[prev in list] [next in list] [prev in thread] [next in thread] 

List:       mutt-dev
Subject:    Re: The future of mutt...
From:       Alexander Gattin <xrgtn () yandex ! ru>
Date:       2013-10-06 19:01:05
Message-ID: 20131006190105.GC32556 () x505 ! ckee
[Download RAW message or body]

On Sun, Oct 06, 2013 at 01:30:09PM -0500, Derek
Martin wrote:
> > My system has several IP addresses and several
> > hostnames (depending on interface/network). 
> 
> Wrong, it has exactly one hostname, as does
> every TCP/IP-networked host.  It may have
> several domain names corresponding to the
> multiple IPs on your machine... those are not
> hostnames, they are domain names.

You have several hostnames or A records or domain
names or whatever. Then you have `hostname`, which
is configured in kernel, at least in Linux
(cat /proc/sys/kernel/hostname), which may match
some A record, or not. Or match partially. Your
method does not even do that. It simply resolves
kernel's notion of hostname in configured domains,
and can still return A record the host doesn't
actually own.

Just imagine that there was no such thing as
`hostname` in kernel. How would you do your patch?
You could get all IP addresses of all interfaces
and try to resolve them via NSS/DNS, then take the
result you like the most (1st one for example, or
the longest one).

Or you could try to find the 1st/cheapest default
route and take source IP address corresponding to
this route (if there's such a notion in your
UNIX), and resolve it.

P.S. I do not propose to implement anything of the
mentioned above, it's JFYI.

-- 
With best regards,
xrgtn
[prev in list] [next in list] [prev in thread] [next in thread]