Bug#811219: transition: netcdf
On 22-01-16 10:42, Sebastiaan Couwenberg wrote:
> On 21-01-16 20:17, Sebastiaan Couwenberg wrote:
>> On 21-01-16 11:20, Emilio Pozuelo Monfort wrote:
>>> On 17/01/16 00:32, Bas Couwenberg wrote:
>>>> NetCDF 4.4.0 final has been released and bumps the SOVERSION to 11
>>>> accounting for the removed symbols in RC4. Fortunately we only have to
>>>> transition netcdf, and not also the C++ & Fortran packages. Only a few
>>>> packages FTBFS:
>>>>
>>>> minc (2.2.00-6) FTBFS, this is the same issue as the for the previous
>>>> netcdf transition (#793885), which has been fixed in libminc, but not minc.
>>>>
>>>> python-scientific (2.9.4-3) FTBFS due to an old issue too (#799195).
>>>>
>>>> mmtk (2.7.9-1) FTBFS due to #798665, and requires python-scientific.
>>>>
>>>> These packages are sid-only already, so shouldn't be much of an issue
>>>> for this transition.
>>>
>>> I was waiting for the python3 transition, which is now done. So go ahead.
>>
>> Thanks. I've just uploaded netcdf (1:4.4.0-1) to unstable.
>
> netcdf-cxx, netcdf-cxx-legacy, netcdf-fortran & netcdf4-python have
> already been rebuilt with netcdf (1:4.4.0-1) on the release architectures.
>
> The transition tracker does still need to be updated to also mark the
> packages still depending on the netcdf (1:4.1.3-7.2) from the previous
> transitions. Using the attached ben configuration (or the one in the
> initial bugreport) will also include gerris, kst, minc & mmtk in the
> tracker.
>
> minc & mmtk both FTBFS and are sid-only as noted before.
>
> gerris links to netcdf but the resulting packages don't depend on any
> netcdf packages.
>
> kst doesn't detect the new netcdf libraries any more, and disables the
> netcdf plugin because of that.
It seems that leftovers from the previous transition are preventing
testing migration again.
I guess we can't remove the old netcdf (1:4.1.3-7.2) packages because
some un- or badly maintained packages still depend on them.
The hints for the previous transition got the netcdf packages migrated
to testing, but it did prevent migration of a subsequent
(non-transition) update because of the same missing binaries issues.
Can we fix this once and for all, or do we have to rely on hints to
migrate netcdf packages every time?
Kind Regards,
Bas
--
GPG Key ID: 4096R/6750F10AE88D4AF1
Fingerprint: 8182 DE41 7056 408D 6146 50D1 6750 F10A E88D 4AF1
Reply to: