On 10-08-15 08:45, Alastair McKinstry wrote: > The big hurdle is the netcdf/gfortran transition: a number of > packages depend on netcdf.mod. As Matthias points out below, this > would involve a package rename anyway: do we wish to combine the > two, given that the netcdf transition has been tested in > experimental? For netcdf-fortran I don't see a problem other than the outstanding bugs to add the new build dependencies [1], etsf-io & oasis3 still need to build depend on the new libnetcdff-dev. Patches are available in the bugreports. The biggest hurdle for the netcdf transition [2] is that we can't update only netcdf-fortan we need to update the C & C++ packages too. The netcdf transition tracker still needs to be updated to handle the package split as mentioned in the transition bugreport (#791215). I'm not sure if it's wise to start the netcdf transition already, even though I would prefer to transition to the new netcdf packages instead of doing a v5 rename. A number of reverse dependencies cannot be built in unstable because the libdap transition (#791114) hasn't started yet, so libstdc++6 still breaks the current libdap version in unstable on which gdal among others (build) depends, and so most GIS package cannot be built because they (build) depend on gdal. This includes gmt, ncl & vtk6 that are part of the netcdf transition. I'd like to ask the opinion of the Release Team whether we should trigger the netcdf transition too, or do a v5 rename in the old packages after all. [1] https://bugs.debian.org/cgi-bin/pkgreport.cgi?tag=netcdf-split-c-f-cxx;users=debian-gis@lists.debian.org [2] https://bugs.debian.org/791215 Kind Regards, Bas -- GPG Key ID: 4096R/6750F10AE88D4AF1 Fingerprint: 8182 DE41 7056 408D 6146 50D1 6750 F10A E88D 4AF1
Attachment:
signature.asc
Description: OpenPGP digital signature