Summary/Minutes from today's FESCo Meeting (2013-09-11)

[Matthew Miller]

On Thu, Sep 12, 2013 at 02:18:23AM +0200, Lennart Poettering wrote:
> Are you suggesting it should pass that switch? If so, I'd prefer if this
> would become the agetty upstream default. Have you tried asking Karel
> for that? Why did you involve FESCO?

This went to FESCo because the bashrc maintainer didn't want to change the
bash prompt otherwise. (And we voted to leave that alone.)

I agree that the FQDN terminology was used loosely and sometimes incorrectly
in the discussion. We can revisit if you (or others) think it should be. I
don't think FESCo makes any claim to infallibility. (At least, I hope not!)

> How is that fqdn determined? Simply the result of gethostname()? Or is
> this the result of "hostname --fqdn"? (the former just reads a string
> from the kernel, which is fine; the latter involves DNS resolving that
> string, and then reverse resolving it again, which is nothing we should
> do when showing a login prompt.)

Based on the bugs presented in the ticket, despite the loose use of FQDN,
this appeared to be about putting \n in /etc/issue (after the initial idea
of calling agetty --long-hostname was discarded). The agetty man page says
that's nodename from `uname -n`, which is the underlying implementation to
gethostname().

So, in other words, this:

> Or are you simply suggesting that a possibly configured long kernel
> hostname should not be truncated on display to become a short name?

but for the specific case of /etc/issue. (Agetty's default is to truncate
the hostname prompt it gives, but I don't think it does that with '\n')

I did read the whole previous mailing list discussion, and the above is what
informed my thinking even though I didn't say so in the discussion. (And I
probably should have.)


-- 
Matthew Miller  ☁☁☁  Fedora Cloud Architect  ☁☁☁  <mattdm at fedoraproject.org>