James Carlson wrote: ...
I think we need a refactoring.
Indeed. If we had a do-over, I think the right model would have been to just let inetd dynamically create service instances directly from inetd.conf, rather than requiring the two-step convert & import which doesn't add any real value. The result would be fairly similar to how the lrc services are handled.
Less drastically, though, we might just have inetconv not bother doing validation when it's being run (perhaps only in the inetd-upgrade case with an environment variable or new command option), and then let inetd handle it afterwards, which would get us closer to what Jim suggested - if the service is incorrectly specified, then it'll end up in maintenance.
Dave _______________________________________________ networking-discuss mailing list [email protected]
