utils.TestMisc)
> ----------------------------------------------------------------------
> Traceback (most recent call last):
>   File "/usr/lib/python3/dist-packages/pyresample/test/test_utils.py", line 477,
> in test_get_area_def_from_raster_extracts_proj_id
>     self.assertEqual(area_def.proj_id, 'WGS 84 / Pseudo-Mercator')
> AssertionError: 'WGS_1984_Web_Mercator_Auxiliary_Sphere' != 'WGS 84 /
> Pseudo-Mercator'
> - WGS_1984_Web_Mercator_Auxiliary_Sphere
> + WGS 84 / Pseudo-Mercator
> 
> 
> ----------------------------------------------------------------------
> Ran 271 tests in 193.325s
> 
> FAILED (failures=1, skipped=2)

I gave a very quick look to the issue and it seems that the problem has
be triggered bu the update to gdal 3.0.3.

-gdal-data 3.0.2+dfsg-1
+gdal-data 3.0.3+dfsg-1
-libgdal26 3.0.2+dfsg-1
+libgdal26 3.0.3+dfsg-1

rasterio could be also involved since the specific test uses rasterio to
retrieve the information about the geographic projection

I don't see any change in gdal data at a first look so probably the
problem is related behavior change in gdal that is not correctly handled
by pyresample or rasterio.

I need to make some additional investigation.


kind regards

-- 
Antonio Valentino