Bug#160596: details?
Ok, so the proper behavior is fr getaddrinfo() to continue resolution; I
understand this. However, it's the order that it's probing the hosts
that is the problem. It begins probing by appending a dot to the host,
and looking up that. If it finds dns information for that host, it
returns that information. /etc/hosts has no notion of fully-qualified
domains; dots should be option. All it understands are aliases. The
dot-appended host should not be looked for in /etc/hosts; or, if it is,
then getaddrinfo() should continue processing w/ the non-dot-appended
host, even if it finds a result for the dot-appended host via the
resolver.
The bug isn't with how it probes for a host; the bug is with the host
that it probes for. Probing for "foo.com." in /etc/hosts, when asked for
"foo.com", is a bug.
If this is how the new lookup model is supposed to work, then it should be
properly documented. Currently, it's not.
On Wed, Sep 18, 2002 at 12:36:51AM -0400, Ben Collins wrote:
>
[...]
> >
> > Basically, how is this not a bug?
>
> It is not a bug, because /etc/hosts is not broken, nor is it's usage.
> The "problem" is that getaddrinfo is a new interface designed with new
> goals. When getaddrinfo does a "lookup" it does not stop when it has
> found the first match. Instead it keeps probing to find all info about a
> particular host, which includes A, AAAA, and reverse for both. This is
> how it is supposed to work. Just because this new mechanism does not fit
> into an old model, does not mean it is broken.
>
>
> --
> Debian - http://www.debian.org/
> Linux 1394 - http://www.linux1394.org/
> Subversion - http://subversion.tigris.org/
> Deqo - http://www.deqo.com/
--
Buying a Unix machine guarantees you a descent into Hell. It starts when
you plug the computer in and it won't boot. Yes, they really did sell you
a $10,000 computer with an unformatted disk drive.
-- Philip Greenspun
Reply to: