In your previous mail you wrote:

   > >> => you can use it with the V6ONLY stuff.
   > >yes, but on rfc2553-compliant system you cannot have both an AF_INET
   > >and an AF_INET6 socket listening on the same port.
   >
   >    (just a picky comment)  RFC2553 does not talk about the behavior
   >    when try to bind(2) to both :: and 0.0.0.0 on the same port.  some
   >    systems reject bind(2) to 0.0.0.0, some does not.
   
   ok. i will try to explain better my thoughts.
   
=> OK because I don't understand what you want.

   i don't want to change RFC2553. i think that V6ONLY is a very good feature
   and may be sufficient to achieve a good af-indipendence. however, this
   moves the focus on the behaviour of bind and getaddrinfo (which is not
   described in detail in RFC2553 - i suppose because there is no consensus).
   
   some implementations of bind(2) allow binding of 0.0.0.0 and :: address
   on the same port, and some do not. i think that this may be a problem for
   the developers which are working in multi-platform environments.
   
=> the V6ONLY stuff is just here in order to give a way to enforce the
same behavior on any compliant platform.

   i'd really like to minimize the negative effect of this behaviour, but i
   understand that there is not much that can be done about this.
   
=> one thing we can not do is to pick up an implementation and to make
all others non-compliant. When there is no way to get a portable code
for every compliant implementation we update the 2553 document. This was
done for BIND 9 filtering and gave V6ONLY stuff. We should not go further
without very good reasons, we should not break old codes too.

   i hope that we will come to a *BSD-like de-facto standard for the
   behaviour of bind and getaddrinfo, but i wonder if having no standard
   is even worse than having a bad standard (linux-like).
   
=> we have a specification (not a standard for the IETF, APIs are never
on the standard track) and many years of experiments with it.
Use it and close this discussion.

[EMAIL PROTECTED]
--------------------------------------------------------------------
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]
--------------------------------------------------------------------

Reply via email to