Subject: Re: re-reading /etc/resolv.conf on change
To: Perry E. Metzger <perry@piermont.com>
From: Bruce J.A. Nourish <bjan+tech-userlevel@bjan.net>
List: tech-userlevel
Date: 12/31/2003 16:24:13
On Wed, Dec 31, 2003 at 05:19:18PM -0500, Perry E. Metzger wrote:
> Another possibility -- maybe we are looking at this wrong? There are a
> variety of conditions we might want to inform programs of ("you've
> just been awakened after a deep sleep", "the network addresses have
> all changed", "your resolv.conf is different"), and perhaps we should
> come up with a more general mechanism for telling programs about such
> things?
> 
> Then, if you alter resolv.conf, run a program announcing "I've altered
> resolv.conf" to the world and all will be well! :)

This touches on a general problem that exists in all the free Unices:
user <-> kernel communication. There is simply no general, event based
infrastructure to communicate changes in the system state.

"The network addresses have all changed," and all the other messages
you mention are in the same class as "someone just plugged in a PCMCIA
card/scanner/whatever," or "we just found a wireless LAN that we can
join."

Freedesktop.org (among others), Linux vendors, and some competent Linux
kernel people are starting to code solutions to this problem. One 
familiar and obvious approach would be to watch the Linux crowd for a 
year or so, and see what floats to the top in that time.

Of course, that doesn't help us _now_. I have no particular opinion on
the resolver hack, but it is symptomatic of a bigger, broader and
growing problem.
-- 
Bruce J.A. Nourish <bjan+public@bjan.net> http://bjan.freeshell.org