From owner-cvs-etc Thu Mar 30 14:22:28 1995 Return-Path: cvs-etc-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id OAA00334 for cvs-etc-outgoing; Thu, 30 Mar 1995 14:22:28 -0800 Received: from saturn.landmark.com (landmark.com [192.246.113.2]) by freefall.cdrom.com (8.6.10/8.6.6) with SMTP id OAA00327; Thu, 30 Mar 1995 14:22:15 -0800 Received: from ren.landmark.com by saturn.landmark.com with SMTP id AA19896 (5.67b8/IDA-1.5); Thu, 30 Mar 1995 17:22:12 -0500 Received: by ren.landmark.com with SMTP id AA18238 (5.67b8/IDA-1.5); Thu, 30 Mar 1995 17:22:11 -0500 To: "Rodney W. Grimes" Cc: phk@ref.tfs.com (Poul-Henning Kamp), ache@freefall.cdrom.com, CVS-commiters@freefall.cdrom.com, cvs-etc@freefall.cdrom.com Subject: Re: cvs commit: src/etc/etc.i386 rc.i386 In-Reply-To: Your message of "Thu, 30 Mar 1995 13:37:14 PST." <199503302137.NAA06580@gndrsh.aac.dev.com> Comments: Hyperbole mail buttons accepted, v3.16. Date: Thu, 30 Mar 1995 17:22:05 -0500 Message-Id: <26684.796602125@ren.landmark.com> From: Harlan Stenn Sender: cvs-etc-owner@freebsd.org Precedence: bulk Is there some reason not to follow the SysV lead, and go for the Start/Kill mechanism using [SK][0-9][0-9].name ? Personally, I'd prefer 3 digits instead of 2, but that's just a nit. The issue of having a clean way for independent packages to get access to Start and Kill services is significant. There are times I think this whole issue should be handled with "make" and well-known labels instead of relying on the (dys)lexicographic order of files in a subdir. If a package needs NFS-mounted disks up before running, it should simply be able to specify this requirement instead of requiring somebody to check that this requirement is met, especially since this is most likely to bite people when an OS is upgraded. H