On Tue, May 07, 2013 at 06:52:36PM +0200, Lennart Poettering wrote: > As mentioned, nss-myhostname provides the scope ID field filled in > correctly to glibc. If this information is somehow lost on the way to > the caller of getaddrinfo(), then this is a fuckup in glibc, and that's > it.
Possibly, but I can't find any documentation, either internal to glibc
or from the IETF, suggesting that getaddrinfo() is expected to return a
non-default (i.e. non-zero) scope ID unless one was explicitly specified
by the caller. I also note that FreeBSD's getaddrinfo() behaves the
same as glibc's, in that the scope ID is set to 0 except in the specific
case where the caller passed a scoped numeric address as the first
argument.
Sections 2.1 and 2.2 of draft-ietf-ipv6-scope-api-00 ("Scoped Address
Extensions to the IPv6 Basic Socket API") are relevant here, as they
define the updates to getaddrinfo() for scoped addresses.
I'm going to re-assign this to the glibc maintainers for their comment,
but again, I don't think glibc is doing anything wrong.
noah
signature.asc
Description: Digital signature
