>>>>> "Hamish" == Hamish Moffatt <[EMAIL PROTECTED]> writes:
Hamish> These problems should be solved by discussion and Hamish> generation of a policy. IMHO it would be better to have a Hamish> consistent approach that didn't solve every problem (or Hamish> had some other flaw) than to have each individual Hamish> developer generate their own scheme. I agree. Would it be feasible to have something like "update-alternatives", but instead of managing files in the file system, it allocates port numbers? That way every service that listens on port, for example 143, will be registered, but only one will be "active" at one time. When a service is activated, then this means installing the inetd line if it is an inetd based daemon or activating an /etc/init.d/* script (by renaming the appropriate links) and starting it. When a service is deactivated, the reverse happens. Or perhaps a running some configurable script for each activation/deactivation might be better, that way a package could request deactivation mean "remove listen config option on this port". Might be useful for packages like apache which can (and sometimes do) listen on other ports. -- Brian May <[EMAIL PROTECTED]> -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]