Subject: Re: pcvt and TIOCCONS
To: None <current-users@NetBSD.ORG>
From: der Mouse <mouse@Collatz.McRCIM.McGill.EDU>
List: current-users
Date: 04/20/1996 07:13:56
gwr quotes and writes:
>>>> The tty_name would be matched against the name of the current tty
>>>> line so all lines (and only those lines) with he tty_name field
>>>> matching would be processed.
>>> I like cgd's idea about supporting globs here.
>> Also agreed.
> This lets you have one record apply to many tty lines.
> Is this something people actually want to do or just a "creaping
> feature?"

And then mrg quotes that and writes
> this should not be done unless /etc/ttys' is also given the ability
> to have globs, IMO.

My opinion: I think globbing should be supported for tty name matching
in the tty_log{in,out} file, or whatever it ends up getting called.  I
see it as fundamentally different from /etc/ttys because /etc/ttys is
not just a lookup table but also the driving list for the getty
spawner (init, currently, but conceptually separate).  Supporting globs
in /etc/ttys would mean that the getty spawner would have to expand the
globs without anything to expand them relative to.  (Except perhaps the
contents of /dev, but I don't like that because losing an entry in /dev
becomes a silent failure mode instead of the present noisy one.)

If we separate the list-of-ttys for the getty spawner from the table
that describes those ttys, I would say the latter _should_ support
globs.

And I think the file under discussion here should support globs, since
it is entirely a lookup table.

One could argue that "globs" should really be "regexps", but I see no
need for that much generality.

					der Mouse

			    mouse@collatz.mcrcim.mcgill.edu