I'm trying to compile a package (LinCVS-0.3beta1 currently, but I've run into the analogous situation with other packages) for installation on our NFS filesystem, and running up against the limits of my knowledge of C. By default, the package expects to install its binaries in /usr/local/bin, shared files in /usr/local/share, etc. That can be changed in the "install" target of the Makefile (for example, replacing /usr/local with /opt), but when I try to run the binary it fails because it's still looking for data in the default locations. I could probably dig through the source and figure out where all of this stuff is set, but it would take an awfully long time and I'd have to repeat the process for every package that I compile for use on our network. This also means that using precompiled RPM's to install software on our NFS server is out of the question. Complicating matters somewat is the fact that our filesystem structure differs between the server and the clients; for historical reasons, the fileserver's /usr/local/export appears as /opt on all of the clients. That will need to be fixed eventually but it's that way for the moment. Is there a known way to approach this issue? Do C progamming practices offer a general solution, or am I really stuck rewriting every package by hand to account for our filesystem layout? _______________________________________________ Redhat-list mailing list [EMAIL PROTECTED] https://listman.redhat.com/mailman/listinfo/redhat-list
