Bug#1039870: release.debian.org: Help eccodes migrate to testing

[Bas Couwenberg <sebastic@xs4all.nl>]

Package: release.debian.org
Severity: normal

eccodes is no longer available on 32bit architectures and FTBFS on big endian architectures, it has been removed from armel,armhf,i386,mipsel,s390x in unstable (#1039537)

This required partial removals of its rdeps too:

eccodes (#1039537)
- cdo (#1039538)
  - python-cdo (#1039554)
- cfgrib (#1039539)
  - python-xarray [arch:all]
- eccodes-python (#1039540)
  - cfgrib
  - satpy [arch:all]
- ecmwflibs (#1039541)
  - eccodes-python
- emoslib (#1039542)
  - magics++
  - metview
- fdb [s390x] (#1039543)
  - metview
- flexpart (#1039544)
  - metview
- flextra (#1039545)
  - metview
- glgrib (#1039546)
- gnudatalanguage (#1039547)
  - coyote [arch:all]
  - idlastro [arch:all]
  - mpfit [arch:all]
- magics++ (#1039548)
  - cdo
  - ecmwflibs
  - magics-python
  - metview
  - metview-python
- magics-python (#1039549)
- metkit [s390x] (#1039550)
  - fdb
  - metview
- metview [s390x] (#1039551)
- pygrib (#1039552)
  - pyaps3 [arch:all]
  - satpy [arch:all]
- ruby-grib (#1039553)

fdb, metkit, and metview were already removed from the 32bit architectures, and only needed to be removed from s390x.

The [arch:all] rdeps didn't need intervention.


britney seems to be having problems performing the removals from testing, the excuses for eccodes include:

 removing libeccodes-dev/2.28.0-1/i386 from testing makes python3-eccodes/2:1.5.0-1/i386 uninstallable
 removing libeccodes-dev/2.28.0-1/i386 from testing makes python3-ecmwflibs/2:0.5.0-1+b1/i386 uninstallable

The excuses for magics++ have similar issues:

 removing libmagics++-dev/4.13.0-1/i386 from testing makes python3-ecmwflibs/2:0.5.0-1+b1/i386 uninstallable

These removing issues are repeated for armel,armhf,mipsel,s390x.


Please help to remove these old binaries to unblock the testing migration of eccodes and its rdeps.


Kind Regards,

Bas