[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: netCDF Strategy

Copying Nico in as cc, as I don't think I have seen him post here.
Please Nico, if you read this only once - it would be great if you
subscribed to this list as well as the "commit/bug" list so everyone
knows what is happening with netcdf (or let me know that you are

I have been working through some of the old bugs against netcdf and I
have noticed a couple of things about the netcdf C package which I have
been testing the bugs against:

1. We claim in the changelog that Multi-Arch is enabled, but I am not
sure we have done this fully according to
We have the files installing to the correct paths, but I can't see the
"Multi-Arch" stanzas in d/control.

2. Whilst trying to get backtraces for a couple of "mayhem" bugs in gdb
(I will come back separately about this if I don't crack it soon), I
noticed the -dbg package does not seem to install the debugging symbols
properly (at least gdb does not pick them up automatically with this
version of netcdf - but it could be my poor use of gdb). The file list
is completely different between 4.3.4 and the old version of the debug
package. I am not sure why this is, as the d/rule & d/control looks
pretty standard.

On 01/17/2015 01:02 AM, Sebastiaan Couwenberg wrote:


> The netcdf-cxx & netcdf-fortan packages are mostly ready in git. I
> only expect minor changes in the near future.
>  ITP: https://bugs.debian.org/775520
>  Git: https://anonscm.debian.org/cgit/pkg-grass/netcdf-cxx.git
>  ITP: https://bugs.debian.org/775524
>  Git: https://anonscm.debian.org/cgit/pkg-grass/netcdf-fortran.git
> The netcdf-dbg package should become a transitional package
> recommending the C, C++ & Fortran debug packages, and the debugging
> symbols for NetCDF C moved to libnetcdf-dbg.
> libCF is still TODO


> http://www.unidata.ucar.edu/projects/index.html#libcf
> We may also want to package python-netcdf4, which has a different
> source from python-netcdf already in the archive.
> https://github.com/Unidata/netcdf4-python

Will do the ITP soon. I have googled around to see what the difference
is to the Scientific Python version. It seems that the netcdf-python
implements a much wider range of the netcdf functions, so it is
definitely worthwhile.



Attachment: signature.asc
Description: OpenPGP digital signature

Reply to: