From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from emory.mathcs.emory.edu ([128.140.2.1]) by archone.tamu.edu with SMTP id <22552>; Wed, 3 Jul 1991 13:48:43 -0500
Received: from audfax.UUCP by
	emory.mathcs.emory.edu (5.65/Emory_mathcs.3.2.3) via UUCP
	id AA29148 ; Wed, 3 Jul 91 14:48:11 -0400
Return-Path: arnold@audiofax.com
Received: by audfax.audiofax.com (3.2/AUDIOFAX-1.1)
	id AA06373 for archone.tamu.edu!rc; Wed, 3 Jul 91 14:46:58 EDT
Message-Id: <9107031846.AA06373@audfax.audiofax.com>
From: arnold@audiofax.com (Arnold Robbins)
Date: Wed, 3 Jul 1991 13:46:56 -0500
In-Reply-To: Byron Rakitzis's 28-line message on Jul  3,  1:12pm
Work-Phone: +1 404 618 4281
Fax: +1 404 618 4581
Fax-Csi: ARNOLD DAVID ROBBINS
Fax-Mailbox: 581
Keep-Headers: none
Ultrix: Just Say NO!
Important-Saying: Premature Optimization Is The Root Of All Evil.
X-Mailer: Mail User's Shell (7.1.1 5/02/90)
To: rc@archone.tamu.edu
Subject: Re:  naming conventions

I'm quoting Byron w/o permission, since I thought the list might
be interested.  I hope Byron won't mind...

> /usr/lib/rcmain is the file that td talks about, and I have REAL problems
> with it.
> 
> BTW, td runs /usr/lib/rcmain on ALL invocations of rc, not just rc -l.

I agree that running it on all invocations is wrong.

> Have you seen what /usr/lib/rcmain does? It does the interpretation of all of
> rc's flags!! (-l, -c, and so on)

No, I haven't seen it; I don't remember it being in the v10 doc.  I'd like
to see a copy.  It's a pretty nifty idea putting that stuff outside the
shell itself.  Makes it *very* easy to add a new option.

Doing -e, -x, & -v must be a neat trick.

> I don't believe in /etc/profile for sh either; it's just a way for fascist
> sysadmins
> to rule the lives of their users. I refuse to incorporate such an addition
> into rc.

Ah well, we agree to disagree.  It's also a great way for a sysadmin to
make life easier for his or her novice users, but it's obviously religion so
I won't waste our time on it.

> As for dashdee, dashell, and so on, I think the meaning of the variable names is
> pretty clear. When in input.c I refer to the variable "dashex" I know that I'm
> talking about an option that was passed to rc at startup, and not just some random`
> boolean.

Okay, this is understandable.  I just know that I personally hate looking
at code I've never seen before and it being necessary to have the man
page (if there is one) handy.  I could argue that other naming conventions
would suffice, but won't push the issue.

Arnold