Hi, It sounds like Dave proposes some updates to the netcdf package. This would be great as it is not actively maintained. Our NCO project has debs which adhere to recent Debian policy. Rorik Peterson wrote most of this support. However, NCO depends on netcdf version 3.6.0 so we have been unable to build our packages for well over a year because the netcdf deb is version 3.5.1, as you can see here
http://nco.sf.net#debian So if you do update netcdf to 3.6.0, then we will be one of your major customers. Actually you should go the whole way and update to 3.6.1 which supports newer netCDF file types. You might look at the NCO debian directory http://cvs.sourceforge.net/viewcvs.py/nco/nco/debian for ideas on how to update the netcdf package. If you have any specific questions, I will try to help. > maybe converting to debhelper and adding f90 support. Reminder that f90 support is tricky because it, like C++, has name-mangling (for the .mod header files). If you decide to include f90 support then you will need to add an F90 compiler dependency. I would prefer g95 over gfortran because g95 is useful for real science now. In any case, these are the flags we build netcdf with: export CC='gcc' export CXX='g++' export CPPFLAGS='-DNDEBUG' export CFLAGS='-O -g -fno-common' export FC='g95' export F90='g95' export FFLAGS='-O -g -fno-second-underscore' export F90FLAGS='-O -g -fno-second-underscore' Note the no-second-underscore option is not-default but improves souce code portability for C<->fortran90 interaction because most other OSs (Solaris, Irix) use a single underscore, not the default double underscore which G77 originally chose. Please ping me if you ever do update netcdf and we will beta-test that your build works by trying to build NCO packages off it. P.S. The Unidata udunits package has a similar structure to netcdf and I encourage you to try packaging if you have success with netcdf. Thanks, Charlie -- Charlie Zender, [EMAIL PROTECTED], Department of Earth System Science 3228 Croul Hall, UC Irvine, Irvine CA 92697-3100. (949) 824-2987 :) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
