Re: Wrap up and last call: sin6_scope_id semantics

[YOSHIFUJI Hideaki / 吉藤英明]

In article <[EMAIL PROTECTED]> (at Thu, 16 Aug 
2001 17:31:26 +0200 (CEST)), Erik Nordmark <[EMAIL PROTECTED]> says:

> When you ship a new release that supports a link ids, site ids, etc
> won't you also ship a getaddrinfo and getnameinfo which can handle those?

We need some paths to keep backward compatibility on systems
such that kernel / library and applications may be developed / distributed
separately by respective maintainers.
ex.) linux kernel, glibc and other utilities such as ping6, net-snmp etc.

Atsushi Onoe wrote:
> I won't oppose that an implementation accepts the values which don't match
> the scope type, but at least it MUST NOT returns such values to applications,
> IMHO.

So just "MUST NOT" is not acceptable.
To leave meaning of 0-scope unspecified (for kernel<->userland) may be ok.

--yoshfuji


--------------------------------------------------------------------
IETF IPng Working Group Mailing List
IPng Home Page:                      http://playground.sun.com/ipng
FTP archive:                      ftp://playground.sun.com/pub/ipng
Direct all administrative requests to [EMAIL PROTECTED]
--------------------------------------------------------------------