On Mon, 12 Sep 1994, L. Todd Masco wrote:

> Date: Mon, 12 Sep 1994 21:05:35 -0400
> From: L. Todd Masco <cactus@bibliob.slip.netcom.com>
> To: amiga@NetBSD.ORG
> Subject: Re: Summary of /etc/rc*
> 
> 
> Charles Ewen MacMillan writes:
>  >  Or, alternately, you can leave a single flag and "if" statement at the 
>  > beginning of /etc/rc.local to boot whatever daemons you please, without
>  > any changes to the current implementation.
> 
> Ah, but that *is* a change to the current implementation.
> 

 Yes, but it is a site-specific change, rather than a change to NetBSD
 source tree.

 I change many things on my own machines, to make them more compatible
 with the Suns, Macintoshes, etc.. that they must run with, that I would
 never suggest implementing as standard features.

>  >  On the BSD machines, there is less to run than on a Sun, so it seems
>  > to me there would be even less use for the setup you are describing, and
>  > in my opinion, not nearly enough usage to warrant the changes to
>  > init, changes to existing startup scripts, and more importantly drift 
>  > from the rest of m68k.
> 
> Maybe you missed it the first two times I said it:

> change in only the Amiga NetBSD.  I'll makw this available through the
> Amiga list, people can test it, and then I'll propose it back to
> everyone for *all* of NetBSD.  If it isn't taken up, then I'll
> distribute is as a patch to NetBSD in general.  IE, I'm not developing
> this just for the Amiga: I want to be able to use it on our VAXen when
> NetBSD is stable for them (You can't complain about bad UNIX
> implementations until you've administered Ultrix machines).
> 
> The changes to init are trivial -- and what you're saying is that we're
> better off with everyone making their own independent changes to
> startup scripts in incompatible ways.  Obviously, I strongly disagree.

 So far, you have:

1) Stated that my arguments were probably "religious" in nature, without 
   apparently, bothering to look at the justifications for those arguments.

2) Stated erroneously, against my protestations, that they were in fact
   "religious" in nature, thereby implying that the arguments could not
    bear any weight.

3) Implied that I do not know how to administer a site, due to a lack
   of knowledge admitted on my part, of who collaborated on SVr4, in
   a further attempt to invalidate my argument.

4) In the end, having bothered to read at least a few lines of what I
   was saying, summarized the argument as stating that we are all:
   "better off with everyone making their own independent changes to"
   "startup scripts in incompatible ways." I fail to see what is 
   incompatible about my implementation of these startup scripts, as
   after all, I have kept these very same scripts through several 
   upgrades.

   I could run any rc.local file that I have, on any machine in use
   on this site, which is I think a far greater compatibility than otherwise.

> 
> It's great that your own setup on your machines is so simple,
> that you never need to bring a machine partially up, and run no
> product- and other- related add-on startup scripts and daemons
> that you might want to modify in a system-safe manner.

 I clearly do run product related scripts, as stated in the previous 
posting. Doing so does not require init levels in any wise.

 Running them in a "system-safe" manner requires making them 
interruptable, which is already available in current.

> Not all of us have such simple setup.  A machine with flexfax, INN,
> named, and framemaker (for example) will have a family of things
> to run for each setup.  Any time you modify the scripts with each
> of these, you can screw your entire rc.local -- with run-levels,
> you'll only screw up the immediate family of daemons.
>

 Obviously, as I mentioned named and "news" explicitly in my next to 
previous posting, I am running both of these daemons, I said "news" for 
those less familiar with INND.

 Both daemons keep track of their own PID's in file format, hence neither 
requires more than three lines (counting the test to check for their 
existence) in any given rc script, and in the case of my system, they 
have seperate rc's which are run "intr."

 This pretty much prevents them from "screwing up any rc's."

 My system setup could hardly be called simple, it is simply efficient.

 
> I don't think it's good design to avoid a scalable startup system
> just because the base system isn't that complex (right now).
> --

 What is ever likely to be added to the base system, on a global level 
that is going to _require_ a scalable startup? Not likely anything.

 There are many things that should be implemented at _user_ or _site_ 
level that people apparently wish to be added into the operating system.

 Perhaps I will start arguing that everyone adopt my libc changes so 
that I need not rebuild it with each complete overhaul.... har.


Charles Ewen MacMillan  | Tezcat.COM - Wicker Park 
  <ilixi@tezcat.com>    | Offering Internet Access 
Modem: 312-850-0112/0117| Via Interactive UNIX to 
Voice: 312-850-0181     |    the Chicago Area.
WWW: http://tezcat.com/ | Mail: info@tezcat.com