> > Well, I suppose the releases effectively Last Called SASL-IR, then. In > that case, I like Alexey's proposal better than Corby's: > > b authenticate (SID 1234432143) plain AGFybnQAdG5yYQ== > > (I'll be happy with either, though. And I wish people would mention it > on the list when they freeze a draft by releasing code widely.) > > Arnt
Agreed. The parens on the SASLIR would be a problem at this stage. If we can guarantee SASL Method and SASL IR pairing as Alexey suggests, then this is a minor implementation detail. Implementation wise it would be cleaner to make sure that the SASL-IR is right after the SASL method since the parser would only have to do a quick logic check. If the next parameter is in parens, then I can safely assume the SASL-IR is not there. Corby
