authd does not require autoconf. If needed, change any defaults by editing the config.h file. To build, simply run "make"
If you're using encryption, put a one line pass phrase in the file "/etc/ident.key" (or another place if you change the default location via a server option), making sure the file is readable by the authd process and NOT readable/writable by others ("chmod o-rw"). If the permissions are not set correctly, authd will refuse to encrypt.
A sample xinetd configuration file has been provided; copying xinetd.conf.auth to /etc/xinetd.d should work for Red Hat distributions. Be sure to make any changes needed to the default values and path as needed then restart/reload the xinetd daemon to use it.
All of the options available can be seen with the "-h" option. Some notes on some of the less obvious options and parameters:
If the "-n" option is also specified, then the numeric user id will be followed by the 2nd up to uint fields providing that uint is greater than two.
$ /usr/sbin/in.inetd 33201,6667
$ /usr/sbin/useradd -s /sbin/nologin -r somebody
$ chmod o-rw /etc/ident.key
authd will refuse to encrypt if this is not done.
$ /usr/bin/openssl enc -d -base64 -aes-128-cbc -pass file:/etc/ident.key
(Change the cipher to what's appropriate if you did not use the default for the "-E" authd option or the default was changed in config.h) Use enc's -in option if the base64 encryption is stored in a file rather than being piped into stdin)
Be aware that many system log daemons are not capable of handling non-ASCII yet, so combining this with the "-l" option may not produce readable syslog messages.
These only appear when authd is launched with the "--xerror" option, because some server administrators do not believe in giving outsiders any useful information regarding the state of their servers. However, the --xerror is useful for diagnostics and troubleshooting.