[ietf-nntp] Re: AUTHINFO/SASL responses
rra at stanford.edu
Tue May 18 00:32:16 PDT 2004
Clive D W Feather <clive at demon.net> writes:
> Russ Allbery said:
>> Not a new code, please.
> Why not?
NNTP has too many codes to keep track of and handle already. It's a real
pain when writing clients.
> It doesn't feel right to me, though I'm not going to die in the ditch on
> this one. But surely it's better to have one return code and have some
> way to distinguish "no data" from "empty data" in the argument?
Think of it from the client perspective. Using this method, it can just
treat 281 as "authentication succeeded" and 283 as "authentication
probably succeeded but I need to check something." If we come up with a
new code that combines the two, it both has to make another check *and*
still handle 281 for legacy AUTHINFO USER/PASS.
Russ Allbery (rra at stanford.edu) <http://www.eyrie.org/~eagle/>
More information about the ietf-nntp