Re: glibc's getaddrinfo() sort order
Anthony Towns writes ("Re: glibc's getaddrinfo() sort order"):
> On Tue, Sep 18, 2007 at 05:09:43PM +0100, Ian Jackson wrote:
> > Andreas Barth writes ("Re: glibc's getaddrinfo() sort order"):
> > > [stuff]
> > So, trying to keep stuff simple, I propose:
> > 1. RFC3484 s6 rule 9 should not be applied to IPv4 addresses
> > by Debian systems, and we overrule the maintainer.
>
> By making "sortv4" default to yes when unspecified in gai.conf? Does
> this need to change for stable too?
AIUI stable works properly already. If not then I think it does need
to change in stable too.
I don't want to mandate how the effect is achieved. If the glibc
maintainer prefers to do it some other way then that's fine.
> > 2. RFC3484 s6 rule 9 should not be applied to IPv6 addresses
> > by Debian systems, but we do not overrule the maintainer
> > on this point.
>
> Why are we weaker on this point? Should we simply say "wrt IPv6 addresses,
> the RFC should be followed", and leave the behaviour we think proper
> up to the results of:
My proposal is weaker on this point because in the case of IPv6 rule 9
represents less of a divergence from historical behaviour - because
there is less of a history of what the behaviour should be. There are
still in most of the world relatively few IPv6 deployments. It is not
wholly unreasonable to change the behaviour when switching to a new
protocol.
That means that to answer this question definitively we would need a
proper understanding of the current state of the IPv6 addressing
architecture and the current view in the IETF and amongst other
implementors (since obviously all systems need to do the same thing).
I don't think we presently have that understanding and I don't think
it's urgent that we should try to acquire it and make a definitive
ruling. (I do think it's clear that glibc upstream has a wholly
flawed understanding.)
> > 3. We recommend to the IETF that RFC3484 s6 rule 9 should be
> > abolished, definitely for IPv4, and probably for IPv6 too.
>
> ?
WDYM `?' ?
Obviously the Debian TC can recommend something to the IETF. What
that means is that the people who brought this question to our
attention can go back to the relevant IETF WG(s) with not just the
arguments we've developed but also a decision from us to prove they're
not a kook.
> I don't disagree, I'm just not convinced that the behaviour is bad
> enough to warrant violating the standard or diverging from upstream.
This behaviour _actually broke our own production servers_.
How bad does it have to get ?!
Furthermore, the broken behaviour is NOT THE STANDARD.
LET ME SAY THAT AGAIN: THE BROKEN BEHAVIOUR IS *NOT THE STANDARD*
LET ME SAY THAT AGAIN: THE BROKEN BEHAVIOUR IS *NOT THE STANDARD*
LET ME SAY THAT AGAIN: THE BROKEN BEHAVIOUR IS *NOT THE STANDARD*
LET ME SAY THAT AGAIN: THE BROKEN BEHAVIOUR IS *NOT THE STANDARD*
LET ME SAY THAT AGAIN: THE BROKEN BEHAVIOUR IS *NOT THE STANDARD*
> I'm suspicious that the standard doesn't specify what's actually
> been implemented for the past decade or so too though, but haven't
> verified that.
Are you doubting what we're saying about what the de facto standard
is ?
Ian.
Reply to: