This archive contains answers to questions sent to Unidata support through mid-2025. Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated. For the most up-to-date information on the use of NSF Unidata software and data services, please consult the Software Documentation first.
Harry Mangalam <address@hidden> writes: > Hmm - OK - I now see where that is mentioned in the web help pages - I > should have looked there first; but why not have configure set them? Well that is the intention, but the real world is surprisingly uncooperative in this area. > > Also, most of my experience has been on linux where there is a decided > preference towards building shared libs. For the netCDF build, as > well as the DODS libs, everything seems to be geared towards static > libs. Is there a reason for this? (I don't know - I'm just > wondering). We all bear the weight of history. NetCDF was developed before libtool, and before automake really got going. Only autoconf was used, and hand-crafted make files, which are rather complex. In 3.6.0 we upgraded to the latest autoconf version (we had been stuck at version 2.13, now we're at 2.59.) In 4.0 we add automake, and it helps a lot. (4.0 is due out this Summer, with oodles of exciting new features, but full backward compatibility with both classic and 64-bit offset netCDF files). Someday I would like to use libtool on netCDF. But when, or even if, that will happen, is something I can't say right now. It will not be in the 4.0 release, nor, I think, the 4.1 release (which is off the end of our schedule right now). > > The Fortran compilers were the g77-3.4, with f2c, and the g95 compiler >>From the g95 site, precompiled for AMD64. > Well I would have thought that would have worked out of the box. I have g95 around here somewhere, and I'll check to see what went wrong. Ed -- Ed Hartnett -- address@hidden