I'm looking at (and have been resisting the change) in xinet.conf two different ways of doing its "configuration".
The can be summed up (using neutral terms) as "Way-A (WA) and "Way-B" (WB) and "Way-C". In WA, we have something like: <local WA setup for xinet> <list of all xinet services specifying disabled or enabled> <then list of all services in config file to specify params for WA services> <WB:> <local WB xetup for xinet> <include <subdir> with separate files, specifying on/off and config for each for each service> <WC:> <local WB setup for xinet> <list of all xinet services specifying disabled or enabled> <include <subdir> with separate config options for each service> Advantages/Disadvantages (that I see). Method WA: everything together in 1 file so admin can see all services in 1 place. At a glance, a human can see what is enabled vs. not. File len = ~300 lines (relatively short). Dis: updating a service requires updating the base file. Method WB: each service in separate file allowing services to be inserted or extracted without any change to the base file. The config file can be included with the installed service -- ease of packaging and installation/removal. Dis: each service file must be examined to see if it is enabled or not. To disable or make changes, user mush cycle through 23 config files to edit or change all services. Method WC: (split method): Top config includes common config, but the enabled/disabled status is in the top level include, with the included directory containing only the service related config information. This allows "at a glance" to see what services are enabled or not, but parameters for each service are still in separate files. This allows 1-point view of enabled services, but puts all the config information separate that can be included in each installed packages' rpm. Dis: if a service installs itself and wants to "enable" itself as it is being installed, it must edit the top-level include. Each has advantages/disadvantages: Method WA is convenient for user to edit. Method WB is convenient for install scripts to add/remove services and automatically turn on-or-off at the as the installation file specifies. And WC, provides user with 1 glance for enabled services, but still requieres separate editing to set params. Install is as straight-forward as WB, unless install/remote service should toggle it on or off in master-service enable list. Even there, the top level include can be altered to automatically enable or disable upon installation. Method "WA" is most like the old inetd.conf in that all services and their control are in one file, but "WB" is easiest for the automated install scripts. WC still allows user to view all enabled or disabled services in 1 file, but has advantage that updating or adding service doesn't change run-params of any other service (though if auto-enable/disable still require editing the main file). Without basing your opinion on what just "happens" to already be in place (that's cheating! :-)), which of the 3 ways do people prefer? All-in-one, split/service (23 extra files in my config), or all-control-in-one file with params being split/service. Comments? Preferences? FWIW, I believe both WA and WB have been "done", but I don't know if WC has been tried. Thanks, -Linda -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
