In your previous mail you wrote:
But IPV6_ADDRFORM could be used only for a PF_INET6 socket.
=> it is easy to extend it to PF_INET but I agree the best solution
is to do a getsockname(). Specs should be modified in order to give
a defined answer as soon as the socket exists. Same for getpeername().
I guess the original question asked a method to distinguish a
"pure" PF_INET socket from a PF_INET6 one.
=> in the IPV6_ADDRFORM rationate there is the case the socket was passed
as an open file descriptor to the application then something is really
needed.
(I know you don't recommend programmers to use such two-socket usage
model, though.)
=> yes, I find it too complex for near no advantages, ie. the KISS
principle should be applied...
[EMAIL PROTECTED]
---------------------------------------------------------------------
The IPv6 Users Mailing List
Unsubscribe by sending "unsubscribe users" to [EMAIL PROTECTED]
