From owner-freebsd-current  Mon Jul 24 22:16:53 1995
Return-Path: current-owner
Received: (from majordom@localhost)
          by freefall.cdrom.com (8.6.11/8.6.6) id WAA09536
          for current-outgoing; Mon, 24 Jul 1995 22:16:53 -0700
Received: from genesis.atrad.adelaide.edu.au (genesis.atrad.adelaide.edu.au [129.127.96.120])
          by freefall.cdrom.com (8.6.11/8.6.6) with ESMTP id WAA09524
          for <current@FreeBSD.org>; Mon, 24 Jul 1995 22:16:45 -0700
Received: from msmith@localhost by genesis.atrad.adelaide.edu.au (8.6.9/8.6.9) id PAA01801; Tue, 25 Jul 1995 15:04:47 +0930
From: Michael Smith <msmith@atrad.adelaide.edu.au>
Message-Id: <199507250534.PAA01801@genesis.atrad.adelaide.edu.au>
Subject: Re: Knobs in /etc/sysconfig
To: jkh@time.cdrom.com (Jordan K. Hubbard)
Date: Tue, 25 Jul 1995 15:04:47 +0930 (CST)
Cc: msmith@atrad.adelaide.edu.au, current@FreeBSD.org
In-Reply-To: <3128.806642718@time.cdrom.com> from "Jordan K. Hubbard" at Jul 24, 95 08:25:18 pm
Content-Type: text
Content-Length: 2803      
Sender: current-owner@FreeBSD.org
Precedence: bulk

Jordan K. Hubbard stands accused of saying:
> > Please ensure that the default "yes" values are documented in the comments
> > preceeding each knob; preferably in some standard format, eg:
> > 
> > # default: <value>
> 
> Sounds reasonable..  We still need to go through and ethnically clense
> our handling of variables in general, actually.  It would be nice if
> there was a STANDARD NAMING CONVENTION that told one whether or not a
> given variable was a Boolean, a list of arguments, a list of things
> that had their own sub-variables (like the network_interfaces var),
> etc.

Hmm, configuration files : the bane of my existence 8)

Before I/we get too rampant on this, what's the feeling (in particular
from Rod, of course 8) on imposing some structure on the layout of
/etc/sysconfig?

A first cut would look something like this :

Lines beginning with '#' are comments.  Lines beginning with '#*' are
structure comments.

#* Section: Keyboard
...
#* Keyword: KBD_repeat_rate
#* OneOf: "No change"=none,"Fast"=fast,"Medium"=medium,"Slow"=slow
#* Default: none

KBD_repeat_rate = none
...
#* Section: Networking
...
#* Keyword: NET_interface_list
#* String:
#* Default: "ed0 lp0"

NET_interface_list = "ed0 ed1"
...
etc.

It's impossible to embed _all_ of the intelligence required into these
comments, and overdoing it is harmful to your sanity, but minimising the
keyword-specific intelligence required for processing would be nice 8)

Some other proposals : variable name prefixes.  I've already proposed 
KBD and NET, I'll add PKG_pkgname for package-specific stuff (in a
Packages section).

As far as types are concerned, I see current use of :

Boolean : ON/OFF
OneOf : one of a,b,c,d
ListOf : none or more of a,b,c,d
String : some string value
StringBoolean : some string value or OFF (or NO or something similar).
FQDN : qualified name (dotted address style)
IPADDR : a dotted-quad value.
Pathname : a pathname

I can see possible use for :

Mailaddr : a mail address
Device : a device name

And I'm sure I've missed some.

The alternative to the structure comments is a hints file, but keeping the
two in sync would be a major nuisance, and would really only hit the
limelight if a proposal similar to the above was rejected.

> Probably the time to do that is now before it gets too big, but I'm
> also not going to hold my breath.  It's work.

Not as much as some other things.  Thoughts?

> 					Jordan

-- 
]] Mike Smith, Software Engineer        msmith@atrad.adelaide.edu.au    [[
]] Genesis Software                     genesis@atrad.adelaide.edu.au   [[
]] High-speed data acquisition and                                      [[
]] realtime instrument control          (ph/fax) +61-8-267-3039         [[
]] My car has "demand start" - Terry Lambert                            [[