similar to Solaris's "-r <repository" or Mac OS X's "-i <infosystem>",
to select the password database (files, nis, krb5). Otherwise, we default
to using whatever PAM decides.
XXX: This avoids the issue of supporting separate -l -y -k, but is the behavior
correct? Should passwd -p disable all other passwd methods? Should it become
the default if compiled in?
If a user wishes to change a password on a system running YP, and
the master server is not running rpc.yppasswdd, passwd(1) would fail,
even if the user had a local entry. Fix this by checking for local
entry if master is not running rpc.yppasswd iff we defaulted to using
YP (not invoked as "yppasswd" or with "-y").
YP server. This is required if the passwd database is to stay in sync
if this program is run on the YP server. Note, local passwd database
operations can still be performed by passing the -l flag.
Also, some minor cleanup and RCS id police.
