>> Hence, I think getaddrinfo implementations SHOULD or maybe
>> even MUST return the site-local address before the global
>> address - otherwise site-local addresses will lose one of
>> their important uses.
>As stated before, given the lack of interface information in
>getaddrinfo, the source addresses selected for the destination
>address comparisons could be incorrect. Implementations may need
>to provide some way to compensate for this problem, even if its
>as ugly as a management knob to force use of global addresses
>when all else fails. That said, I have no problem with SHOULD.
my take was that, if the node/library obeys ipngwg-addr-select-01,
getaddrinfo/gethostbyname will (or should) return the addrsses ordered
appropriately.
for a normal UNIX implementation, ipngwg-addr-select-01 defines the
behavior of the kernel (source address selection) and the name
resolver library in libc (destination address seletion).
itojun
--------------------------------------------------------------------
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]
--------------------------------------------------------------------
